Here - Colgate University

The Colgate Academic
Review
Volume VII
Spring 2010
i
This marks the seventh edition of The Colgate Academic Review, and we at the Student
Lecture Forum could not be more pleased with how well the project this publication
represents has grown and evolved over its nearly four years of existence. The ensuing 270
pages contain arguably the best of what Colgate’s student body has to offer in academic
work from the Spring, 2010 semester. Papers were selected based on the results of an
academic conference, held April 18, 2010 in the Ho Science Center. Students submitted
abstracts of their work and presented their findings to panels of student and faculty
judges in three of Colgate’s four major academic divisions: Social Sciences, Natural
Sciences, and Humanities. The fourth category, the liberal arts CORE Curriculum,
received no submissions this semester and so has, unfortunately, been omitted. Three
submissions were selected as the best in their respective division, with one additional
paper judged overall “best in conference.” This winner, Allison Patchen, was awarded the
George E. Stevenson prize from the Colgate Speaking Union.
The fact that so many authors in the current edition hail from the Class of 2010 is
somewhat bittersweet: those of us who remain at Colgate will miss their insights and
acumen in the semesters to come. However, we celebrate the fact that the wider world is
now the beneficiary of their talents, and we are confident that new lights will reveal
themselves in short order. We eagerly anticipate cataloguing their future academic
accomplishments.
We hope you enjoy engaging with the ideas inside this volume as much as we have.
-Andrew Pike, Class of 2012
Editor
ii
The Student Lecture Forum dedicates this volume of The Colgate Academic
Review to Conor Tucker, Class of 2010. During his time as a student at
Colgate, including two years as President of the Student Lecture Forum, Conor
exhibited exceptional dedication to the ideals upon which this organization
was founded: scholarship, inquiry, and a fierce curiosity extending far beyond
the bounds of the classroom. We wish him all the best in his future endeavors.
iii
iv
Table of Contents
Best in Conference:
Page
Allison Patchen, Threats and Opportunities: Experiences
in Public Housing in Utica, New York
1
Best Social Sciences Submission:
Mark Kanter, Implicit Discrimination in Major League
Baseball: Marginality Decisions Based on Minor League
Offensive Production
27
Best Natural Sciences Submission:
Courtney Walsh, Narrative Processing Predicts WellBeing and Communion in Victims of Intimate Partner
Violence
44
Best Humanities Submission:
85
Max Counter, The Dynamics of Doubt
Other Superior Entrants:
Colleen Tubridy, Occupational Diversity of Immigrants
in the United States
108
Dan Lasker, Southern Racism and Reality: A Case Study
134
Jess Mawhirt, Segregation and Incarceration: How Life in
the Ghetto Leads to Life in Prisons for Young Black Men
152
Marissa Dombovy-Johnson, The Effects of Taurine and
Caffine Alone and in Combination on Locomotor Activity
in the Rat
169
Michael LoFaso, Measuring the Persistence of Output
Shocks: A Study of Output Behavior using ARMA and
Monte Carlo Methods
188
v
Vildana Hajric, A Generalized Leslie Matrix Model for
Projections of Area V Humpback Whale Population
Demographics
210
Mary Liz Brady, Validating Foreign Language Issues:
What are the Effects of Non-English or Bilingual
Population on an MSA?
224
Kate Gundersen, www.youarewhatyoueat.com: The
Religious Nature of Healthy Eating Blogs
238
vi
Threats and Opportunities: Experiencing Public Housing
in Utica, New York
Allison Patchen, Class of 2010
One of the most challenging effects of living below the poverty line is difficulty finding safe
and affordable housing. In the early 1990s, the US government sought to remedy what
social scientists call “concentrated poverty,” said to cause heightened crime rates, lower
employment levels, and lesser educational attainment in cities. The result was Housing
Opportunities for Everyone (HOPE VI), which is designed to replace low quality housing or
abandoned buildings with new housing structures. For many people who live in cities,
public housing is the only affordable option. Utica, a city that houses a large number of
refugees, was one of the select cities in the United States to receive funding from the HOPE
VI project. Through six diverse focus groups with residents of the Utica Municipal Housing
Authority, I ask why public housing residents are not yet self-sufficient. In my analysis, I
find that residents prefer to live in public housing for a variety of reasons, that many
residents are under-informed about opportunities available to them, that they depend on
family networks for support, and that there is a need for support services such as
citizenship classes for refugees.
“I don’t even think about (moving out]. I think I’m gonna die there.
I don’t want to pay heating and lights.” – Doris, 58
Introduction
Many families living at or below
the poverty line struggle to find decent
and affordable housing. Some are just
one paycheck away from homelessness.
Public housing became a part of the
mission of the US government during the
Urban Migration that took place from
the 1940s through the 1960s. These
housing units quickly became “urban
ghettos” (Sugrue 1996). In the early
1990s, the US government sought to
remedy what many social scientists
called “concentrated poverty,” which
causes negative effects on a community
such as heightened crime rates, lower
employment
levels,
and
lesser
educational attainment. The result was
Housing Opportunities for Everyone
(HOPE VI). HOPE VI, initiated in 1993, is
designed to place new housing structures
in places where low quality housing
already exists or where abandoned
buildings are currently standing. The
goal is not only to change the nature of
the space, but also to change the
economic structure of the community by
bringing in residents of higher income
levels. Furthermore, it includes a
“community and supportive services
component designed to help raise the
incomes of public residents and move
them on a path towards financial selfsufficiency” (Reig, Liebow, and O’Malley
2006, 192).
1
Public
housing
has
been
established throughout the country to
provide decent and safe rental housing
“for eligible low-income families, the
elderly, and persons with disabilities”
(“HUD’s Public Housing Program”). An
individual’s eligibility depends on his or
her annual gross income, status as
elderly, disabled, or as a family, and US
citizenship or eligible immigration status
(1). Rent or Total Tenant Payment (TTP)
is based on a family’s anticipated gross
annual income less deductions for
dependents, elderly family members or
persons with disabilities. TTP is
determined as the highest of either (a)
30% of monthly adjusted income, (b) 10%
of monthly income, (c) welfare rent, if
applicable, or (d) minimum monthly
rent, as established by the Housing
Authority, at times as low as $25.
According to the HUD website, an
individual may stay in public housing as
long as he or she complies with the lease.
An individual or family will not be
required to move from public housing
unless there is affordable housing
available on the private market, based on
income.
refugees and immigrants occupy a
significant proportion of the subsidized
housing units of the Utica Municipal
Housing Authority. I thought these
residents might have different limiting
factors that cause them to remain in
public housing than the residents who
have been living in public housing for
many years. As of 2007, the median
household income in Utica was $30,971
with 24% of people living below the
poverty line.
In this paper, I argue that few
residents have goals of self-sufficiency
because many have larger issues that
they are dealing with and because push
factors for leaving public housing are
relatively absent. Essentially, my research
question is: What is keeping current
public housing residents out of the
private market? Why are they not “self
sufficient”? These questions had been
negotiated with the grant-writer, Mr.
Furman, and executive director, Dr.
Herbowy, of the Housing Authority prior
to the beginning of the research. Each of
the six focus groups was targeted at a
specific population, which gave me
information about senior citizens, young
adults, residents of each housing area
owned by the Utica Municipal Housing
Authority, and the two major refugee
populations in the area, Russians and
Somali Bantu.
For this project, I examine the
efforts of the Utica Municipal Housing
Authority (UMHA) in meeting the needs
of residents of its housing units and the
HOPE VI target area. The UMHA
administers approximately one thousand
public housing units as well as the HOPE
VI target area. The public housing units
range from one-bedroom apartments to
six bedroom apartments. There are
attached apartment buildings, high-rise
apartment buildings, and duplexes that
serve families, elders, and people with
disabilities. The large population of
In this paper, I focus on urban
poverty, its diverse manifestations
among people, and how the Housing
Authority attempts to tackle the related
issues. Across the focus groups, issues of
cost and safety are two of the most
common reasons people cite for living in
public housing. First, I will discuss the
2
theoretical framework I employed in the
analysis of these focus groups, focusing
primarily on theories of concentrated
poverty, social capital and human capital.
Next, I review the research previously
completed on the HOPE VI project and
Family Self Sufficiency across the country
and in Utica. Finally, I identify my
contribution to the body of research on
the topic.
Latinos, immigrants, and refugees, all of
whom are part of the urban underclass.
Despite the change in the reality of
experiences for the urban poor, many
policies and programs, including the
family self-sufficiency program of the
UMHA, remain rooted in the concept of
public housing as a transitional step in
the road to homeownership.
Social disorganization can also be
defined as a lack of social capital.
Coleman’s (1988) concept of social
capital “allows taking resources and
showing the way they can be combined
with other resources to produce different
system-level behavior or, in other cases,
different outcomes for individuals” (s101).
The concept of social capital is useful in
assessing the threats and opportunities
for self-sufficiency present in public
housing. His theory of social capital
involves
three
subsets.
Social
relationships that provide information
that facilitates action are one type. The
second form is norms and effective
sanctions. Prescriptive norms, or “norms
that one should forgo self interest and
act in the interests of the collectivity”
strengthen families and generally
encourage people to work for the public
good (S104). This theory states that
effective norms can reduce deviant
actions that harm others. The final form
of social capital according to Coleman
depends on two essentials: reliability of
the social environment, which implies
that commitments will be upheld, and
the actual extent of commitments held.
Impoverished urban centers are often
lacking in all or nearly all of these
categories. In Utica, the majority of
public housing units are part of large
complexes physically located apart from
Theoretical Framework
Social disorganization has been
discussed as a threat in poor urban
neighborhoods since 1925 (Park &
Burgess, 1925). In that era, public
housing was seen as a transitional and
functional part of the metropolis “where
new immigrant groups would pass
through for a temporary period of time”
(Curley, 2005, Park & Burgess, 1925).
However, since the mid-1960s, social
conditions for the “urban underclass”
have worsened. Wilson (1987) and
Sugrue
(1996)
argue
that
the
suburbanization of jobs and a lesser
demand for low skilled labor contributed
to a “downward spiral for urban blacks.”
Furthermore, economic integration of
urban areas lessened with the onset of
the “democratization of suburbia.” Here,
the loss of entry-level jobs, the loss of
middle class community members (both
black and white), and capital flight to the
suburbs causes concentrated poverty.
Wilson (1987) argues that as time
progresses, the urban blacks became ever
more isolated from informal job
networks, employed role models, and
mainstream patterns of behavior (Curley,
2005). In Utica, concentration effects
may effect more than just the urban
black population, but also poor whites,
3
private housing, which may contribute to
a lack of social capital as discussed
above.
negatively stereotype all of its occupants
or decrease the quantity or quality of
resources they invest there (Galster,
Cutsinger, Malega 2008). These effects
can be seen in Utica in the poor
condition of the roads surrounding some
communities of the UMHA, and the poor
quality of parking lots and public space.
The effects of social capital and
social disorganization can be best
understood through the framework of
concentrated poverty, which was the
primary motivation for the development
of the HOPE VI project. Popkin, Levy,
and Buron (2009) define concentrated
poverty as “communities with poverty
rates of 40 per cent or more.” Galster,
Cutsinger, and Malega (2008) find that
those living in disadvantaged, inner-city
neighborhoods with high levels of
poverty and social disorganization have
poorer health, lower levels of educational
achievement,
fewer
employment
opportunities, increased likelihood for
gang recruitment, and heightened
exposure to violence when compared
with people living in more advantaged
neighborhoods.
Motivation, a prominent issue in
the discussion of poverty, is essential in
examining self-sufficiency in the context
of public housing. Although some argue
that federal policies serve to subsidize
dependent behavior that keeps people
poor, the motivation theory gives a
different perspective.
Learned helplessness implies a
motivational deficit (people come to
believe action is futile), cognitive
interference (people will have
difficulty learning that action can
produce favorable results in new
situations) and an affective reaction
(depression or resignation). Further
[…] the perception of helplessness
can
be
inferred
from
the
environment,
without
direct
experience of failure (Kane 1987: 411)
Concentration
effects,
often
associated with public housing, are a
major threat to an individual’s growth.
South (1999) argues that concentration
effects are the results of the high level of
social
isolation
from
mainstream
patterns of behavior due to the lack of
sustained interactions with “steadily
employed persons from stable families.”
Another side effect of concentrated
poverty is institutional decay, or
decreased
financial
support
and
leadership in a community due to the
emigration of residents who used to
provide
these
services
(Galster,
Cutsinger, Malega 2008). Stigmatization
of a community, another concentration
effect,
occurs
when
institutional,
governmental,
or
market
actors
Related to this is the self-determination
theory, used to explain an individual’s
lack of motivation or internalization. It
entails examining individuals’ social
contexts
their
developmental
environments to examine the degree to
which their needs for competence,
autonomy, and relatedness are being or
have been thwarted (Ryan and Deci
2000, 74). Thus, an individual may
become passive or unmotivated as a
result of the situation in which he or she
grew up. Self determinacy theory and
motivation
theory
examine
the
4
“therapeutic value of choice” and the
difficulties an individual may have
making choices, especially if one does
not expect that he or she can personally
help attain it (Kane 1987: 410). How do
neighborhoods affect an individual’s selfmotivation and expectancy? What effects
does public housing have on how
children are socialized?
VI funds, buildings must be in need of
drastic improvement, a neighborhood
must show signs of distress, and
substantial matching funds must be
available to the community (Rosenthal,
2004). Only 28 housing authorities
nationwide received HOPE VI funds. The
Utica Municipal Housing Authority
received a grant totaling $11,501, 039 and
sought to use this money to replace
Washington Courts with new subsidized
housing units in the Cornhill district.
The premise of the project is to decrease
the concentration of poverty in at-risk
urban centers and increase social capital
through “scattered site” housing that
would promote the economic integration
of these communities (Gilderbloom
2008). The HOPE VI project also sought
to help residents attain financial selfsufficiency (Popkin et al 2009). This
differed from the traditional goal of
public housing: to provide decent,
affordable housing; now public housing
could become a “means to economic
independence” (Newman and Harkness
2002: 21).
Massey and Denton (1993) argue
that federal housing policy promoted
disinvestment
in
black
urban
neighborhoods and further segregated
cities by placing public housing projects
in
predominantly
poor
black
communities. They point out that this
hyper segregation leads to social
isolation, a theme in social theory on
urban plight. As one of the expressed
goals of HOPE VI is mixed-income
housing, the hope is that social isolation
will
be
combated.
Theoretically,
improving the SES of the most distressed
neighborhoods will result in a greater
diminution in social dislocations than
proportionally
comparable
improvements in other types of
neighborhoods (Crane 1991, as quoted in
South 1999). Crane argues that social
problems are contagious and are spread
through peer influence; this suggests that
social problems will be alleviated
through positive peer influence. To what
extent has socioeconomic integration
occurred thanks to the efforts of HOPE
VI? Has it had the desired effects?
Recent empirical data suggests
that there are some benefits of mixedincome housing, but it does not have
“the transformative effects policy makers
and scholars had envisioned” (Popkin et
al. 2008). For the HOPE VI project, many
residents had to be relocated during the
rebuilding or renovation of existing
housing projects. Popkin et al. (2008)
found that the largest and most
important effect of HOPE VI relocation
was the reduction in the fear of crime.
This study shows that there were no
changes in self-sufficiency (employment
rates) for residents, for whom health
problems are some of the major obstacles
Literature Review
Extensive
research
and
evaluations have been done on HOPE VI
throughout the country over the past
sixteen years. In order to receive HOPE
5
to employment.
“A housing-only
intervention may not have been
sufficient to address the serious, chronic
health needs of these vulnerable
residents” (Popkin et al, 493).
remains very distressed economically.” A
major critique of HOPE VI in Utica was
“the perceived failure of HOPE VI to
building anything but houses” (OwensManley 16). At the time of the study,
education and job training were
significantly lacking in the community.
One of the major issues Cornhill
Residents had with HOPE VI was that
they were not involved enough in the
planning process and were not kept
informed its proceedings. Residents
stated what has been demonstrated
throughout HOPE VI projects across the
country: policy makers seek to minimize
cost and maximize efficiency and are less
concerned with the furtherance of public
values (2003, “When Hope falls Short”).
HOPE VI is structured as a set of
contractual relationships with local
public housing authorities, which in turn
contract
private
developers
and
management companies, which was the
case in Utica. Although HOPE VI
requires the input of the residents,
private
companies
and
Housing
authorities often refuse to give them
meaningful opportunities to provide
their input (2003).
Reig, Liebow, and O’Malley (2006)
examine the outcomes of HOPE VI in
Seattle, where the mix of public housing
was similar to that of Utica, NY,
including senior housing, public housing
apartments, affordable homeownership
units, and a community center. They also
had commercial space. Like Utica, the
Rainier Vista, the section of Seattle that
received funding, residents included
immigrants from East Africa as well as
African American and American born
white residents. They find that in Seattle,
the HOPE VI project was combined with
Jobs-Plus, another program focused on
helping low-income individuals become
self sufficient. They find that a “formal
body representing tenants on matters
concerning housing and community life
in their development,” the a Rainier Vista
Resident Leadership Team (RVLT) was
developed. After elections, the RVLT
board of directors was representative of
the development’s ethnic and national
groups (197). Although this form of
community
development
seemed
successful, residents who participated in
the
job-plus
program
left
the
development and moved to a different
part of the city (199). This study shows
that helping residents find and maintain
jobs is a complicated and difficult
endeavor.
The
Owens-Manley
study
identifies another major problem related
to HOPE VI: the perception that one
must be in the “in” crowd to be
considered
for
the
reconstructed
housing. That is, certain individuals are
favored others based on an official’s
discrepancy. Marquis and Ghosh (2008)
also find that there is a bias in the
selection of which individuals are
accepted for resettlement in the new
communities and which are relocated to
other communities. This coincides with
the finding that “hard to house” residents
Research overseen by OwensManley (2007) has indicated that HOPE
VI has “fallen short of the original hopes
an[d] expectations, and the community
6
who have to deal with multiple complex
problems such as mental illness, severe
physical illness, substance abuse, large
numbers of young children, weak labor
market histories and criminal records are
less likely to realize meaningful
improvements in quality of life (Popkin
et al 2009, 494).
private housing, often in the form of
homeownership (2005). Despite the
reported successes of FSS programs,
some experts criticize the concept of
“self-sufficiency.”
Hawkins
(2005)
questions the legitimacy self-sufficiency
as a goal because it is difficult to define,
difficult to evaluate, and “limited,
unattainable,
and
insufficient for
policymaking” (79). However, Hawkins
argues that self-sufficiency should reflect
four dimensions: 1) autonomy and
responsibility, 2) financial security and
responsibility, 3), family and self wellbeing, and 4) basic assets for living in the
community (2005). He argues that
decreasing long-term poverty and
welfare depends on a combination of
many factors, so all these must be
treated. Hawkins however, does not
evaluate current FSS or PFS (personal
and family sustainability) programs.
Some of these variables do not
play a part in Utica, as individuals with
criminal records are not permitted to live
in the public housing units. In 2002, a
study was published that examined the
long-term effects of public housing on
self-sufficiency. Newman and Harkness
(2002) find that people who had lived in
public housing between the ages of 10
and 16 had higher annual incomes and
were about 7% more likely to be working
than other individuals who were raised
in families with comparable incomes.
However, Sanbonmatsu et al (2006)
studied the effects of the Moving to
Opportunity program, which showed
that “there was no difference in the
children’s achievement test scores or
measures of school or behavioral
problems for any age group” (Mouw 98).
This implies that living in “concentrated
poverty” may have little or no effect on
children’s academic outcomes.
The needs assessment for the
Utica Municipal Housing Authority
builds upon the work of Owens Manley
(2007), Anthony (2005), and Hawkins
(2005) and examines the new needs and
obstacles residents face after this
community reconstruction. It evaluates
the current attitudes of the residents of
the UMHA’s subsidized housing units
through the use of focus groups. The
study focuses particularly on the services
currently being provided to residents in
addition to the services the residents
think would be beneficial to them. The
concepts of concentrated poverty and
social capital are examined by inquiring
about individuals’ length of stay in the
housing units and their reasons for living
there. Finally, it explores the added
challenges refugees and immigrants face
in public housing.
Self-sufficiency is increasingly a
primary objective of local housing
authorities. Anthony (2005) was one of
the first to research the outcomes of a
Family Self Sufficiency (FSS) program.
He finds that any participation at all in
an FSS program will increase one’s
earnings over the next few years, but
completion of the program is associated
with a much greater wage increase and
increased likelihood of movement to
7
group consisted of refugees from
Somalia, a Haitian immigrant, and an
African American citizen. The sixth focus
group had 14 Russian refugee immigrants
who lived in Adrean terrace. See
Appendix
A
for
demographic
information on group participants.
Methodology
I chose to work directly with the
Utica Municipal Housing Authority
(UMHA) for this project because it
would allow me direct access to the
residents of the subsidized housing units.
It also gave me access to knowledge
about all services currently being offered
to residents and the extent of use of each
service in the community. I was also
allowed to work within the structure of
the housing authority by collecting
information in the three largest housing
facilities and tapping into the networks
that already existed between UMHA
employees and community members.
Participants were recruited for
the focus groups in two ways. The case
managers for Gilmore Village, Adrean
Terrace, and Perretta Twin Towers were
informed of the focus groups and asked
to invite people to attend. Beyond this,
the maintenance workers at Adrean
Terrace and Perretta Twin Towers
distributed fliers to each household.
Jason Beck, AmeriCorps Vista worker,
and I handed out the fliers door-to- door
on two occasions in Gilmore Village and
Adrean Terrace to invite residents to
participate. Our goal was to attract as
diverse a population as possible for the
focus groups in the limited amount of
time we had to conduct them. In Adrean
Terrace, we had an over-representation
of the refugee populations and an under
representation
of
native-born
populations.
The method of data collection for
this project was six focus groups that
took place on Thursday October 15,
Thursday October 22, and Tuesday,
October 22. The intention was to have
four focus groups, but attendance was so
great at the two sessions on October 22
that I had to split the group into smaller
groups. Participants in each focus group
lived in public housing. Although the
focus group participants were not a
random sample of the population living
in the units, the attendance was focused
on subgroups of the population who lives
in public housing. The first focus group
contained ten young adults between the
ages of 18 and 25. The second focus group
was made up of four residents of Gilmore
Village who were participants in the
family self sufficiency program. The third
and fourth focus groups were each made
up of 8 participants who were residents
of Perretta Towers, the senior and
disabled living space. The fifth focus
Each focus group met once, for
approximately 45 minutes. I facilitated
four focus groups and Jason Back
facilitated two. We followed a discussion
guide that had been developed through
collaboration
with
Dr.
Herbowy,
executive director, and John Furman,
grant-writer, at the Municipal Housing
Authority
(Appendix
B).
We
conceptualized “self-sufficient” as living
outside public housing, so I asked
participants why they currently live in
public housing, how long they plan to
stay, and what would move them to leave
8
public housing. In order to better
understand their needs and how they
were being met, I asked participants
about the services they receive and what
they would like to receive.
it was difficult to hear participants and at
times they did not seem engaged, as they
frequently made side conversations with
their neighbors. However, many people
had come to the room with the purpose
of participating in the focus group and
were interested in sharing their
experiences.
The first focus group was made up
of ten young adults, aged 18-25 who were
participants in the Job Crops program,
designed as a job training program
funded through the federal stimulus. All
of these participants live in the HOPE VI
target area, and some of them live in
public housing. The focus group took
place during their last day with the
program, so all were invited to
participate; only one person opted not to
participate. For the first focus group,
participants were thrown off by the first
question, “How long have you lived in
subsidized housing?” because few of
them live in subsidized housing, and
even fewer of them call it by that term.
The participants did not understand the
purpose of the research as well as other
groups did, and I did not understand
exactly how these people fit into my
project at first, but they all reside in the
HOPE VI target area.
At the last focus group session at
Adrean Terrace, approximately fifty
people came. The majority of them were
Russian refugees who did not speak
English, and there was a substantial
minority of Somali-Bantu refugees who
did not speak English. I randomly split
the Russians into two groups. Mr.
Furman facilitated one group, and Jason
Beck facilitated the other. There was a
translator from the Russian population
for each group. As my tape recorders ran
out of power, neither of these two groups
was recorded. As Mr. Furman did not
take notes during or after his focus
group, I was not able to use the
information he gathered. Jason recorded
the responses to the questions as they
were provided, which was made easier
due to the delay of translation. The
Somali-Bantu focus group had more fluid
attendance; the majority of participants
arrived on site approximately twenty
minutes late, and one woman, from
Haiti, arrived twenty minutes into the
session. Despite some side conversations
during the focus group, participants
seemed eager to answer my questions.
This group was translated by Jama, a 17year-old son of one of the participants.
The second focus group consisted
of four people, all of whom are
participants in the Family Self Sufficiency
Program (FSS). This one fostered the
most in depth discussion. Information
was shared among group members as
they explored the nuances of each
question.
The focus groups with the senior
citizens took place in an open
“community room.” Two focus groups
were held at the same time, and there
were other people in the room. At times
As a final element of the research,
and as an attempt to gain a more holistic
understanding of the issues related to
public housing, I interviewed Mallory
9
I would say, you know bettering
yourself as a person in all aspects um,
and, having enough income in your
household to support your whole
family outside of public housing.
Oakes, the family self-sufficiency
coordinator for the UMHA. As this
research aims to examine why people are
not yet self sufficient, Ms. Oakes was the
best person to interview as she works
directly with residents hoping to become
self-sufficient. During an interview that
lasted approximately 40 minutes, I asked
Mallory about some of the questions that
had arisen during the focus groups.
Oakes has worked for the housing
authority for nearly 18 years, so she was
able to provide in-depth answers to my
questions.
The purpose of her position is to “help
motivate the residents and if they’re
already employed maybe look for more
gainful employment, if they’re not
employed, help them get employed. Um
also help them as far as training,
education-wise or on-the-job training.”
With the help of the theories of
concentrated poverty, social isolation
and social capital, I provide residents’
explanations for living and staying in
public housing, and then provide the
motivations for leaving that a few
participants articulated. Next, I explore
thought processes of participants with
regard to obtaining jobs and having
concrete goals. Finally, I discuss the
supports and services identified by the
refugee/immigrant and the senior
groups, which had the greatest needs.
After completing each focus group
session and the interview, I transcribed
the data. The results were then coded an
analyzed for this paper. The primary
codes I used were loving public
housing/not wanting to leave, wanting to
leave, convenience, paying bills, security,
explaining rationale (for living in public
housing),
opportunities/needs
for
children/youth, being/expecting to be
coddled, attaining self-sufficiency, not
being informed, and identifying needs of
others. In my final analysis I selected the
codes that best fit into my research
question
and
had
the
richest
information.
Reasons for Residence
The vast majority of participants
were content to live in public housing.
Many had complaints about heating or
maintenance issues, but they would
rather live here than in the private
market. The waiting list for entering
public housing is long; an individual may
wait up to five years before moving in,
according to Housing Authority Officials.
In the six focus groups conducted,
consisting of approximately 60 people,
only five people stated that they wanted
to leave; three young adults, one senior
citizen, and one Somali Bantu.
Overwhelmingly, participants stated that
they enjoyed living in public housing and
Results
Based on my findings it is not
possible to fully explain the reason the
residents living in public housing are not
self sufficient. However, the focus group
discussions provide helpful insight into
the threats and opportunities associated
with gaining self-sufficiency from public
housing. Here Mallory Oakes, FSS
coordinator, defines self-sufficiency:
10
Tashmica: Well yea, you had the
center in the back, then you had the
park, then you had my church and I
didn’t have to go nowhere. Now you
gotta travel here and here and there
just you know to go to church, to find
something to do, you know […]
that they had no intention of leaving.
This may be related to the lack of
prescriptive norms Coleman (1988)
describes as a lack of social capital.
Although it might be “better” for an
individual who has enough income to
move out of public housing to provide a
space for another needy person, the
individual will remain in public housing
out of self-interest.
A: If you could go back to subsidized
housing, you would?
Tashmica: I would, I loved to live in
Washington Courts (group 1)
When comparing one’s current
experience in public housing with a
previous experience in the private
market, an individual’s rationale for
staying in public housing becomes
clearer:
This is a common thread throughout the
responses to this question. Over and
over, participants expressed that they did
not want to leave, that “it’s better here.”
Even Mallory stated that it made more
sense to live in public housing than in
the private market for a number of
reasons. It is easier, more convenient,
much more affordable, and more secure.
This may be indicative of the lack of
other viable opportunities for quality
housing so the reliability of thee social
environment that the UMHA offers to its
residents is an element of social capital
that they were not willing to give up.
Leantine: Where I come from I say
this is luxury, seriously. Where I grew
up I never grew up with stairs in the
house, I grew up with elevators, and
stairs was like twenty floors like one
staircase so this is luxury. It’s
beautiful it’s quiet you know,
barbeque with your family and stuff.
You can do your own thing, it’s
beautiful, it’s beautiful. […] I’m not
moving. (Group 1)
One of the primary reasons people
implied for remaining in public housing
was fear. A fear of job loss, a fear that
they could not make ends meet, or a
specific fear of paying utilities,
surrounded the conversations about
living in public housing.
One woman who was displaced with the
reconstruction associated with HOPE VI
shared her experience with the group:
Tashmica: When I was living there, I
was forced to move, I wouldn’t a
moved.
A: Why did they force you?
Doris: We thought about that
[buying a house], me and my
husband, but then I was like, you
know you gotta pay your bills,
your taxes and if something go
wrong it gonna come out of your
pocket. (Group 2)
Tashmica: cause they tore down—
they tore it down!
A: That’s right, they tore it down. I’m
sorry—so is it a real burden for you
to have been displaced?
11
Doris was paying the flat rate at UMHA:
$600 per month for a one-bedroom
apartment, the maximum payment. She
clearly made enough money to rent on
the private market or even to purchase a
house, but she was afraid to take the risk
because of the added responsibilities that
are attached to living outside the public
market.
incomes. This is consistent with other
findings that since the 1970s, public
housing has moved from being a
“temporary way station” for families to a
permanent stop (Bratt 1989, Anthony
2005).
The fear attached to moving to
the private market may be the result of
social isolation. If a resident is not
regularly exposed to people with steady
jobs who own their own homes, it may
be difficult for them to think it would be
possible for them to own a home. A lack
of relationships with people who can
provide information about having made
the transition from public to private
housing may be related to residents’
reluctance to take what they may
perceive as a huge risk in moving out.
Sheer expense was also a major
rationale for living in public housing.
Residents of Adrean Terrace related their
experiences on the private market,
explaining that they paid more than
twice as much. One Somali-Bantu
Participant related his story:
Jama: He say before here he used to
pay rent some other place. He used
to pay bills like a lot more than now.
He used to pay different payment, for
heat, electricity, water. Now, so he
came here, and me, myself I came
here. He say the electricity and water
and the house payment is all
connected, it’s together. He say he
don’t have any arguments about it,
he’s fine, he’s happy, he likes it all.
He say he used to pay $400
something before when he used to
live somewhere before. And $400 for
the house and the other payments so
that is $400 for electricity and $400
for house payment and that is $800.
(Group 5).
Staying Informed and Getting Ahead
In the subsidized housing units in
Utica, there are various services available
to residents. However, some residents
are not as well informed as other
residents due to their lack of social
relationships, or a deficiency in social
capital. Social Capital theory, as
described by Coleman (1988) emphasizes
the importance of social relationships
that provide information that facilitates
action. For example, Diane has been
living in public housing for a little over a
year. She knows she would like to leave
public housing within a year or two and
she sees that education will help her
achieve her goal. In November she began
a BOCES program that will result in her
certification in nursing. At various times
throughout the focus group, it was clear
that she was not as well informed as the
other three residents, who had lived their
Other participants agreed that, despite
working, they could not afford to pay for
housing, utilities, and food. For the
Somali-Bantu, public housing is not the
5-year “stepping stone” Mallory Oakes
states is the purpose of public housing,
but rather a permanent solution to a
problem of not having the ability to
support their large families on limited
12
ten or fifteen years, and had other family
members living in the complex as well.
to be the first time she is on this type of
public assistance. She does not seem to
know how things “work” as well as those
who are more established in the
community and who have spent
extended periods of time on public
assistance. In discussing homeownership,
Diane was the only participant who was
interested in purchasing one in the
future. Doris, however, knew about a
support that the UMHA provides to its
residents:
Diane: I was tryin’o find out.
Dominick told me they had this
program where they help you with
school…”
Cascilla: That’s the one she’s in …
Diane: But who do you have to
contact for that?
…Ie’sha: Just tell Mallory and she give
you papers and you take it to
whoever and they look at the books
and you fill out the paperwork and
you get a check every month
Doris: I think she explained it to me
like if they raise your rent, whatever
money was raised goes into an
ESCROW account and that is saved
for like one year or two, but then at
the end you gotta be specific about
what you wanna do with that money.
But it goes into a ESCROW account.
This is a case where Diane had heard
about a program but did not know how it
worked or how to get more information
about it. This program would provide her
with a government subsidy for going
back to school. Furthermore, it would
provide
financial
counseling
and
homeownership classes, elements of the
program she did not know about, but
was interested in when she heard about
them during the focus group. Less than a
minute later, Cascilla mentioned a
reading program the UMHA had for kids
during the summer; Diane’s response
was, “Oh, they did?” Although it is likely
that she received some passive
information about the program, such as
fliers in the mail, she clearly did not take
advantage of it. Even though Cascilla
does not have young children, she still
knew about the program because she is a
well-established community member,
having lived in public housing for more
than a decade.
… Diane: Oh, really?
Often, what the residents save in an
ESCROW account is spent on cars or
homes, the participants in the focus
group told me. According to Mallory,
“They can actually take it out and use it
at any time during the 5 years for specific
reasons, to help them maybe achieve
their next goal.” As Diane is an aspiring
homeowner, it would be useful for her to
take advantage of this program. With
more information she might be able to
judge the benefits it may provide her
family. While discussing health care,
Diane complained of having to pay for
her daughter’s prescriptions out of her
own pocket. Immediately, the other
three women present asked her which
kind of Medicaid she had. They told her
to change to Fidelus because it’s better
and less expensive. All agreed that
Fidelus is the best service provider.
Diane has lived on her own since
she was sixteen years old, but this seems
13
Diane’s case is probably not unlike
other short-term users of public housing.
New to the system, and without strong
social relationships in her community,
she does have as much knowledge about
the programs available to her and her
family. These programs could be used to
support her and help her get on her feet
more quickly if she were informed
enough to take advantage of them.
challenges may arise that prevent them
from completing their education.
In my first focus group, all
participants but one had completed
either high school or their GED. One
student had finished only eleventh grade,
but about half of them had one semester
or more of “college.” However, for the
majority of the participants, who were
between the ages of 18 and 25, the goal
was to go to college. Many of the young
people wanted to get good jobs so they
could pay their way through college.
College was a general term used by all
participants; they did not say what
college they wanted to attend or that
field they wanted to pursue.
Lacking Concrete Goals
The Family Self Sufficiency
Program is concerned with helping
individuals make enough money to
support
themselves
financially
independent of public assistance of any
kind. Participation in such a program
requires a significant time commitment
for job-skills training, job searching, and
working (Anthony 2005). Another major
element of the program involves goal
setting, short term and long term. This
proves difficult for many residents of the
Housing Authority and the HOPE VI
target area.
Upon probing participants the
young adult focus group about what
their aspirations were, the majority of
responses were vacant stares. Only one
participant, Vladimir, stated that he
wanted to get a good government job
and move out on his own:
Vladimir: That’s what I’m lookin’ for…
A good job with benefits, a state job,
or a government job.
For many young people in the
United States, “getting in to college” is a
concrete aspiration that they reach
toward for the first eighteen years of
their lives. Upon entering college, a
student may aspire to succeed in school,
but a larger goal is needed to follow
graduation. At an elite university, that
goal could be more education or an
internship. These goals aspirations
discussed frequently in the community
and are expectations for students when
they first enter school. For those living in
poverty, getting to college may seem the
ultimate goal. Upon arrival there, new
A: A state job. So how are you going
to get that job?
Vladimir: A civil service test.
A: And then you can get it? So
you’re ready, you’re on track?
Vladimir: A couple of references from
the boss, that’s it.
He also was the only one with a concrete
plan about how he would meet his goals
– take the civil service test. However,
even that concrete plan is unclear,
because Vladimir did not express what
14
type of work he wanted with the
government.
something she thought the housing
authority could do for her:
This speaks the fact that these
young people were raised in a different
social and cultural context than people
who attend elite universities, where the
vast majority of students intend to
pursue some sort of professional work
and have an idea of the field they would
like to enter. The participants in the jobtraining program provided by the UMHA
to young unemployed people were
unable to state a job that they would like
to do. They were unable to articulate
problems they were facing and could not
state in a concrete manner what their
aspirations in life were.
Ie’sha: Some type of training, any
type of training. Anything… I would
do construction too. I have three
kids, I would do anything it takes for
my babies. Construction, plumbing,
anything.
This is evidence that there is abstract
willingness to learn a new skill or to take
a class where she will be able to make
enough money to support her children.
However, she does not say what type of
work she would like to do. Although she
is willing to do “anything” she is not
currently gainfully employed.
Despite this lack of concrete
goals, a strength of the participants in
the first two focus groups is that they
enrolled in programs to help them
formulate and attain goals. These
participants are already steps ahead of
their peers who have not enrolled. The
participants in the Job Crops program
learned skills such as lawn mowing and
painting, and the participants in the FSS
program received subsidies to support
their return to education or job skills
training. At the very least, they were
interested in making money and were
motivated to try to find ways to make
more money, which did not seem to be
the case among the senior or refugee
populations.
These findings are evidence of
Ryan and Deci’s self-determination
theory and the facilitation of intrinsic
motivation, social development, and
well-being. They argue that “human
beings can be proactive and engaged or,
alternatively, passive and alienated,
largely as a function of the social
conditions in which they develop and
function” (Ryan and Deci 2000, 68). It
seems that the environment in which
these young adults were raised has
caused them to develop into more
passive members of society who are less
capable of determining for themselves
what sort of future they would like to
pursue, compared to those of a more
privileged background.
Immigrant Issues
Participants in the FSS program at
Gilmore Village (focus group 2) were
more able to state goals, but plans were
still not very concrete. Ie’sha enrolled
herself in a GED program and stated
Refugees enter the country with a
specific cultural background and a
history of hardship. Two major refugee
populations in Utica are Russian and
Somali Bantu. Since they are refugees
15
understand what you says to them
but it is difficult for them to put it in
writing. But it would be ok for you to
teach them how to write, if you teach
them how to read, so if you open like
uh English classes, it would be better
for them. (Group 5).
and not immigrants, the circumstances
of their pasts are typically traumatic,
which makes it more difficult for them to
readjust once they come to the United
States. Among the participants of the
three focus groups involving refugees
from Russia and Somalia, fewer than one
third could speak English at a level even
close to conversational. This obviously
limited their ability to communicate
their needs and desires to the group. Not
speaking English may carry with it a
stigma that alienates these individuals
from other parts of society. However,
both ethnic groups were engaged in the
conversations and were interested in
sharing their points of view. Among the
participants were four immigrants; three
of them lived in Perretta Twin Towers—
two from Puerto Rico and one from
Italy—one, from Haiti, lived in Adrean
Terrace.
Milda has since learned how speak, read,
and write in English and completed a
college degree. All of the Somali-Bantu
participants of this group as well as the
Russian participants of Group 6 were
very interested in ESL (English as a
Second Language) classes. However,
among the senior citizens, the situation
was different. Although the three
participants who were immigrants could
speak English, it was not clear whether
non-English
speakers
would
be
interested in classes. Carmen, from
Puerto Rico, stated:
Carmen: I donno if he wanna [learn
English] because sometimes he don’
wanna go to school.
English Language Learning classes
and citizenship classes were of great
interest to the refugees and immigrants.
Milda, a Haitian immigrant who is 31, has
been in the United States for fifteen years
and has lived in Adrean Terrace for 9
years. She was able to provide some
insight into the situation of a nonEnglish-speaking immigrant:
A: But what if somebody came here?
Carmen: Oh, I don’t know
The situation in Perretta towers may be
different because it is targeted toward
the senior citizen and disabled
population. There may be a greater sense
of community among separate ethnic
groups in the building, as described in
Irene’s comment:
Milda: So like some of them, not
everyone can get into this country.
Some of them they don’t know how
to write, some of them they don’t
know how to read. I taught myself all
these things. When I first came here,
I was the same thing. Some of them
they don’t know how to write. Some
of them, they’re very intelligent, but
the problem is they don’t know how
to write and they don’t know how to
read. Just, you can talk to them, they
Irene: It isolates them. They stay to
themselves right, because they do not
understand English very well. Like I
communicated with one of the ladies
over there really slowly but she really
don’t understand what I’m saying at
all times. Because I was uh, I knew, I
had known German, and those
16
people had known German at one
time, but you know I haven’t used my
German in 30 years, so.
translate one person’s comments, “She
say she wanna know then the classes are
gonna start.” Similarly, Milda explained
what she knew about the process and
expressed her interest in becoming a US
citizen.
A: So it’s really difficult to
communicate with the Bosnians. And
the people who speak Spanish…
Each of the ethnicities seems to
embody its own community within the
housing units of the UMHA. Although it
seems that native-born residents of
public housing are concerned about
them and would like to better
understand them, the refugees I spoke
with were unable to communicate.
Among
the
Somali-Bantu,
social
cohesion can be seen in the community
of families who depend on each other for
emotional support. Similarly, among the
Russian population, there seems to be a
sense of family among the many
participants. In both communities, there
were multiple cases in which the older
members of the communities looked to
the young people to translate for them
when I went door to door inviting
residents to attend the focus group; at
times the children were as young as five
or six.
Irene: None. But they speak Spanish
among themselves and that bugs a
lot of people a lot of times, because
they think they are talking about
people, and maybe they are, maybe
they’re not. So they primarily stay to
themselves and the English people
stay by themselves. 3 Groups… within
a community. (Group 3).
Here, Irene expresses her desire to be
able to communicate with her neighbors
in the community and her frustrations
with not being able to communicate.
Although Carmen stated that she does
not think her husband would like to
learn English, Irene seems to think it is
important in the creation of a cohesive
community. Similarly, Randy, the African
American participant who was part of the
Somali Bantu group, said, “English
classes, you know, we need someone to
come in and teach English classes…”
(Group 5). He peered around the table
with a look of concern and compassion,
as if he truly wanted to be able to
communicate with these people and
make their lives better.
Perhaps as a result of refugees’
cultural experiences and lack of English
knowledge it is more difficult for them to
fathom “self sufficiency.” As they are new
to the country, they are less likely to have
informal networks they can tap to try to
find work or to ask for advice about
possible decisions they make. Networks
of non-English-speaking refugees are
very likely limited to people of their own
ethnicities. This seems to limit social
interactions and may perpetuate the
problem of social isolation in the
community in the form of poverty, lack
Beyond the debate about learning
English, refugees and immigrants also
expressed interest in becoming US
citizens. Both the Russian participants
and the Somali Bantu participants were
eager to begin the process of obtaining
citizenship. When the classes were
mentioned in group 5, everyone began to
speak at once. Jama was only able to
17
of jobs, and lack of English skills in the
community.
winter, had a security guard to keep out
non-residents or unwanted people, and
there is a string to pull in each apartment
if a resident has a medical emergency.
Here again, the reliability of the
environment seemed a very positive
element in participants’ experiences with
public housing, but one resident shared
her frustrations with medical care:
Senior Issues
The senior living spaces offered by
the housing authority bring together
many low-income senior citizens and
people with disabilities. Between Perretta
Towers and Chancellor, there are over
two hundred senior citizens utilizing
UMHA housing
units,
which
I
hypothesized may cause concentration
effects. As Galster, Cutsinger, and
Malenga
(2008)
theorize,
the
consequences of concentration effects
are poorer health, lower levels of
educational
achievement,
fewer
employment opportunities, and social
isolation. Obviously the case is
somewhat different for senior citizens,
who live on a fixed income because few,
if any are still employed. Seniors tend to
have more severe and frequent health
complaints because of their age, but the
participants in the focus group seemed to
have few severe health issues; all were
well taken care of in terms of health care
and were satisfied with their level of
education. Only one participant said she
wanted to go back to school, and most
said they had completed at least their
GED. Social isolation did not seem to be
a problem either because various
participants compared their experiences
with other seniors or stated they had
connections outside public housing, such
as their families.
Sharon: Alright, I’m with Secure
Horizons, now I can’t get my
medication… because it’s called co…
something gap. They took away my,
uh, the “d” on the social security…
Ken: Part D is medicine isn’t it?
Sharon: Part D on the Medicaid and
stuff. They took that away. I can’t get
prescription drugs, co pays are
ridiculous, I can’t even pay for them!
I’m in a stoop. And pissed. (Group 3)
Here, as had happened in Group 2, one
of the other women asked her why she
didn’t change her policy to Fidelus.
Sharon responded, “Because I have no
idea of anything!” This is a case where
social capital in the form of social
relationships is lacking, which has
consequences for her emotions and
health. Each of the other participants in
the focus group had enough medical
coverage to pay for their doctor
appointments and medicine without
trouble.
The seniors in both focus groups
took advantage of the services offered to
them through the housing authority, to a
much greater extent than in the other
housing units. They participated in
BINGO, and used the computer.
Participants in both groups said they
were interested in learning how to use
Senior citizens reported general
satisfaction with their experiences in
public housing. They felt safe because
the housing authority kept the area
around the complex free of ice during the
18
the computer, and that they would like a
larger computer lab. There are two
computers for about 120 apartments.
Another common desire in both groups
was to have exercise machines or a
person to come in and teach exercises to
the group. Finally, both groups spend
considerable time discussing trips they
would like to take as a group.
opinion there were not enough buses
available to give him the mobility he
needed. It seemed like he did not have a
family network to lean on in Utica if he
needed something. Gina voiced the
importance of a family network in
getting transportation:
Gina: Once in a while I depend on my
sister. She’s an ____. I need
transportation in the wintertime,
because I’m a walker in the summer,
I can walk miles and miles it doesn’t
bother me. In the wintertime, I’m
terrified to put my feet outside. I
can’t walk to that sign (points). If I
have to go to the doctor, or the bank,
or to get my pills…We need someone
like that (Group 4).
Theresa: Well I think trips would be
nice. The other seniors, you know
how their programs are […] it would
be nice if they brought us to different
places, like to northern Utica, like to
different places (Group 3)
It seems like Theresa, and the other
participants who nodded their heads in
agreement after her statement, wanted
to be treated as senior citizens were
treated in private senior living facilities.
Nearly all participants had at least one
idea about a trip they would like to take.
Thus, family networks seem to be an
important issue related to whether a
senior resident feels satisfied with the
amount of transportation offered by the
housing authority.
Transportation seemed to be an
important issue:
Unlike other residents of public
housing, fear of job loss is not a factor
keeping them in public housing. Rather,
fear for the safety of their bodies is more
prominent. These individuals know their
weaknesses and depend on the housing
authority to provide them with the
support they need. Due to the reliability
of the facility and its employees, it seems
that the majority of residents are content
to live there and do not hope to move
out.
Irene: From these housing complexes
you have grocery buses going so if
you can’t get out or walk to a grocery
store you can go once a week to
Hanoford, once every two weeks to
um Chanantry’s and once a month to
Walmart.
Now
it
is
free
transportation the housing has
gotten for us. (Group 3)
Irene was satisfied with the services she
received here, but she also has a
daughter who works across the street
who could take her shopping or on other
errands if needed. On the other hand,
Jose (Group 4), who had lived in New
York City before coming to Utica, found
transportation a large issue. In his
Discussion and Conclusion
In this research, I sought to learn
more about how residents in public
housing construct their experiences
there and why they are not yet “selfsufficient.” The goal of the UMHA at the
19
outset of the study was to find out why
residents were not self-sufficient and
what services could be provided to help
them become self-sufficient, in order
that Mr. Furman can use the data as
rationale for applying for grants to meet
these needs. The findings of this study
can be used in multiple ways. It will be
useful to the housing authority itself.
Although the managers of the housing
authority often have “community
meetings” with residents, discussions
primarily revolve around maintenance
issues. This research further provides
important insight into the reasons people
live in public housing. Far from feeling
stigmatized for living in public housing,
the vast majority of participants felt
extremely satisfied, some even privileged,
to live there. Lack of income is only part
of the reason people live in public
housing. The more important element
perhaps is the security attached to public
housing residence: if you lose your job
you will not get tossed out on the street.
A $50 minimum rent is unmatched in the
private market. The dependability of the
social environment that Coleman
describes as an element of social capital
seems to be of particular importance in
this community.
intends to remain in public housing for
the rest of their lives because they cannot
afford the private housing market.
Concentration effects do not seem to be
as present among seniors, who struggle
to find meaning and purpose in a fixed
income experience. There is “no way out”
for seniors.
Concentration effects are evident
in the results regarding a lack of
motivation. It seemed that most
participants were not motivated to move
away from public housing. As Kane
(1987) suggests, “the perception of
helplessness can be inferred from the
environment, without direct experience
of failure.” Without positive role models
and shared experiences of friends and
relatives who live on the private market,
residents may feel helpless. Carmen,
from the senior housing site, said, “the
people that move out of here, they come
back” (Group 3). Residents felt they
couldn’t afford to move out of public
housing, even if they had not tried it.
Others remarked that public housing was
“luxury” compared to any other type of
housing they had previously inhabited.
The residents’ failure to effectively
support themselves and their families in
the private housing market may have
destroyed any motivation an individual
may have had to live outside public
housing. This larger issue poses a threat
to the success of self-sufficiency
programs in the UMHA.
The senior citizens living in public
housing who participated in this study
wanted to be treated with respect. The
requests they had with regard to
additional services would improve their
quality of life and may take little effort
on the part of the housing authority. The
problem with this is, as always, is lack of
funds. A difference between Perretta
Towers and the rest of the UMHA
housing units is that self-sufficiency is
not a goal for any residents. All but one
Self-sufficiency, as defined by the
housing authority (note Mallory’s
quotation at the beginning of the results
section), is not on the agenda of many of
the participants in the focus groups.
When I began this study, Mr. Furman
20
wanted to know why residents were not
self- sufficient and what the housing
authority could provide to them to make
them self-sufficient. Consistent with
Anthony’s (2005) findings, young adults
(aged 25-40) tended to have a greater
interest in attaining self-sufficiency than
older adults (aged over 40). Refugees in
public housing did not seem to
understand the concept of selfsufficiency and stated that they would
like to remain in public housing for a
long period of time, perhaps for the rest
of their lives. The disconnect between
the goals of the administrators and those
being administered to is problematic. As
Mallory stated, “You can bring a camel to
the water, but you can’t make it drink.”
The opportunity exists for all non-senior
residents to participate in the Family Self
Sufficiency program, but only a small
minority participates. Despite the efforts
of the housing authority, if residents are
not self-motivated enough to participate
they will probably not achieve the goal of
self-sufficiency.
There is much need for further
research, as this project faced some
limitations. One limitation of this
research is that certain populations
within the housing authority were underrepresented: US citizens who do not take
part in self-sufficiency programs. My
identity as a young, white, female
student of an elite university may have
influenced how people responded to my
questions. They might have omitted
experiences that they thought I could not
relate to. For focus groups five and six, it
may have been better to conduct the
entire discussion in the native language
of the participants. In-depth interviews
with residents of public housing who
participate in job training programs such
as the Ross program or FSS could provide
more insight about their motivations for
participation in the programs. Perhaps
this type of information could be helpful
to aid in the motivation of other
residents to participate. Although
quantitative data exists that compares
participants in FSS programs with people
who do not participate, their data would
be strengthened by qualitative data. Indepth interviews could demonstrate
qualitative differences between the
participants and non-participants, as well
as between those who complete the
programs and those who do not.
Additionally, a longitudinal study could
evaluate the effectiveness of FSS
programs.
The findings here build on the
work of Owens-Manley, Coleman, and
Curley because it provides qualitative
data to explain why some people choose
to remain in public housing. The
participants’ inability to articulate
aspirations or goals suggested that social
isolation was a problem in the
communities. As Curley argues, residents
of public housing seem to be isolated
from positive role models and to thus
continue in the cycle of poverty.
Although FSS is available to all residents
of public housing, Mallory stated that
only 77 families take part, out of over 600
housing units (this excludes the senior
residences).
Another space for more research
is in the long-term outcomes of HOPE
VI. It would be interesting to find out
five or ten years after a resident has been
relocated, as Tashmica was, what their
educational and occupational outcomes
are. Poverty still exists in Utica, and
21
although the UMHA seems to be
working hard to improve the situation, it
is essential to keep investigating the
successes and failures of programs
alongside the needs, desires, and
aspirations of residents. Finally, it would
be interesting to investigate how the
people on the UMHA waiting list survive
in the time between they qualify to be
placed on the wait list, and their
placement in a UMHA apartment.
Perhaps if residents of public housing
knew more about the people waiting,
they would be more likely to move out in
order to make space for them.
should be researched and that
information used to develop a culturally
relevant “manual” that will help with
readjustment
and
provide
recommendations and resources for
gaining self-sufficiency. Third, invest in a
community organizer. Fourth, create a
strong community council that comes
together around issues other than
maintenance, and finally, introduce
computer labs and classes accessible for
all residents.
The purpose of this paper was to
understand why residents of the Utica
Municipal Housing Authority were not
yet self-sufficient. Self-sufficiency, as
defined by the head of the FSS program
is bettering yourself as a person in all
aspects and having enough income in
your household to support your whole
family outside of public housing. That
kind of self-sufficiency was not an
explicit goal of any of the participants in
the focus groups. Diana was interested in
moving out of public housing because
she wanted to buy a house, but she was
one of the few who expressed this
interest. Therefore, the framing of public
housing must change. The definition of
self-sufficiency and the establishment of
self-sufficiency as a goal for every
resident are essential if that is the goal of
the employees of the UMHA. If it is not
possible for all residents to become self
sufficient (according to the definition
above), perhaps a new definition is
needed. Most importantly, there must be
alignment between the residents and
employees to achieve the best results.
The de-concentration of poverty is
going to take more than just rebuilding a
few buildings in one community. It will
take a serious multifaceted investment in
a large community of people. It will
probably be a process that takes a long
period of time. Despite the offerings of
the FSS and Ross programs, a relatively
small number of people take advantage
of them. Something has to change for
these residents to take the step to change
their minds about living in public
housing.
Recommendations
There are infinite possibilities for
addressing the issues presented in this
paper. The public housing authority has
become an important source of social
capital for residents. Therefore, in order
to better serve the residents’ needs and
improve their self-sufficiency, I will make
five
recommendations.
First,
the
definition of self-sufficiency must be
discussed and aligned from the time a
resident enters public housing. Second,
the cultural background of residents
In this effort, it is absolutely
essential to think critically about “selfsufficiency potential.” Which residents
22
have the potential to be self sufficient,
and which do not? Senior citizens who
live on a fixed income, for example, have
little potential compared to a household
with two able-bodied adults and few
children. Therefore, I propose that every
new resident has a consultation in which
a caseworker works with him or her to
establish an action plan to improve his or
her self-sufficiency. This would result in
an agreement between the resident and
the UMHA regarding how to improve
self-sufficiency. The income they save
from paying a lower rent could go to
their children’s education, savings for a
car or house. At this time, the
caseworker would also inform the new
resident about the Family Self Sufficiency
Program and the Ross program, as well
as other opportunities the UMHA has to
offer. If the resident is a senior citizen,
the UMHA should also have a
consultation with that individual to
ensure that they have medical care, are
aware of existing programs in the
UMHA, and perhaps provide ideas for
nearby volunteerism or social justice
projects.
Increased
community
involvement could promote a better
standard of living among senior citizens.
This consultation could also be useful in
establishing a plan for English language
learning for residents who came to the
United States recently, either as
immigrants or as refugees. The
caseworker should maintain profiles of
all residents: family size and ages, skills,
aspirations, education needs, and selfsufficiency action plan. Although this
will only help new residents, eventually
all residents will have had this service
when they move in. The purpose of these
efforts is to ensure that upon a resident’s
placement in public housing, he or she
clearly understands the purpose of public
housing and how their life will improve
as a result of living there.
The establishment of a culturally
sensitive manual for residents who came
to the United States as refugees can also
serve an important purpose in leading
them
toward
self-sufficiency.
A
community organizer from the refugee
population could be hired to compile the
manual. First, the HA must understand
the background of the refuges. Do they
want to adopt the American dream? If
the writer of the manual understands the
culture they are dealing with, they will be
able to align the text with the values of
its recipients. If there is a misalignment
between the federal desires for residents
of public housing and the abilities of the
residents, the issue must be addressed.
Once that research has been completed,
the manual would include important
contact people in different fields, steps to
follow for taking English classes,
completing a GRE or other educational
opportunities, where to turn for
psychological support, and other things.
This manual could then be personally
distributed and explained in the
residents’
native
language.
The
community organizer would stay on as a
staff member and would serve as a
caseworker for members of their ethnic
group. More than one person could
occupy this position, depending on the
number of ethnic groups in the
community.
A community organizer could aid
in the strengthening and deepening of
inter-residential ties has the potential to
aid in self-sufficiency. The UMHA could
23
document stories of individuals who
have moved out of public housing and
found success, in the form of a monthly
newsletter or on a bulletin board located
where people often go. This could spark
other residents’ desire to move out.
Another tactic involves taking time to
recognize skills within the community
and encouraging residents to use existing
skills creatively to get income can
increase their self-confidence and
improve community relations. The
community organizer could conduct a
capacity survey with residents to see
what experiences and local knowledge
they possess. Additionally, associations
within the community could be
established to capitalize on these skills,
or provide childcare to working mothers,
or provide support for newcomers to the
community. A more casual, open forum
could occur on a monthly basis where
individuals can discuss what they are
struggling with. The recognition of
existing
knowledge
within
the
community and provision of space to
share that information could alleviate the
financial or emotional struggles an
individual is going through.
Finally, access to computers is
essential to low-income individuals of
the current epoch. From children’s
homework assignments to adult’s job
application process, a computer and a
printer is extremely important in the
self-sufficiency of residents of public
housing. Students reported receiving
lower grades because their homework
assignments were handwritten rather
than typed, and adults stated that they
needed computer access in order to get a
job. Computer classes could also build
the skill sets of individuals. Once an
individual has mastered word processor,
excel, and PowerPoint, he or she will
have
access
to
jobs
such
as
administrative assistant, data entry, or
secretary. Furthermore, internet games
are useful in teaching children as well as
adults new vocabulary words, a helpful
support for English Language Learners.
There is much to be done, but
there are a multitude of exciting options
for the improvement of public housing
for residents. This improvement could
lead to greater self-sufficiency among
residents, which would then allow the
UMHA to serve a larger number of
people. However, these efforts alone may
prove to be lacking. It is still better to
live in public housing than in the private
market. To address this issue, I would
recommend lifting the cap on what
residents pay for their apartments. That
way, if a resident is making so much
money that they do not want to pay 1/3
of it to UMHA for their apartment, they
can chose to move out. Then, they could
possibly find a private residence that is
less expensive. A more gradual move
from public to private housing could also
be helpful: after a certain point, residents
A strong community council
would be the final step in strengthening
and deepening inter-residential ties. It
could also improve communication
between the UMHA and residents. If the
community council had some power to
make decisions about their community,
residents might take more ownership
over the place they live. They would feel
more agency within the community and
have the potential to become more
independent.
24
could have to pay for utilities or for
maintenance so that they would have the
skills to keep a budget. The issue is
many-sided but every side offers an
opportunity.
Many thanks to Jason Beck for his
substantial contributions to this project,
to the Utica Municipal Housing Authority
for partnering with me, Professor Benson
for her advising, and my friends at the ESS
2010 conference for helpful feedback.
Works Cited
2003. “When Hope Falls Short: HOPE VI,
Accountability and the Privatization of
Public Housing.” Harvard Law Review.
Vol 116 no 5. pp 1477-1499.
Curley, Alexandra M. 2005. “Theories of
Urban Poverty and Implications for
Public Housing Policy.” Journal of
Sociology and Social Welfare.
2007. “HUD’s Public Housing Program.”
U.S. Department of Housing and Urban
Development.
<http://www.hud.gov/renting/phprog.cf
m>
Ellen, Ingrid Gould. 2008. “Spillovers and
Subsidized Housing.” Pp 1440158 in
Revisiting Rental Housing: Policies,
Programs and Priorities. Ed Nicolas
Retsinas and Eric S. Delsky. Washington,
DC: Brookings Institution Press.
Anthony, Jerry. 2005. “Family Self
Sufficiency Programs: An Evaluation of
Program Benefits and Factors Affecting
participants’ Success.” Urban Affairs
Review. Vol 41 no 1. Pp 64- 91. Sage
Publications.
Gilderbloom, John Ingram with Michael
Brazley and Michael Anthony Campbell.
2008. “Hope VI: A Dream or Nightmare?”
pp 115-143 in Invisible City. Austin, Texas:
University of Texas Press.
Halpern, David. 2005. Social Capital.
Polity Press, Malden, MA.
Bratt, Rachel. 1989. “Rebuilding a lowincome housing policy.” Philadelphia:
Temple University Press.
Hawkins, Robert Leibson. 2005. “From
Self-Sufficiency to Personal and Family
Sustainability: A New Paradigm for Social
Policy.” Journal of Sociology and Social
Welfare. Vol 21, no 4. pp 77-91.
Coleman, James S. 1988. “Social Capital
in the Creation of Human Capital.”
American Journal of Sociology, University
of Chicago.
Kane, Thomas J. 1987. “Giving Back
Control:
Long-Term
Poverty
and
Motivation.” Social Service Review. Pp
405-418
Crane, Jonathan. 1991 “The Epidemic
Theory of Ghettos and Neighborhood
Effects on Dropping Out and Teenage
Childbearing.” The American Journal of
Sociology. Vol 96, no. 5. Pp 1226-1250.
University
of
Chicago
Press.
<
http://www.jstor.org/stable/2781341?seq=
4>
Katz Reid, Carolina, Edward Liebow, and
Gabrielle O’Malley. 2006 “Building
Community
during
HOPE
VI
Redevelopment: Lessons from a Seattle
25
the Facilitation of Intrinsic Motivation,
Social Development, and Well-Being.”
University of Rochester. American
Psychological Association, vol 55, no. 1 pp
68-78.
Case Study.” Human Organization. Vol
65, no 2.
Kingsley G. Thomas, Jennifer Johnson,
and Kathryn L.S. Pettit. 2003. “Patterns of
Section 8 Relocation in the HOPE VI
Program.” Journal of Urban Affairs. vol
25, no 4. pp 42-447.
Rosenthal, Gilbert A. 2004. “Hope after
HOPE VI.” Journal of Housing and
Community Development. Vol. 61 no. 3.
Pp 19-26.
Massey, D.S., and Denton, N.R. 1993.
American Apartheid: Segregation and the
Making of the Underclass. Cambridge:
Harvard University Press.
South, Scott J. and Kyle D. Crowder.
1999. “Neighborhood Effects on Family
Formation: Concentrated Poverty and
Beyond.” American Sociological Review,
Vol. 64, no 1 pp 117-132.
Marquis, Gerald P. and Soumen Ghosh.
2008. “Housing Opportunities for People
Everywhere (HJOPE VI): Who gets back
in?” Social Science Journal. Vol 45 no 3.
pp 401-418.
Sugrue, Thomas J. 1996. The Origins of
the Urban Crisis: Race and Inequality in
Postwar Detroit. Princeton University
Press: Princeton, NJ.
Mouw, Ted. 2006. “Estimating the Causal
Effect of Social Capital: A review of
Recent Research.” Annual Review of
Sociology. Pp 79-102.
Wilson,
W.J.
1987.
The
Truly
Disadvantaged. Chicago: The University
of Chicago Press.
Newman, Sandra J. and Joseph M.
Harkness. 2002. “The Lonog Term effects
of Public Housing on Self-Sufficiency.”
Journal
of
Policy
Analysis
and
Management. Vol 21 no 1 pp 21-33
Owens-Manley, Judith ed. et al. 2007.
“The HOPE VI Project- Utica, NY. Year 4
Program Evaluation.” The Arthur Levitt
Public Affairs Center, Hamilton College,
Clinton, New York.
Popkin, Susan J. Diane K. Levy, and Larry
Buron.
(2009).
“Has
HOPE
VI
Transformed Residents’ Lives? New
Evidence from the HOPE VI Panel
Study.” Housing Studies. Vol 24, no 4. pp
477-502.
Ryan, Richard M. and Edward L. Deci.
2000. “Self Determination Theory and
26
Implicit Discrimination in Major League Baseball:
Marginality Decisions based on Minor League Offensive
Production
Mark Kanter, Class of 2010
While Major League Baseball employs a wide variety of races and ethnicities, including
white, African-American, Hispanic, and Asian players, it is possible that professional
rosters may not be organized based on talent alone. Although studies have shown that
salary discrimination is no longer present in Major League Baseball, this paper explores the
possibility of a marginality bias in the sport, allowing for whites of comparable ability to
non-whites to receive a roster spot due to an implicit bias in managerial and organizational
positions. This possibility of marginality is explored within the MLB for the 2009 season by
regressing a variable indicating whether a player was called up to a major league team
against a player’s race and variety of control variables. Results from OLS regressions
indicate that on average, neither blacks nor Hispanics must exhibit production numbers
superior to whites in order to be “called up” to the major league level. However, it does
appear that non-white players perform, on average, to a greater offensive level than their
white competitors within the minor league system. While a player’s race may not
contribute to management’s decision to promote them, this study does find a variety of
variables that do significantly affect such a decision.
Introduction
Jackie Robinson’s debut for the
Brooklyn Dodgers in 1947 opened the
doors for the addition of non-white
players into Major League Baseball.
Once this invisible barrier faded, Negro
baseball leagues began to function more
as a minor league system for AfricanAmerican players desiring to sign
professional contracts, before finally
disappearing completely by 1960 (Mills,
2003). While Hispanics made up only 3
percent of MLB rosters in 1950, this
representation increased to around 20
percent by 1990 (Gonzalez, 1996). Today,
27 percent of MLB ballplayers are
Hispanic, with blacks representing about
10 percent of the league (Lapchick, 2009).
Although whites have made up between
58-60 percent of rosters since the late
1990s, there still exists a worry about two
separate phenomena—“centrality” and
“marginality”. In a 1970 study done by
Loy and McElvogue (which focused on
professional football), they proposed that
whites are more likely to be employed in
central positions, such as quarterback,
center, and kicker; minorities, they
argued, could be found in positions
relying not on leadership, but athleticism
and instinct, such as running back or
receiver.
This same idea of centralization
has been applied to baseball.
John
Phillips defined centrality in his 1980
study as “the tendency to exclude
qualified black players from positions
which involve interactions with and
control of the action of teammates.” In
1999, Margolis and Piliavin introduced
27
this issue to baseball, claiming a
disproportionate amount of whites
centered in the roles of pitcher and
catcher, with blacks usually in one of the
three outfield positions. They created a
set of control variables (such as size,
power, fielding range, speed, and age) to
include alongside race, and ran a
regression to determine the relationship
between race and position centrality.
Each position was given a centrality
ranking – for example, they determined
catchers to have a 1 rating (most central),
corner outfielders a 5 rating, and the
least central, the designated hitter, a 6.
The regression determined that on
average, whites exhibit greater centrality
over blacks by 1.18 points in the 6-point
scale (this result was statistically
significant). However, this measure of
centrality was an arbitrary number
decided upon by the authors, and
although they reached a significant
conclusion, their results leave little to be
expanded upon.
to argue that running backs and
centerfielders shouldn’t be considered
vital to their respective sports. So, these
positions where measuring ability is
difficult (which they conclude are
positions in which batting is least
important) are likely to have a lesser
proportion of blacks (they don’t look at
Latinos) – essentially, they propose that
when managers have no specific criteria,
they instead consider race. They claim
that catcher, shortstop, and second base
are the three positions where batting is
least considered. To account for this,
this study will use offensive production
statistics as independent variables to
remove this aspect from the call-up
decision,
additionally
running
a
regression which considers data for
outfielders only.
Similar research by G. Gonzalez
(1996) supports the stacking hypothesis
for Latino baseball players. The paper
objects that position stacking and
centrality is solely a black and white race
issue, and uses statistics (not a
regression) to analyze Latino positioning
over time. They find Latino players have
been stacked in the positions of second
base and shortstop from the period of
1950-1992, but also conclude that this
goes against the centrality hypothesis, as
these
infield
positions
require
“interaction and control” central to the
game (Gonzalez, 1996). Yet this still
raises a problem – even if centrality is no
longer the greatest issue, if black players
are disproportionately stacked in the
outfield, with Latinos in the middle
infield, there is likely to be increased
competition with those of the same
ethnicity for a select amount of
positional openings. If managers are
My study also aims to expand on
specific research published in 1994 by
Lavoie and Leonard, which proposes a
separate explanation for discrimination
in baseball. Rather than believe that
players are moved to specific positions
based on their ability to lead (as
determined
by
subconscious
discrimination,
the
aforementioned
“centrality” hypothesis), they propose a
new “uncertainty hypothesis,” stating
that positional segregation occurs most
when it becomes difficult to objectively
measure performance (Lavoie and
Leonard, 1994). Their central concern
with the centrality hypothesis is that
those positions defined as “central” have
been done so “ex post facto” – it is tough
28
considering Latino players specifically
adept at these roles, it will likely be
tougher for an up-and-coming Latino
outfielder or pitcher to arrive to the
majors in that same position. In my
research, I am not looking to consider
stacking by position, but rather the issue
of roster competition.
If certain
ethnicities are struggling more than
others to reach the majors, this bias
should be revealed.
one prove that whites are being selected
over more-qualified members of a
different race? Phillips relied purely on
statistics to analyze the possibility of
marginality, comparing batting averages
and slugging percentages of whites and
blacks for the 20 year period of 19601980. He concluded that the practice of
marginalization was indeed present for
the two decades, with blacks on average
having a batting average around 15-25
points higher than whites, and slugging
percentages about 40 points higher.
Thus,
extrapolated
out,
Phillips
concluded that on average, blacks from
1960-1980 would gain about 15 to 25 more
hits than whites per 1000 at-bats. He
does note that the separation between
statistics for the races begins to weaken
as the years progress, with a difference of
under 10 points by 1980.
Although
Phillips reports some data on Hispanic
ball-players, he states that even in 1980,
there were only 37 Hispanic players on
professional teams, so he did not wish to
derive any type of conclusive evidence
specific to them.
Even if Margolis and Piliavin have
shown that blacks are in positions less
central to the game, it has been noted
that blacks are overrepresented on
baseball rosters, experience no salary
discrimination, and on average, are more
productive than white baseball players
(Lavoie and Leonard, 1994).
In Diamonds in the Rough: The
Untold Story of Baseball, Zoss and
Bowman also discuss this idea, looking at
data from 1968 that shows an
overrepresentation of blacks among the
league’s top hitters as compared to the
average player (1996). However, it seems
strange that blacks are more productive
than their white counterparts; might this
mean that more marginal whites are
making it to the majors? What if even
more blacks should be participating in
MLB, but for some reason, are stuck in
the minor leagues?
Phillips’ study has three potential
drawbacks. First, his conclusions are
drawn only from a variety of statistics,
rather than basing them on an actual
model. Additionally, his statistics are
from players at the professional level,
after they have been chosen to compete
in MLB as everyday players, thus
prohibiting him from looking at how
these players ended up in the major
leagues. Many players may perform
substantially better once out of the
minor leagues, and with his study, there
is no way to observe why certain players
were chosen to compete at this level.
Finally, his focus on black players
Although Phillips considered
centrality in his study, his greater
emphasis focused on marginality,
defined as “the tendency to exclude
blacks of marginal ability in favor of
whites” (1980). Unlike salary data and
productivity, measuring marginality
becomes a more difficult task. How can
29
compared to white is less relevant today,
due to the large increase in Hispanic
players from 1980 to the present time.
However, although the percentage of
blacks in MLB may have decreased, a
recent vocal presence from a few black
major leaguers has brought the issue of
race back into consideration. Orlando
Hudson, a black infielder still active in
the major leagues, claims that due to
their race, qualified black free agents are
not getting as many quality contract
offers as their talent deserves (Passan,
2010). While others argue that it is these
players’ age, declining ability, and the
changing desires of major league
management that are to blame for the
lack of a suitable contract, Hudson
affirms that “there are some things in the
game that shouldn’t be going on,”
alluding to racism without explicitly
stating the word. Additionally, veteran
outfielder Torii Hunter, expressed
discouragement that many Hispanic
ballplayers are thought by fans to be
black, going as far to refer to his Hispanic
coworkers as “imposters” (Lacques, 2010).
While Hunter’s qualm rests with the
excessive hiring of Hispanic players, it
also indicates that there is an observable
amount of discontent within the nonwhite baseball community.
Finally,
Milton Bradley’s claim from the 2009
season that he was often the victim of
racial abuse from the outfield at Wrigley
Field suggests that even with complete
integration, unobservable factors may
exist affecting non-white major leaguers.
to the major leagues at some point in the
year, as well as those who were not. Each
of these additions should provide a
better indication of the differences
between players of different races and
ethnicities.
In my study I am expanding on
the research in Major League Baseball
surrounding centrality and marginality.
Specifically, I compare white players with
black players and those of Hispanic
background in an effort to determine if
non-white minor league players must
have offensive skills significantly better
than their white competitors in order to
be placed on a professional roster. If
black and Hispanic players must perform
better to be called up to a major league
team, this becomes relevant for both
players and baseball organizations.
Although it would be difficult to prove
that preferences for white players were
actually being exhibited, this study
would allow non-white players to benefit
from becoming aware that an implicit
discrimination was taking place, with
their white counterparts of comparable
ability being chosen instead of them.
Additionally, if organizations are not
choosing players based completely on
talent, they are limiting their potential to
win, thus not maximizing possible profits
for their business. Considering that fans
do not care about the racial make-up of
their team (Phillips, 1980), there is no
reason not to employ those players most
likely to improve a team’s chances at
success.
My study includes an empirical
model employing both black and
Hispanic players, and uses offensive
production of players at the minor league
level, including those who were called up
I use individual player data taken
from
www.thebaseballcube.com
of
players from the 2009 MLB season who
played on a team’s AAA affiliate (each
30
professional team’s most competitive
minor league team) at some point
throughout the season.
I regress a
variable indicating whether a player was
called up to the major leagues against a
variety of offensive production variables,
a player’s race, and variables measuring a
player’s age and their personal
experience in the major leagues. While I
expect to see offensive production highly
correlated with a player being called up,
the main focus of the regression is to
determine if a relationship exists
between players entering the majors and
their respective race. The results of my
OLS regression indicate that black and
Hispanic baseball players do not face a
significant disadvantage at gaining
promotion when compared to their
white counterparts. The results suggest
that a marginality bias is no longer
apparent today, appearing to allow nonwhite players a similar path to the major
leagues.
excellent fielding skills may be
considered valuable to a professional
team over one showing offensive success
with poor defensive capabilities. These
catchers may end up transferring to a
different position or finding a job as a
designated hitter. Only four catchers at
the major league level reached the
number of at bats needed to qualify for
the batting crown in 2009,1 with only
three of them batting .300 or above.
Additionally, of catchers qualifying for
this study, over two-thirds were white,
and without any of the remaining
catchers being black, including them in
the study seemed more of a hindrance
than a constructive addition.
Asian
baseball players were also excluded from
the study due to an insufficient amount
of players fitting the necessary
qualifications. A total of 164 players fit
the criteria discussed above for this
specific season. The sample includes 97
white players, 28 black players, and 39
players who have been identified as
Hispanic.
Data and Variables
The website from which data was
obtained contains all MLB team rosters
from the 2009 season. By analyzing each
roster, players with 50 or more minor
league at bats at the AAA level who also
obtained 10 or more at bats at the
professional level for the 2009 season
were included in my data. Pitchers were
not included in the sample, as their
offensive production is not considered by
organizations when determining the best
level of baseball in which they should
play. Additionally, catchers have been
excluded from this study for two reasons.
For one, a catcher’s offensive production
is not traditionally seen as their main
contribution to the game. A catcher with
These 164 players make up only
half of the data necessary to complete
the study. As each of these players was
called up to the major leagues in 2009, a
group of minor league players who were
not moved from AAA to the majors was
needed in order to identify if certain
races were favored. After finding the
players who received a call-up, a random
number generator was used to select five
players from each AAA team who
remained in the minor leagues for the
entire 2009 season. Like the previous
sample, pitchers and catchers were
1
Ironically, Joe Mauer, one of the four
qualifiers, won the title with his .365 average.
31
excluded, and players needed at least 50
at bats to be included, as offensive
statistics for players beneath this
threshold are likely to exhibit much
greater variance that is unrepresentative
of their actual talent. Of these 150
players randomly selected, 102 were
white, 14 were black, and 34 were of
Hispanic origin.
specific regression. It is also noted
whether a player is a rookie or veteran.
A player is considered a rookie if they
had played in less than 50 MLB games at
the end of the 2009 season.
Organizations are sometimes judged by
fans and columnists based on the
number of prospects they move through
their minor league system; if a team is
able to have their previous draft picks
succeed in the majors, the organization
and its scouts are likely to receive
continued support. Due to this, rookies
may not need as impressive minor league
statistics as players who shuffle in and
out of the major leagues; however, for
older players who have never had
significant playing time in the major
leagues, this attribute may be harmful to
their value, hurting their chances for a
call-up. A player is considered a veteran
if they have played in more than 500
major league games. Veterans may have
been sent down to AAA to work on a
specific aspect of their game or for
rehabilitation purposes, and their
production in the minor leagues may be
less important than a player trying to
establish themselves on a professional
roster. Overall age is also considered, as
older players may find it more difficult
than young talent to reach the major
leagues.
Each player’s profile on The
Baseball Cube includes a photograph of
the player and the location of their
birthplace. Black players were identified
by their photograph, while Hispanic
players were found based on a mix of
their
birthplace
and
physical
characteristics. All players born in Latin
America, South America, or other
Spanish-speaking
countries
were
included in the sample.
Players
identified as black Hispanics were placed
into the appropriate category based on
their birthplace – for example, a black
Hispanic born outside of the U.S. was
considered Hispanic, while those born
within the U.S. were considered black.
Three
variables specific to
offensive production were included for
each player. Batting average, slugging
percentage, and on-base percentage were
each considered as independent variables
(see definitions in the glossary).
Additionally, each player’s position was
noted, as perhaps there are differences in
the effects of a marginality bias due to
this – for example, outfielders are
generally considered to be better
offensive players, while infielders
(specifically shortstops and secondbasemen) may be awarded with playing
time based more on their defensive
ability. This issue is addressed later in a
Also included are a few variables
measuring diversity within each team’s
organization.
The study considers
whether the manager of a player’s team
is non-white, as well as the team’s
general manager, as those people higher
up in the system may often be making
choices independent of a coach’s desires.
It is possible non-white managers and
general managers treat potential non32
white players differently than white
managers, either not exhibiting any bias
whatsoever, or perhaps even a bias in
favor of non-white players. Non-white
owners were not considered in the study,
as only one professional team has a nonwhite majority owner.2
these teams exhibited an ease of
promotion different than white players.
Summary Statistics
Before introducing an empirical
model, it may be beneficial to analyze
summary statistics to get an idea of the
differences
between
each
race’s
production within the sample. Totaling
the race breakdown for the two samples
(players called up and those who spent
the entire year on a AAA affiliate), 199 of
the 314 players were white, 42 were black,
and 73 were Hispanic. The 42 black
players in the sample represent 13.3% of
the total players identified for the study,
while the 73 players of Hispanic origin
are about 23.2% of the total players.
Whites make up the remainder, at 63.3%.
The amount of non-whites in the sample
is
a
bit
different
than
their
representation in all of the MLB, with
blacks here representing 13, rather than
10 percent, and Hispanics at 23 percent,
rather than 27; however, these
differences appear small enough to still
likely portray an accurate picture of the
league’s racial make-up.
Finally, the study observes
whether a team plays within a city
containing a high percentage of nonwhite citizens. Perhaps management
takes the fan base into consideration
when constructing a team, preferring to
draft or promote non-white players to
allow the team’s diversity to coincide
with the city itself. With the majority of
MLB teams playing in large, diverse
metropolitan areas, to be considered a
diverse city for this study teams must
play in a metropolitan area with nonHispanic whites representing about a
third of the population. Specifically,
teams playing in cities with a nonHispanic white population below 37% are
considered diverse for this study.3 For
variables indicating a non-white manger,
non-white general manger, and diverse
city population, new variables were
constructed for the regression interacting
a player’s race with the aforementioned
variables; in doing this, the analysis will
show whether non-white players on
First, as the basis of this study is to
conclude whether non-white players
being called up must exude greater
offensive productivity than whites, it may
be beneficial to look at the data for those
players who have been called up. The
average offensive production at the AAA
level for these players in the sample is
outlined below in Table 1. It is clear from
the data that whites in the sample
exhibited a batting average of about 20
points lower than the non-whites. To
illustrate how significant this 20 point
difference is, it is interesting to consider
2
As of the 2010 season, Arte Moreno,
majority owner of the Los Angeles Angels of
Anaheim, is the only non-white owner of an
MLB team.
3
While this percentage appears arbitrary, the
majority of cities clustered in either the 50-70%
or 20-40% range; in order to capture a
representative amount of extremely diverse cities,
the 37% cutoff was implemented.
33
batting averages in the major leagues for
the 2009 season. For players with an
average of at least 3.1 at bats per team
game, the top 56 players in the league hit
better than .290, while the top 100 hit
.270 or better4. Considering that only 102
batters in the MLB qualified for the
necessary at bats, a difference of 20
points can propel a player from the
lowest percentile to right around the
league average.
study. It appears that the disparity
between races may not be as extreme as
it appeared before. While blacks and
Hispanics still have an average batting
average greater than whites, the
difference is reduced to 5-10 points.
Additionally, white slugging is no longer
lowest, now centered relatively evenly
between Hispanics and blacks. Again,
Hispanics experienced the lowest
slugging percentage, this time to a much
greater extent than before. Hispanics
who were not called up exhibited an
average slugging percentage a full 40
points below whites, and over 60 points
below blacks. The difference of slugging
for Hispanics not called up and those
who were turns out to be almost 100
points, an enormous difference, that if
anything, shows that management may
keep a particular eye out for Hispanics
with the ability to hit for power, as this
category is shown to be weakest for
Hispanics in both samples. A table
summarizing statistics for call-ups and
non-call-ups can be found in the
appendix. All three measures included in
the tables above are included as
independent variables within the later
regression, as the decision-makers on
each ballclub may weight certain
production measures greater than others.
For on-base percentage and
slugging percentage, on average, whites
and Hispanics appear to have more
similar statistics, with whites lower in
on-base percentage and higher in
slugging, while blacks exhibit greater
averages in all categories. However, the
most notable difference appears to be
batting average, where the .271 average
for whites stands far below the numbers
for other races.
However, a better sample to
consider might include the players who
were not called up. Doing so should
present a clearer picture as to whether
non-whites who didn’t get called up
achieved offensive statistics superior to
whites who were called up.
Before
furthering the study, this might help in
predicting
the
possible
outcome.
Production numbers for all players not
called up are listed below in Table 2.
The average age of each player in
the total sample is 26.9, with whites
being the oldest at 27.3 years of age,
followed by Hispanics at 26.4, and blacks
at 26.1. While these numbers may not
tell us much, it is interesting to consider
that blacks and Hispanics may develop
more quickly than whites, or that teams
are more willing to bring whites into the
league at a later age, thus allowing them
more time to develop in the minor
First, the numbers reveal that
neither blacks nor Hispanics who were
not called up were statistically superior
in any category to the white players who
were called up; such a difference would
have been extremely revealing for the
4
Data obtained from www.mlb.com
34
leagues. However, it is important to note
that Hispanic players born outside of the
U.S. are eligible to be signed without a
high school education, while U.S.-born
players must first graduate high school
before becoming eligible, thus giving
some Hispanic players a greater time to
develop in the professional system. Of
the 152 rookies in the sample, 100 were
white, 19 were black, and 33 were
Hispanic. For veterans, of which there
were only 27, the majority was white,
with 17 in the sample; of the remainder, 4
were black and 6 Hispanic.
This equation seeks to discover if
there is a link between a player’s race and
their propensity to be called up to the
major
league
level.
Offensive
production, as well as other variables
throughout the equation, are used to
control for the many factors influencing
an organization’s decision to promote a
player. Using an OLS regression, the null
hypothesis being tested is that β 1 and β 2
= 0. Under normal circumstances, it is
assumed that a player’s race should not
have an effect on the probability they are
promoted from AAA to the major
leagues. As this study is testing whether
Hispanics and blacks have a decreased
probability of getting called up, this will
be a one-tailed test, as results causing the
null hypothesis to be rejected will be
indicated by negative coefficients on β 1
and β 2 . However, the significance of all
other variables in this test is measured
based on a standard two-tailed test. If
the two coefficients from the null
hypothesis are significant and less than
0, this would suggest that for black
and/or Hispanic players, their race alone
inhibits them from having the same
probability as an equally talented white
player from reaching the majors. All
variables from β 7 to β 11 are binary,
functioning as either a 0 or 1 depending
on if the player fits specific criteria. Age
is measured as a whole number, based on
the player’s age at the end of the 2009
season. It should also be noted that
values for batting average, on-base
percentage, and slugging percentage
were converted into whole numbers for
the analysis. Table 3, as seen in the
appendix, reveals the results for the
regression using Equation 1.
Of the 30 MLB teams, 8 employed
non-white managers for the 2009 season,
while 5 had non-white general managers.
A total of 14 teams qualified for the
diverse city measure, with Detroit’s 12.3%
non-Hispanic white population making it
the least white city in the sample.
Empirical Model and Results
My empirical model is based on
one specific equation, which is run first
for all players within the study. As the
dependent variable calledup is binary,
coefficients on the independent variables
will represent each variable’s influence
on the probability a player will be called
up to the major leagues.
My initial
regression is based on the following
equation:
(1) Called Upallplayers = β0 + β 1black + β
2hispanic + β 3battingaverage + β 4obp +
β 5slugging + β 6age + β 7rookie + β
8veteran + β 9NWmanager * NWplayer +
β10NWgm * NWplayer + β11diversecity *
NWplayer + u.
Note: ‘NW’ stands for Non-white.
35
The majority of the significance
within the results takes place on the
control variables.
However, before
considering each of these, let us first
consider the outcome for the variables
from the null hypothesis. We fail to
reject the null hypothesis that β 1 and β 2
are equal to zero; in fact, both
coefficients proved to actually be
positive, suggesting that if the results
would have been statistically significant,
blacks and Hispanics may have an
advantage over whites in reaching the
major leagues. According to the results,
a black AAA player would be about 3%
more likely to reach the majors for the
2009 season, while Hispanics would
receive a 1% advantage. Of course, due
to their insignificance, these coefficients
cannot actually be considered to actually
reach this conclusion. When comparing
white, black, and Hispanic player
outcomes, it appears that a marginality
bias in favor of whites is not present in
the 2009 season.
slugging, the results are similar,
suggesting a 0.1% increased chance of
being called up for each one point
increase in slugging percentage. It is
extremely interesting to note that the
results for batting average were both
negative and insignificant. One modern
columnist writes that batting average
“remains
the
one consistently
understood aspect of the game”
(Morrissey, 2007), as even the most lay
fan has little trouble interpreting the
measure. However, the explosion of the
more obscure statistic as made publicly
popular in Michael Lewis’ Moneyball, as
well as the way in which modern baseball
has “create[d] some stunning numbers
for at bats [and] home runs” (2007), each
suggest that both organizations and fans
alike may be less focused on a simple
statistic such as batting average, paying
more attention to other possible
outcomes at the plate. When there exist
seven ways to reach first base,5 why
would people want to limit their
observation to simply hits per at bat?
On-base percentage is able to include
additional ways in which a player may
get on base, and slugging adequately
measures a player’s ability to hit for extra
bases. A handful of the MLB’s top
current players are known for having
traditionally low batting averages while
maintaining much larger on-base and
slugging percentages.6 Of the three
offensive measures, it seems appropriate
that batting average did not provide
Of the offensive productivity
control
variables,
both
on-base
percentage and slugging percentage
proved statistically significant at the 5%
and 1% level, respectively.
Both
coefficients were positive, as would be
expected – it seems natural that the
greater a player’s production, the more
likely they are to be promoted. For every
point increase in on-base percentage, a
player, regardless of race, should expect a
0.2% increase in the probability of being
called up; so, if an average black player is
hitting 20 points higher than an
otherwise identical white player (for
example, .280 compared to .260), there is
a predicted 4% greater chance they are
promoted to the major league level. For
5
Hit, walk, hit by pitch, fielder’s choice, error,
catcher interference, and dropped third strike
6
Adam Dunn and Lance Berkman each finished
within the top 15 for OBP and the top 35 in slugging
for the 2009 season, while neither batted over .275
36
significant results, as it seems it has been
overlooked in recent years, and players
may be exuding talent in alternative ways
to impress organizations.
remember that all players within the
study
have
already
established
themselves at the AAA level, where they
are only one step away from the major
leagues.
The variable for age showed
significance, indicating that for every
year older a player becomes, their
chances of being called up decrease by
about 2%. While this coefficient seems
appropriate, as one would assume teams
would be less likely to bring up an older
player with less potential and fewer
successful years ahead of him, it is
unlikely that age possesses a completely
linear relationship. It is more likely that
this variable would be negative for the
youngest players and become positive for
a brief time when players are at the
average period of sufficient development
to be ready for the majors. After drafting
a player, a team is allowed to keep him
within the minor league system for six
years before they must either promote
him to the major leagues or find an
alternative solution, such as releasing or
trading the individual. With this in
mind, it is likely that players aged 24-27
are being called up the majors more than
those falling out of that range.
While the results for rookies
appear legitimate, more caution should
be considered when analyzing the results
for veterans. Although the coefficient
suggests that having played in at least
500 major league games garners a 25%
increase in the likelihood of promotion,
of the 27 veterans in the study, only 7
were not promoted to the majors in the
2009 season. It is unlikely that all 27 of
these players were in the minors fighting
for a chance to move up, as many could
have been either recovering from an
injury or called up at different intervals
throughout the year to act as pinch
hitters or fill a spot on the roster. Also
noteworthy is that veteran players, as
defined by MLB as those with five or
more years of service, must approve of
being sent down to the minor leagues – if
they don’t, the team can either keep
them on the major league roster or
release the player and continue paying
their salary. With this in mind, it is
possible some of the veteran players in
this study agreed to play in the minor
leagues, perhaps acknowledging that
their talent level has decreased.
The significant coefficient on the
variable for rookies illustrates that for
players who have not played in at least 50
games at the major league level, their
chance at gaining promotion is hindered
by almost 15%. It makes sense that a
player with past major league experience
would be more likely to have the skills to
reach a professional team again.
Additionally, while one may consider this
coefficient to be inflated due to
extremely young players without present
hopes of promotion, it is important to
Unfortunately, none of the
interaction variables accounting for
diversity within management or the
team’s city proved significant.
The
results indicate that non-whites playing
under diverse management actually had
a more difficult time reaching the
majors. However, it does appear that
non-whites in the most diverse cities
37
were more likely to be called up, a
statistic that was significant at the 25%
level.
However, the coefficients for on-base
percentage and slugging are only
significant at the 10% level, where before
they reached significance at the 5 and 1%
levels, respectively. The coefficient for
on-base percentage has increased from
about .002 to .003, indicating that this
statistic may be of more importance to
management for outfielders than
remaining position players. Slugging,
however, appears almost unchanged,
moving from .0013 down to .0012. Older
players may have a more difficult time
being called up for outfield positions,
with a negative coefficient of -.047 for
the age variable, suggesting that an
outfielder has about a 5% less chance of
being called up for each year older they
become. Again, it is important to realize
that it is unlikely age represents a linear
relationship. Although 51 of the 111
outfielders in the sample qualified as
rookies, the -.046 coefficient, over three
times as large as that of the first
equation, was insignificant. However, it
appears veteran outfielders, of which
there were 7, found it almost twice as
likely as non-veterans to be called up,
although it is unwise to over-analyze
this, as only one veteran outfielder in the
study was not called up. One final area
of note for this second equation is the
higher R-squared value of .2605 as
compared to the initial .1906, suggesting
that it may be easier to predict call-ups
for outfielders than for all players as a
whole. Of course, this makes sense, as
by limiting the model to outfielders, we
have eliminated the possibility of a
player’s position affecting the results.
In Phillips’ study, to control for
the fact that in 1980 many black baseball
players played predominantly outfield
positions, he decided to control for this
and directly compare statistics for white
and black outfielders. In doing so, any
variation in offensive productivity based
on position is eliminated. Similarly, I
decided to run a separate regression
which includes only players involved in
one of the three outfield positions. Of
the 111 outfielders in the study, 70 were
white, 27 were black, and 14 were
Hispanic. The regression was run based
on the same equation as before, only
with a different sample of players. The
equation
is
reprinted
below.
(2) Called Upoutfielders = β0 + β 1black +
2hispanic + β 3battingaverage + β 4obp
β 5slugging + β 6age + β 7rookie +
8veteran + β 9NWmanager*NWplayer
β10NWgm*NWplayer + β11diversecity
NWplayer + u.
β
+
β
+
*
Again, using an OLS regression, the null
hypothesis being tested is that both β1
and β 2 = 0. All of the same control
variables are used, with the only
difference in regressions being that nonoutfielders have been removed from the
data. Results are shown in table 4 in the
appendix.
The
results
appear
extremely
consistent with those from the first
equation. Each variable significant in
the previous results, except for the
rookie variable, is again significant here.
Conclusion
38
Unlike previous studies, which
often evaluated players who were already
established in the major leagues, this
study examines the path that players of
different races must take in order to play
on a professional roster. Players who
moved from a AAA affiliate to a major
league ballclub during the 2009 season
were included in the data, as well as
randomly selected players who were not
promoted within the season.
A
regression using whether a player was
called up as the dependent variable used
a variety of variables to find what
organizations consider when promoting
a player; specifically, the study desired to
find if non-white players exhibited a
more difficult time being called up from
AAA to a major league team. Although
many of the control variables proved
significant, neither blacks nor Hispanics
appeared to face any additional barriers
when seeking promotion.
performing well are likely to have similar
statistics to the blacks and Hispanics
being promoted. In this situation, a
lower percentage of black and Hispanic
players in the minor leagues are
performing as poorly as whites, with a
greater portion from each race promoted
than whites. So, wouldn’t this mean that
there are more whites in the minor
leagues than the majors? By referring
back to the summary statistics for those
players who were not called up, it
appears the figures for this study support
this idea. About 68% of the 150 players
not called up were white, while whites
represented only 60% of players in the
MLB last year. Similarly, Hispanics not
called up were only 22.6% of the sample,
over 4% lower than their MLB presence.
Although blacks not called up were
proportionately only 1% lower than the
MLB average, those who did make the
move to the majors overrepresented the
league average by 3%.
With such results, one must
return to the drastic differences in
offensive production between white and
non-white players. What is to account
for the discrepancies in offensive
production between the races? Perhaps
one idea to consider is the raw number of
individuals of each race within the
system. From the table in the appendix
of all players in the sample, we see that at
least for batting and on-base percentage,
white minor leaguers tend to perform
inferior to other races. If the number of
whites in the minors is significantly
greater than other races (which it is,
making up about two-thirds of the
league), then the distribution of offensive
production makes more sense. With a
large portion of whites pulling down
these average numbers, those who are
While the results of this study do
not support the idea of a marginality bias
currently present in the MLB, a great
deal is still captured by these results. It
appears that more in-depth offensive
statistics, such as on-base percentage and
slugging percentage, are being given
more weight when deciding a player’s
promotion. Additionally, regardless of
the manager, general manager, or racial
make-up of the city in which a team
plays in, it still does not appear that race
is a factor in promotion. Although
Orlando Hudson correctly points out
that both Jermaine Dye and Gary
Sheffield remain unsigned, successful
white players, including pitcher Jarrod
Washburn and power-hitting third
39
baseman Joe Crede, remain on the
market as well.
It would be interesting to expand
this study in a few different ways.
Adding variables measuring defense
could be worthwhile, as well as a variable
specific to language. While this study
examines race, it is likely that a large
range of fluency in English exists among
the players included. The data also does
not capture if there was a specific reason
a player was called up; situations in
which a player is needed to replace an
injured starter are not included in this
data. Additionally, teams out of playoff
contention may bring up minor league
players for a period of evaluation, having
no real intention of keeping them on the
roster for the following season.
Of
course, with the constantly changing
trends in baseball, be it statistically or in
the league’s racial make-up, there is no
reason a similar study at a later date
couldn’t produce revealing results
different than those observed here.
40
Appendix:
Table A-1: Average Minor League Offensive Production By Race (All Players)
Batting Average
On-Base Percentage
Slugging Percentage
White
0.26046
0.33297
0.41064
Black
0.28088
0.35567
0.43383
Hispanic
0.27597
0.33390
0.38945
Table 1: Average Minor League Offensive Production By Race (Call-ups Only)
Batting Average
On-Base Percentage
Slugging Percentage
White
0.27172
0.34601
0.44225
Black
0.2913
0.36839
0.44971
Hispanic
0.29505
0.35208
0.43346
Table 2: Average Minor League Offensive Production By Race (Non Call-ups Only)
Batting Average
On-Base Percentage
Slugging Percentage
White
0.24975
0.32056
0.38059
Black
0.26014
0.33021
0.40207
Hispanic
0.25409
0.31306
0.33897
41
Table 3: Regression using Equation (1)
(1)
Dependent
Variable:
Called Up (All Players)
β 1 Black
.0307(.0950)
β 2 Hispanic
.0113 (.0831)
β 3 BattingAverage
-.0007 (.0012)
β 4 OBP
.0021** (.0010)
β 5 Slugging
.0013*** (.0005)
β 6 Age
-.0237** (.0103 )
β 7 Rookie
-.1381** (.0605 )
β 8 Veteran
.2529** (.1062)
β9 NWmanager*NWplayer
-.0164 (.1041)
β 10 NWgm*NWplayer
-.0412 (.1291)
β11Diversecity*NWplayer
.1082 (.0918)
β 0Constant
.1182 (.3579)
Number of Observations
314
R-Squared
.1906
Adjusted R-Squared
.1611
***Significant at the 1 percent level.
**Significant at the 5 percent level.
*Significant at the 1 percent level.
42
Table 4: Regression Using Equation (2)
(2)
Dependent
Variable:
Called Up (Outfielders)
β 1 Black
.0293 (.1352)
β 2 Hispanic
.1804 (.1742)
β 3 BattingAverage
-.0016 (.0020)
β 5 Slugging
.0012*(.0007)
β 6 Age
-.0469**(.0193)
β 7 Rookie
-.0459(.1052)
β 8 Veteran
.4482**(.2007)
β 9 NWmanager*NWplayer
.0421(.1712)
β 10 NWgm*NWplayer
-.1323(.2300)
β 11Diversecity*NWplayer
.0576(.1628)
β 0Constant
.6261(.6482)
Number of Observations
111
R-Squared
.2605
Adjusted R-Squared
.1783
***Significant at the 1 percent level.
**Significant at the 5 percent level.
*Significant at the 1 percent level.
Glossary
On-Base Percentage (OBP): Number of
Hits, Walks, or Hit-By-Pitch / Number of
At Bats + Number of Walks + Number of
Hit-By-Pitch
Batting Average: Number of Hits /
Number of At Bats
Slugging Percentage: Number of Bases /
Number of At Bats (for example, a
double would count as two hits, a triple
three, and a home run four)
Designated Hitter: In the American
League, a player who is part of the nine
person lineup but does not play a
43
position in the field (the pitcher does not
bat)
Multivariate Analysis." Sociology of Sport
Journal 16 (1999): 16-34. Print.
Works Cited
Mills, P. "Negro League History 101."
Negro League Baseball. P. Mills, 2003.
Web.
8
Apr.
2010.
<http://www.negroleaguebaseball.com/h
istory101.html>.
Gonzalez, G. Leticia. "The Stacking of
Latinos in Major League Baseball: A
Forgotten Minority?" Journal of Hispanic
Higher Education 1.4 (1996): 320-28.
Print.
Morrissey, Mo. "MLB's Top 50 Batting
Averages: The Lost Importance of the
Batting
Average."
Associated
Content, 30 Nov. 2007. Web. 29 Apr.
2010.
<http://www.associatedcontent.com/arti
cle/465048/mlbs_top_50_batting_averag
es_the_lost.html>.
Lacques, Dave. "Torii Hunter Upset His
Comments Caused a Stir as Ozzie
Guillen, Others Weigh in." Daily Pitch.
USA Today, 10 Mar. 2010. Web. 29 Apr.
2010.
http://content.usatoday.com/communiti
es/dailypitch/post/2010/03/torii-hunterscomments-cause-stir-as-ozzie-guillenresponds/1>.
Passan, Jeff. "Hudson Hints at Racism for
Blacks in Free Agency." Yahoo! Sports MLB. 13 Apr. 2010. Web. 29 Apr. 2010.
<http://sports.yahoo.com/mlb/news?slu
g=jp- dyehudson041210>.
Lapchick,
Richard,
Alejandra
Diaz-Calderon and Derek McMechan.
The 2009 Racial and Gender Report Card:
Major League Baseball. UCF Catalog.
University of Central Florida. College of
Business Administration, 15 Apr. 2008.
Web.<http://web.bus.ucf.edu/sportbusin
ess/>
Phillips, John C. "Race and Career
Opportunities in Major League Baseball:
1960-1980." Journal of Sport and Social
Issues 7.2 (1983): 1-17. Print.
Zoss, Joel, and John Bowman. Diamonds
in the Rough: The Untold History of
Baseball. Chicago, Ill: Contemporary,
1996. Print.
Lavoie, Marc, and Wilbert M. Leonard II.
"In Search of an Alternative Explanation
of Stacking in Baseball: The Uncertainty
Hypothesis." Sociology of Sport Journal 11
(1994): 140-54. Print.
Lavoie,
Marc.
"The
"Economic"
Hypothesis of Positional Segregation:
Some Further Comments." Sociology of
Sport Journal 6 (1989): 163-66. Print.
Margolis, Benjamin, and Jane A. Piliavin.
""Stacking" in Major League Baseball: A
44
Narrative Processing Predicts Well-Being and
Communion in Victims of Intimate Partner Violence
Courtney Walsh, Class of 2010
Narratives may play an important role in understanding and healing from traumas among
victims of intimate partner violence. This study investigated whether narrative processes of
overcoming and integrating traumas into overall life stories (i.e., narrative resolution,
complexity, and growth) relate to female victims’ overall psychological well-being and
capacities for close, meaningful relationships following intimate partner violence (N=32).
To assess predictor and outcome variables, a narrative protocol and questionnaires
examining participants’ well-being, relationship attitudes, and relationship engagement
were used. Results showed that victims’ levels of narrative processing were generally
predictive of their intimate relationship attitudes, positive relations with others, and
overall well-being. Decreased rumination partially mediated the relationship between
narrative resolution and well being. Overall, findings indicate the importance of narratives
in trauma recovery in the case of intimate partner violence.
Storytelling is pertinent to an
individual’s self-insight and development
both in the therapy setting and in life
overall.
Storytelling,
or
narrative
processing,
is
the
narration,
interpretation, and incorporation of past
experiences into individuals’ life stories
(Pals, 2006). Consequently, narrative
processing is important in making
meaning of past events, particularly
traumatic experiences, within the
context of the life story as a whole.
People’s stories vary in their inclusion of
narrative
components,
such
as
complexity, resolution, and growth.
Overall, the sharing of one’s experiences
is linked to identity development, wellbeing, positive relationship development,
and meaning in one’s life (e.g., King,
2001; Linley & Joseph, 2003). Past
research has shaped the current study’s
focus on a particular trauma in women,
Intimate Partner Violence (IPV), and its
relationship to narrative processing,
which has the potential to be a powerful
coping mechanism in trauma victims.
Understanding coping in IPV victims is
essential, as being able to identify
successful coping mechanisms has
positive implications for recovery and
therapy for victims. The present study
centers on one main question: What
narrative processes (i.e., complexity,
resolution, and growth) predict IPV
victims’ well-being and capacities for
close relationships?
The Role of Life Stories and
Narratives in Personal Development
Life stories are complex, evolving,
psychosocial constructions (i.e., both
individuals and their cultures author and
shape these stories) associated with
identity development (McAdams &
Bowman, 2001). Individuals constantly
revise their unique life stories to comply
with cultural expectations and to
incorporate past, present, and future
events into these stories (McAdams,
1998, in McLean, 2008). Furthermore,
meaningful storytelling, reflection on
45
experiences, and differences in life
experiences are pertinent to identity
development within a culture (McLean,
2008).
Essentially, identity is an
internalized,
evolving
life
story
(McAdams, 2001).
process difficult life experiences. For
example, a study by Pennebaker and
Seagal (1999) found that narrative
writing helps organize, understand, and
create a coherent, integrative account of
complex experiences over time. In other
words, speaking about an emotionally
stigmatizing experience (e.g., IPV) can
turn a complex event into a simpler,
more efficient story. On the other hand,
failure to address difficult experiences
and traumas has negative effects on the
individual. Specifically, the silencing of
abuse, both individually and culturally,
leads
abusive
experiences
and
accompanying stories to be perceived as
uninteresting,
irrelevant,
or
unacceptable, which further leads to
reluctance in individuals to speak about
their experiences. It is vital, however, for
individuals to talk about the difficult
events in their lives, as it gives those
events meaning and enhances selfconcept and self-understanding (Fivush,
2004). Overall, narrative processing is
significant
in
personal
identity
development and also increases one’s
ability to understand, resolve, and
incorporate trauma into an overall life
story.
Past research has shown the
importance of how individuals convey
and conceptualize narratives within the
context of their overall lives. In a study
by Sanderson and McKeough (2005), the
experimenters found that narrative
thought is useful in understanding life
events, human intentions/actions, and
different event outcomes. Specifically,
the results revealed that, relative to a
therapy setting, narrative language helps
clients work through how events are
causally related and how one’s internal
‘intention
state’
(i.e.,
emotion,
motivation) shapes one’s external actions
(Sanderson & McKeough, 2005). These
findings are related to this study’s
examination of how narrative processes
of growth, complexity, and resolution
relate to a person’s relationship goals,
needs, and overall feelings of well-being
(i.e., life satisfaction, daily functioning,
and positive emotion) (King, 2001). In
addition to the importance of the
content of a narrative, Sanderson &
McKeough
also
emphasize
the
importance of individuals’ manner of
conveying their narratives; this is
important because the narrative can
highlight participants’ unresolved issues,
meaning making abilities, and manner of
incorporating stories and negative events
into their overall lives.
Individual Differences in Narrative
Processing: Implications for WellBeing
Narrative Processing: Importance to the
Life Story Model of Identity
Although life narratives have the
potential
to
promote
positive
development in individuals, narratives
vary substantially in their ability to do so.
Various processes are integral to one’s
narrative and illustrate individual
Narrative processing is also
beneficial for one’s ability to better
understand, positively transform, and
46
differences in thought processes and
emotional maturity. The current study
mainly addresses three of these
components:
narrative
complexity,
narrative
growth,
and
narrative
resolution. First, narrative growth is a
positive change in thinking that occurs
when reflecting on a difficult event,
leading to maturity, strength, and
positive functioning (Mansfield, 2009).
Second, narrative resolution occurs when
an individual makes peace with difficult
life experiences within the context of his
or her life story (Mansfield, 2009). Third,
narrative complexity refers to individuals’
depth of thought in narratives and their
abilities to see different perspectives and
outcomes
(Mansfield,
2009);
this
narrative process is reflected in plots,
themes, characters, and settings that are
intertwined into a person’s overall life
story (Singer, 2005). In the current study,
IPV experiences can be classified as self
defining memories. Such memories are
significant scenes in the life story, as the
construction and interpretation of these
memories in the midst of personal, longterm goals gives life unity and purpose
(Singer & Salovey, 2003; Singer, 2004;
Tomkins, 1987, in McAdams et al., 2006).
Narrative resolution may be especially
important for handling traumatic self
defining memories such as IPV. Each of
these three aspects of narrative
processing—narrative
complexity,
narrative
growth,
and
narrative
resolution—may play a critical role in the
development of well-being, particularly
following trauma, as reviewed next.
goals, and find redemption in their
traumas is related to an individual’s
overall well-being. While a few studies
have shown growth to be marginally
predictive of well-being (see Mansfield,
2009), other studies have found growth
to be significantly linked to well-being
(see Bauer, McAdams, & Sakaeda, 2005a;
Bauer, McAdams, & Pals, 2006; Bauer &
McAdams, 2004). Bauer, McAdams, and
Pals (2006) discuss how high levels of
eudaimonic
well-being
(i.e.,
an
individual’s overall ego development,
psychosocial integration of difficult
experiences, and psychological wellbeing) are linked to an accentuation of
personal growth, development, and
viewing of traumatic experiences as
transformative and insightful in life
stories. As such, a narrative focus on
growth is related to an overall sense of
well-being in an individual.
The presence of redemption
sequences in individuals’ life stories is
also indicative of growth from the
trauma. Redemption sequences occur
when a negative life experience is
followed by or transformed into a
positive life experience scene (McAdams,
Reynolds, Lewis, Patten, & Bowman,
2001).
Conversely,
contamination
sequences occur when a positive life
experience is followed by or transformed
into a negative life experience scene. For
example, a redemption sequence occurs
if a person who descends into heavy
drinking then decides to make the
positive life change of becoming sober;
contrarily, a sober individual deciding to
heavily drink would be an example of a
contamination sequence. In samples of
both midlife adults and undergraduate
students, McAdams, Reynolds, Lewis,
Narrative Growth
The ability of individuals to
positively change their thinking, set
47
Patten, and Bowman (2001) found that,
in contrast to contamination sequences,
redemption sequences in narratives were
correlated with high levels of well-being,
life satisfaction, self-disclosure, and self
esteem. Recovery from traumas such as
IPV may similarly be linked to an
enhanced self-view and overall changed
life philosophy. Ultimately, the presence
of redemption sequences in narratives is
related to higher levels of well-being,
enabling healing and resolution to occur.
of narrative resolution, including
narratives’ presence of crystallization of
desire vs. discontent and coherent
positive resolution.
Individuals’ perception about
important events in their lives,
particularly traumatic ones, is important
to their well-being levels. In looking at
adults’ narratives of life changing
decisions, Bauer, McAdams, and Sakaeda
(2005b) found that crystallization of
desire (focusing on “approaching a
desired future”) leads to a greater level of
well-being/life satisfaction, more positive
self-view, and less trauma avoidance
than does crystallization of discontent
(focusing on “escaping an undesired
past”). For example, following negative
experiences, crystallization of desire
leads individuals to seek happiness,
pleasure,
and
success
whereas
crystallization of discontent leads
individuals to escape failures and
negativity. Thus, a dichotomy exists
between a positive approach orientation
(i.e., crystallization of desire) and
negative avoidance orientation (i.e.,
crystallization of discontent).
Growth also results from life
goals, which are linked to narrative wellbeing and identity development. When
setting life goals, an individual seeks to
become the best version of him or
herself, which is referred to as one’s ‘best
possible self’ (King & Hicks, 2006). In
narratives, the ease, vividness, and
emotional depth with which individuals
think about their best possible selves is
pertinent to growth. Several studies have
found that investing in attainable goals
and letting go of unattainable goals is
strongly associated with well-being and
ego development (King & Hicks, 2006).
Overall, positive change after difficult
experiences
(i.e.,
growth)
and
establishment of achievable goals are
linked to individuals’ well-being levels.
Within this context, confronting
trauma in a positive manner is also
linked to increased trauma resolution
and well being. Pals (2006) describes a
two step process in narrative reasoning
and post-trauma growth: exploring the
impact of a meaningful, negative
experience in depth and constructing a
resolved, positive ending of selftransformation. In a longitudinal study of
adults who have had difficult life
experiences, Pals found that positive self
transformation is likely to lead to wellbeing. She looked at “coherent positive
Narrative Resolution
Like growth narratives, narrative
resolution is also linked to positive wellbeing. In this sense, resolution indicates
individuals’ ability to make peace with
their traumas, create a positive solution
for their traumatic experiences, and
disallow such traumas from continually
negatively affecting their daily lives.
There are variations in individuals’ levels
48
resolution”, which occurs when a well
integrated life story of a difficult
experience ends positively and leads to
adult identity development. It follows
that “coherent positive resolution” leads
to higher levels of overall well-being
(e.g., resiliency, life satisfaction), which is
an important outcome measure in the
present study.
eudaimonic well-being, they view
traumatic experiences as transformative
and capable of offering insight;
furthermore, in addition to having
positive feelings, individuals with high
levels of well-being are able to perceive
and integrate difficult events into their
overall self-view. Overall, the integration
and incorporation of thoughts, feelings,
and experiences into one’s life story is
the hallmark of narrative complexity.
Furthermore, in studies assessing
post-trauma recovery and counteraction
of past traumas, present control (i.e.,
control over recovery) and positive life
changes were found to be associated with
less distress, fewer depressive symptoms
(e.g.,
social
withdrawal),
greater
cognitive restructuring, and higher
effectiveness in recovery than past (i.e.,
self-blame) or even future control
(Frazier, Mortensen, & Steward, 2005;
Frazier, 2003; Frazier, Tashiro, Berman,
Steger, & Long, 2004). Overall, victims’
levels of peace with their trauma, control
over their recovery from trauma, and
focus on positive solutions to their
trauma are related to their psychological
well-being. In victims, approaching a
desired future and creating a coherent
and resolved narrative are linked to
increased positive identity development,
life satisfaction, and resiliency.
Happiness and maturity are also
linked to narrative complexity and wellbeing. King (2001) discusses how
happiness and maturity are two aspects
of the good life. The good life results
from the incorporation of difficult life
experiences into one’s narrative (e.g.,
including change, suffering, regret, and a
positive ending). Individuals with
complex narratives are able to see and
experience differing outcomes, including,
regret, agony, and positive change.
Contrary to having an outcome of regret,
the ability to find redemption after
considering different perspectives in a
difficult life event is linked positive
psychological well-being (Ryff, 1989, in
King, 2001). Particularly in adult women
who have lived in an unequal, limiting
environment, difficult life experiences
have the opportunity to enhance
complexity of life perspectives, which is,
in turn, related to the ‘good life’ (see
Helson, 1992, in King, 2001).
Narrative Complexity
In the midst of processing and
understanding difficult life events,
narrative complexity can occur, which is
linked to well-being. Following from the
previous discussion, narrative complexity
is an important predictor of well-being,
even in the presence of narrative
resolution (Mansfield et al., 2009). If
individuals have
high
levels of
In individuals facing difficult
experiences, specific words in narratives
denote depth of processing and predict
overall well-being. Using a meta-data
analysis
from
previous
studies,
Pennebaker, Mayne, and Francis (1997)
found that positive-emotion words in
49
narratives are predictive of increased
psychological health and individuals with
a mixture of many positive-emotion
words and some negative-emotion words
have the highest amount of well-being.
These narratives are considered to have
high levels of narrative complexity, as the
inclusion of both positive and negative
emotion words illustrates individuals’
abilities
to
see
differing
perspectives/outcomes and to intertwine
differing emotions in their narratives.
Overall, the integration of fragmented
negative and positive emotions is
indicative of narrative complexity and
consequent overall psychological wellbeing.
2004; Pennebaker & Seagal, 1999).
Overall, individuals’ integration of
multiple emotions and perspectives into
their trauma narratives is predictive of
increased well-being, life satisfaction,
and positive identity development.
The Relationship between Narrative
Processes and Communion
Communion can be defined as
individuals’ experiences with love and
intimacy as well as their connections to
others
shown
through
helping,
nurturing, and affection (Bakan, 1966, in
McAdams, Hoffman, Mansfield, & Day,
1996; Singer, 2005). In my study,
communion is examined as an
individual’s
intimate
relationship
attitudes and positive relations with
others. Although communion is mainly
an outcome measure in this study,
McAdams, Hoffman, Mansfield, and Day
(1996) found that communion themes in
narratives
(i.e.,
love/friendship,
care/help) are positively associated with
intimacy motivation/attainment and
personal needs for close, warm affiliation
and nurturance. In essence, communion
themes
encompass
individuals’
sentiments towards their interpersonal
relationships; these themes are then
associated with the externalized outcome
of communion, which is the playing out
of individuals’ relationship attitudes and
behaviors in their daily lives. Therefore,
an individual’s relationship attitudes are
both contained in and highly correlated
with narratives.
Failure to acknowledge difficult
life events leads to a lack of narrative
complexity, well-being, and identity
development. Because negative events
are emotionally difficult for individuals
to explain within the context of the life
story, the most common response is to
deny or discount such events; this
repression can cause individuals to have
positive
illusions
about
difficult
experiences, leading to a focus more on
potential positive meanings of events
rather than acknowledging their negative
aspects (McAdams, 2008). Failing to
process and understand the negative
aspects
hinders
individuals
from
confronting the reality of these
situations. Furthermore, these illusions
will not last indefinitely, resulting
ultimately in lowered levels of wellbeing. An individual who pushes
traumatic
experiences
out
of
consciousness is more likely to have a
long-lasting obsession with trauma (i.e.,
rumination) and lack an integrated self,
identity, and coherent life story (Fivush,
Several studies have demonstrated
that positive relationship development is
linked to growth and well-being.
Individuals who experience positive
50
change after trauma alter their life
philosophy
and
enhance
their
relationships (Linley & Joseph, 2003).
Bauer, McAdams, and Sakaeda (2005a)
found that happy people (i.e., people
with connections to other people/things
beyond self) interpret and understand
their lives according to meaningful
relationships; the study also found that
having meaningful relationships is
correlated strongly with psychological
well-being. Bauer, McAdams, and Pals
(2006) discuss how high levels of
eudaimonic well-being are linked to,
among other things, a richer, more
mature
understanding
of
one’s
relationship to self and others. Therefore,
growth and happiness are linked to
positive relationship development and,
in turn, these relationships are linked to
overall positive well-being.
beliefs/values). In this study, IPV was
operationally defined to include physical,
emotional, verbal, sexual, and financial
abuse. Although extensive past research
on narrative processing has helped
structure the current study’s focus (see
studies in King & Hicks, 2006), the
rationale for this study stems from past
research gaps on IPV’s relationship to
narrative processing. IPV victims are a
particularly important population to
study because they have experienced
challenges such as emotional and
interpersonal trauma that they attempt
to fit into their life stories. These stories
provide insight into how individuals
positively
recover
from
difficult
experiences and how these traumas
affect individuals’ outlooks on life. The
dependent measures in the present
study—psychological well-being and
communion—reflect individuals’ overall
functioning levels both internally and
externally. These measures provide
information about an individual’s overall
well-being (e.g., independence, life
purpose, self-acceptance, general affect)
and
interpersonal
relationship
functioning
and
attitudes
(e.g.,
relationship fear, motivation, and
stability). Overall, I hypothesized that
higher levels of complexity, resolution,
and growth in IPV women’s narratives
would be related to enhanced well-being,
intimate relationship attitudes, and
positive relations with others.
The Present Study
The present study seeks to
contribute to narrative processing
research on Intimate Partner Violence by
empirically investigating the relationship
between processes in individuals’
narratives (i.e., growth, complexity, and
resolution) and their sense of well-being,
relationships with others, and attitudes
about
intimate
relationships.
To
accomplish
this,
the
participants
completed questionnaires about their
social
relational
engagement,
relationship attitudes, and overall
psychological
well-being.
The
experimenter
also
conducted
an
interview with the participant, which
focused on participants’ experience of
intimate partner violence within the
context of their overall life stories (e.g.,
treatment decisions, future goals,
This study also has several
subsidiary predictions about intimate
relationship motivation and perpetrator
punishment satisfaction. In the context
of
narratives,
I
examined
how
participants’ incorporation of post-abuse
intimate relationship willingness and
51
contentment
with
their
abusers’
punishments were related to the
outcome measures of well-being,
intimate relationship attitudes, and
positive relations with others. Based on
past research (e.g., see McAdams,
Hoffman, Mansfield, & Day, 1996; Bauer,
McAdams, & Sakaeda, 2005a), I predicted
that increased willingness to enter into
new intimate relationships would be
related to increased well being, positive
intimate relationship attitudes, and
positive relations with others. Secondly, I
also predicted that increased satisfaction
with perpetrator punishment would be
positively correlated with these outcome
measures. This prediction is based on
past research that has found that, within
a certain time frame of their
perpetrator’s
punishment,
victims’
revenge feelings were significantly
decreased if punishment severity were
higher (Orth, 2004).
(e.g., in this case, IPV) (Berry,
Worthington, O’Connor, Parrott, &
Wade, 2005). Therefore, I predicted that
rumination would be a mediator between
narrative
processes
and
outcome
measures of well-being and relationship;
specifically, I predicted rumination
would mediate narrative resolution’s
effect on both well-being and intimate
relationship attitudes. Past research
particularly points to rumination’s role in
the context of narrative resolution (e.g.,
forgivingness), well-being (e.g., mood
state),
and
intimate
relationship
attitudes (e.g., vengeful/negative feelings
post-abuse).
Method
Participants
The participants (N=32) for this
study were recruited from non-profit
agencies in upstate New York (N=16) and
Texas (N=16) that provide services for
victims of domestic violence. In keeping
with confidentiality concerns, therapists
and advocates at both non-profit
agencies gave their clients information
about this study and the clients then
volunteered to take part in this study.
The
experimenter
contacted
the
potential participants by telephone and
set up a time to meet in-person to run
the study session. The participants were
compensated for taking part in this study
with a $15 gift card.
I also posed the question of
whether participants’ level of rumination
about their abusive trauma is a mediator
between the effects of the narrative
processes and participants’ outcomes.
Past research has shown rumination’s
significant relationship to negative mood
(e.g., depression and anxiety) and
negative correlation with individuals’
forgiveness of themselves, others, and
situations; specifically, goal interruption
and lack of processing emotions in
individuals
potentially
causes
rumination, which, in turn, leads to
negative mood states (Segerstrom, Tsao,
Alden, & Craske, 2000; Thompson, et al.,
2005). Further research has also shown
that rumination is a mediator between a
person’s forgivingness level and revenge
motivations following transgressions
The participants in this study
ranged in age from 22 to 57 years
(M=39.2, SD=10.7). In terms of education,
15.6% had less than high school level,
18.8% had a high school/GED degree,
37.5% had completed some college, 12.5%
had an Associates degree, 12.5% had a 452
year college degree, and 3.1% had a
Master’s degree. Most of participants in
the study were either white or
black/African American (56.3% and
31.3%, respectively), and a smaller
percentage were Hispanic (9.4%) and of
mixed race (3.1%). In terms of marital
status, 6.3% of the participants were
legally married, 25.0% divorced, 31.3%
separated, 6.3% widowed, and 31.3%
single. Occupationally, 34.4% were
workers for pay, 18.8% maintained
households, 3.1% were students, 3.1%
were retired, 28.1% were unemployed,
3.1% were both students and maintained
households, and 9.4% were both
students and workers for pay. In terms of
income, 71.9% of the participants had
income less than $15,000, 22% had
income between $15,001-$55,000, and
6.2% had income over $55,000.
were casual partners, 50.0% were
spouses, and 9.4% were other types of
perpetrators. Therapy-wise, 18.8% of the
women were part of a support group,
65.6% were attending individual therapy,
15.6% had never attended therapy, and
15.6% were past recipients of therapy.
Procedures
Data were gathered over a period
of 4 months via an individual, one-time
meeting session at either an upstate New
York or Texas non-profit agency. The
hour-long study session occurred in an
enclosed room at the agency to ensure
confidentiality. After finishing a consent
form and a demographics questionnaire,
participants
completed
various
scales/questionnaires that addressed
well-being, control over healing from
trauma, relationship engagement and
attitudes, and overall functioning levels.
Participants filled out the scales and
questionnaires by hand unless they
preferred an alternative means or had a
reading impairment, in which case, the
experimenter and participant verbally
filled out the questionnaires and scales
together. Then, participants verbally
participated in an interview, which
examined their overall life reflections
and IPV experiences. The experimenter
recorded all interviews via an audiotaping device, so that trained research
assistants could later code the interviews.
Specific scales/questionnaires and the
narrative protocol are discussed in detail
below (see Table 1 in the appendix for
descriptive information).
In terms of intimate relationships,
21.9% of the participants were currently
in relationships (12.5% still with
perpetrators and 9.4% with new
partners). The mean amount of time the
abuse had been occurring, if occurring
currently, was 135.8 months (Range=3.0192.0 months, SD=89.0). If the abuse was
not current, the mean length of the
abuse was 108.1 months (Range=.75-360.0
months, SD=108.7) and the mean amount
of time that it had been since the abuse
stopped was 11.8 months (Range=.50-54.0
months, SD=16.0). Many of the
participants had experienced multiple
types of abuse, including physical
(84.4%), verbal (96.9%), emotional
(90.6%), sexual (50.0%), and financial
(65.6%). In terms of types of perpetrators
(sometimes spanning numerous abusive
relationships of differing types), 50.0% of
the perpetrators were boyfriends, 3.1%
Measures of Narrative Processing
The interview protocol used was
an adapted version of McAdams’ “Life
53
Story Interview” (McAdams, 1995;
Helson, 1992). McAdams’ original
interview is a longer protocol that is
divided into life chapters, key scenes,
future scripts, life challenges, personal
ideologies, life themes, and an overall life
reflection. This study’s adaptation was
more specific to the IPV experiences of
participants.
Specific
questions
addressed participants’ role models, IPV
experiences,
treatment/therapy
decisions, perpetrator punishment, life
beliefs, future chapters, and overall life
reflection (see Table 2 for more detail).
narrative growth (r=.80), perpetrator
punishment satisfaction (r=.95), intimate
relationship willingness (r=.93)).
Questionnaire Measures
Coding Procedure
IPV experience. Targeted questions were
used to obtain specific data about
participants’
IPV
experiences.
Participants were asked to complete
questions about the kind of IPV
experienced, whether the violence was
currently on-going, type of perpetrator
involved, and the extent of therapy that
had been sought.
Upon
conclusion
of
data
collection, the narratives were coded by
four research assistants who were all
undergraduate students. A coding
manual
was
prepared
to
score
participants on their depth of thought
and perspective taking (narrative
complexity), ability to make peace with
their abuse (narrative resolution), and
redemption from
their traumatic
experience (narrative growth) (see
Appendix A). The narratives were also
scored on participants’ satisfaction with
punishment of their abusers and also
motivation to engage in new intimate
relationships. The coders listened to the
interviews and scored each interview
according to the manual’s parameters.
The two coders assigned to each
interview rated interviews independently
first and then verbally agreed upon an
overall score for each narrative measure,
ranging from a score of 1-5 for each item.
There was a high level of inter-rater
reliability between scorers for all
narrative measures (narrative complexity
(r=.76), narrative resolution (r=.77),
Well-being. To obtain an overall outcome
measure of well-being, Ryff et al.’s (1994)
Scales of Psychological Well-Being and
the Hopkins Symptom Checklist-10 (Syed,
et al., 2008) were combined. Ryff’s
original 84-item scale (14-items per
subscale) was shortened to serve as this
study’s 9-item scale (α=.85). Three
subscales with 3 items each were used
from the PWB scale: Autonomy, which
measures independence (e.g., “Being
happy with myself is more important to
me than having others approve of me”);
Purpose in Life, which measures life
direction and belief system (e.g., “I have
a sense of direction and purpose in my
life”); and Self-Acceptance, which
measures overall self-attitude (e.g.,
“When I look at the story of my life, I am
pleased with how things have turned
out”). Participants answered each item
on a scale from 1 (strongly disagree) to 6
(strongly agree). Secondly, the Hopkins
Symptom Checklist-10 contains 10 items
that pertain to feelings of fear, weakness,
tension, fatigue, worthlessness, and
hopelessness that have occurred in the
54
previous week (α=.92). Participants filled
out each item on a scale from 1 (not at
all) to 4 (extremely). The symptoms
checklist was scored in the opposite
direction to reflect an absence of
symptoms. In order to combine the two
scales, standardized values (z-scores)
were obtained for each score, and the
participants’ scores on the two scales
were averaged (α=.81 for these two
scales).
Positive relations with others. The
Positive Relations with Others subscale
from Ryff et al.’s (1994) Scales of
Psychological Well-Being (α=.73) was
used
to
ascertain
participants’
experiences of engagement, empathy,
and intimacy in current relationships.
This subscale consisted of 8 items on a
scale from 1 (strongly disagree) to 6
(strongly agree), including “Maintaining
close relationships has been difficult and
frustrating for me” (reversed) and “Most
people see me as loving and
affectionate”.
Intimate relationship attitudes. The
outcome measure of relationship
attitudes was assessed via an adapted
version
of
the
Multidimensional
Relationship
Questionnaire
(Snell,
Schicke, & Arbeiter, 2002). The original
scale had 61 items with 12 subscales.
However, the 20-item questionnaire used
in this study had 3 subscales:
Relationship Esteem, which measures
one’s
confidence
in
intimate
relationships (e.g., “I am confident about
myself as an intimate partner”);
Relationship Motivation, which measures
one’s desire to be in an intimate
relationship (e.g., “I strive to keep myself
involved in an intimate relationship”
(reversed)); and Relationship Fear, which
measures one’s aversion to becoming
involved in intimate relationships (e.g., “I
sometimes have a fear of intimate
relationships”). Participants responded
to each item on a scale from 1 (not at all
characteristic of me) to 5 (very
characteristic of me). The Relationship
Fear scale was reversed so that high
scores indicated a relative absence of fear
in relationships. To compute an overall
score for intimate relationship attitudes,
the scores on the 3 subscales were
averaged (α=.81).
Rumination. To assess levels of
rumination about IPV experiences, the
rumination subscale from the Event
Management Questionnaire (Mansfield,
2009) was used (α=.90). The scale had 4
items, including statements such as “I
often find myself thinking of this event”
and “When I think about this event I feel
sad”. Participants responded to items on
a 1 (completely disagree) to 5 (completely
agree) scale.
Results
The present study assessed
whether the narrative processes of
resolution, complexity, and growth
(additionally, relationship willingness
and punishment satisfaction) were
predictive
of
key
outcomes:
psychological
well-being,
intimate
relationship attitudes, and positive
relations with others. To assess these
relationships, a rumination measure was
also used to determine its mediating
effect on outcomes. Descriptive statistics
for all measures are included in Table 1.
55
Relationships
between
Processes and Outcome
Narratives and Well-being
predicts
participants’
relationship
attitudes and relations, I obtained
correlations between narrative processes
and participants’ intimate relationship
attitudes and positive relations with
others (see Table 3). I found that
narrative
resolution
moderately
correlated with participants’ intimate
relationship attitudes; interestingly,
however, none of the other narrative
processes (including the questions on
satisfaction with perpetrator punishment
and
motivation
to
engage
in
intimate/romantic relationships) had a
significant
relationship
with
this
outcome
measure.
Yet,
narrative
complexity,
resolution
(moderately
strong), and growth were strongly
correlated with the positive relations
with others outcome measure. Victims’
level
of
satisfaction
with
their
perpetrator’s
punishments
and
enthusiasm to engage in intimate
relationships
were
also
strongly
correlated with participants’ report of
their level of positive relations with
others. These results suggest that the
ability of victims to rise above and
understand their abuse is linked to their
overall sentiments towards intimate
relationships (e.g., fear of relationships,
motivation to engage in relationships).
Furthermore, victims’ ability to make
peace with their abuse, grow from their
trauma, and understand multiple
perspectives about their experiences
predicts their general experiences with
relationships. Participants’ contentment
with their perpetrators’ punishments and
engagement in intimate relationships
also relates to the overall positivity of
their relationships with others.
Narrative
Measures
In seeking to answer the central
question of whether narrative processing
predicts participants’ self-reported wellbeing, I obtained the correlations
between these two sets of measures (see
Table 3). I found strong correlations
between
narrative
complexity,
resolution, and growth and participants’
well-being. Additionally, there was a
moderate correlation between well-being
and participants’ overall willingness and
enthusiasm to engage in romantic
relationships
post-IPV.
However,
surprisingly the narrative measure of
punishment
satisfaction
did
not
significantly correlate with well-being.
These findings suggest that participants’
depth of reflection, multiple perspective
taking, optimism, and ability to make
peace with their intimate partner
violence experiences is related to higher
levels of overall emotional and
psychological well-being. Furthermore,
these results suggest a victim’s
motivation to engage in an intimate
relationship is positively linked to wellbeing and is an indicator of a person’s
overall functioning level. Unexpectedly,
these results also suggest that a victim’s
level of satisfaction with the punishment
of her perpetrator does not significantly
relate to her overall psychological wellbeing and livelihood.
Narratives and Relationships: Intimate
Relationship Attitudes and Positive
Relations with Others
In seeking to answer the central
question of whether narrative processing
56
growth
and
resolution.
Thus,
participants’ depth of thought about
their abusive experiences and ability to
take
multiple
perspectives
when
contemplating such experiences are very
important in predicting their quality of
interpersonal relationships.
Regression Analyses Examining Multiple
Aspects of Narrative Processing on
Participants’ Outcomes
To probe the relationship between
narratives and well-being, I conducted a
linear regression analysis to look at
narrative
processing
variables
simultaneously as predictors of wellbeing. Taking all the three variables of
narrative complexity, resolution, and
growth together, narrative resolution
most strongly predicted well-being (see
Table 4). This finding suggests that the
ability of intimate partner violence
victims to make peace with their
situations is a more important predictor
of
their
overall
emotional
and
psychological functioning than narrative
growth and complexity.
Regression
Analyses
Examining
Rumination’s Role as a Mediator between
Narrative Resolution and Outcome
Measures
I examined whether rumination
accounts for the relationships that might
emerge between narrative resolution and
well-being and between narrative
resolution and intimate relationship
attitudes. I conducted two separate
regression analyses predicting intimate
relationship attitudes and well-being. In
both cases, narrative resolution was
entered as a predictor in Step 1 and
narrative resolution and rumination were
entered simultaneously as predictors in
Step 2. This protocol tested whether
narrative resolution still had a significant
effect on well-being or intimate
relationship
attitudes
when
also
considering the relationship between
rumination and those variables.
Additionally, I conducted a linear
regression analysis to ascertain the
relationship between narrative processes
(i.e., narrative complexity, resolution,
and growth) as simultaneously predictors
of intimate relationship attitudes (see
Table 5). I found that narrative
resolution was the most important
predictor of these attitudes, suggesting
victims’ abilities to rise above their
traumatic abuse experiences is more
predictive of their outlooks on romantic
relationships than narrative complexity
and growth.
Following standard procedures for
examining possible mediators, I first
established narrative resolution’s strong
correlation with rumination (r=-.50,
p<.01), and, in turn, rumination’s strong
correlation with well-being (r=-.69,
p<.01). I then conducted a mediator
analysis in which narrative resolution
significantly predicted well-being scores
(=.59, p=.00); however, when I entered
rumination in the second step, I found
both resolution (=.33, p=.03) and
Lastly, I conducted a linear
regression analysis to determine the
relationship between all three narrative
processes and participants’ positive
relations with others (see Table 6).
Surprisingly, narrative complexity was
the most predictive factor of this
outcome measure compared to narrative
57
rumination (=-.52, p=.001) to be
significant predictors of well-being, with
resolution’s significance now being less
robust. Therefore, I found rumination to
be a partial mediator of the link between
narrative resolution and well-being. This
finding suggests that resolution has an
impact on well-being partly because of
its effect on rumination (e.g., as narrative
resolution
increases,
participant’s
rumination about IPV decreases, and
well-being increases).
hypothesized that higher levels of
narrative complexity, resolution, and
growth in the women’s narratives would
predict higher levels of well-being and
communion (i.e., more positive intimate
relationship attitudes and relations with
others). I also hypothesized that
perpetrator punishment satisfaction and
romantic/intimate
relationship
willingness (both included in the
narratives) would also be predictive of
more positive outcomes. Lastly, I
hypothesized that rumination would
mediate the effects of narrative processes
on outcome measures; particularly, I
predicted that rumination would be a
mediator between the effect of narrative
resolution on both well-being and
intimate relationship attitudes. My
results
generally
supported
my
hypotheses as most narrative processes
were related to my outcome measures
and, furthermore, rumination was a
partial mediator between narrative
resolution and well-being.
Additionally, I established that
rumination was moderately correlated
with participants’ intimate relationship
attitudes (r=-.44, p<.05). I then
conducted a mediator analysis in which
narrative
resolution
significantly
predicted intimate relationship attitude
scores (=.42, p=.02). However, when I
entered rumination into the model as a
predictor at Step 2, I found that
rumination was not a significant
predictor of intimate relationship
attitudes (=-.30, p=.12). Therefore, this
second mediator analysis showed
rumination to not be a mediator when
looking at the relationship between
narrative resolution and intimate
relationship attitudes. This finding
suggests that narrative resolution’s
relationship with intimate relationship
attitudes is not dependent on its effect
on rumination.
Well-being
Generally, as predicted, most of
the narrative measures were moderately
to strongly predictive of participants’
well-being. Specifically, I found that
narrative complexity, resolution, and
growth were all strongly related to
victims’ well-being levels. Participants’
willingness to engage in romantic,
intimate relationships after their abusive
experiences (as denoted in their
narratives) was moderately correlated
with their levels of well-being. However,
the narrative measure of satisfaction with
the perpetrators’ punishment did not
significantly correlate with participants’
well-being levels.
Discussion
The present study sought to
examine whether different aspects of
narratives were predictive of intimate
partner violence victims’ well-being,
positive relationships with others, and
intimate relationship attitudes. Overall, I
58
Narrative complexity’s strong
relationship with well-being suggests
that a victim’s ability to reflect deeply on
her abuse, integrate the abuse and selfview into an overall life story, and to see
other individuals’ perspectives (e.g.,
children’s, perpetrator’s, family’s) is
linked to higher levels of overall
psychological well-being. This finding is
supported by past research that has
found that individuals who push
traumatic
experiences
out
of
consciousness are more likely to have a
long-lasting obsession with trauma and
lack an integrated self, resulting in
lowered levels of well-being; however,
the integration of negative and positive
emotions in narratives is indicative of
narrative complexity and consequent
overall well-being (Pennebaker, Mayne,
&
Francis,
1997;
Fivush,
2004;
Pennebaker & Seagal, 1999).
adversity (versus regret and rumination
on the negative) is related to heightened
self understanding and positive life
philosophy, which allow for healing to
occur (see also McAdams, Reynolds,
Lewis, Patten, & Bowman, 2001).
The
moderate
correlation
between
participants’
intimate
relationship willingness and well-being
suggests that intimate partner violence
victims’ eagerness or openness to engage
in an intimate relationship post-abuse is
at least modestly related to their overall
psychological functioning and welfare.
This finding is related to past research
that has shown the positive relationship
of happiness and well-being to
meaningful relationship engagement as
well as post-trauma growth’s link to
relationship improvement (Linley &
Joseph, 2003; Bauer, McAdams, &
Sakaeda, 2005a).
The strong link between narrative
resolution and well-being suggests that a
victim’s ability to make peace with the
event by viewing her present situation or
state of mind in a positive light is related
to higher levels of overall psychological
functioning. For example, ‘coherent
positive resolution’, which occurs when a
well integrated life story of a difficult
experience ends positively, is likely to
lead to higher resiliency and well-being
levels (Pals, 2006).
Contrary to my predictions,
participants’ satisfaction with the
punishment of their perpetrators was did
not significantly correlate with their
levels of well-being, although a moderate
correlation did exist between these two
measures. This finding may be due to the
fact that when victims have higher levels
of other narrative processes (i.e.,
complexity, resolution, and growth), they
may care less about their perpetrator
having a more intense punishment. For
example, if a victim has made peace with
her experiences of relationship violence,
she may not be as emotionally invested
in the intensity of her abuser’s
punishment than someone who still is
quite vengeful about her intimate
partner violence (see Orth, 2004).
However, a moderate relationship does
Narrative
growth’s
large
correlation with participants’ well-being
levels indicates that victims’ ability to
find redemption in their traumas and to
focus on a positive, goal-oriented future
is linked to their overall sense of wellbeing. Therefore, an intimate partner
violence victim’s perception of benefit in
59
exist, so there is some link between
increased well-being and increased
punishment satisfaction. Importantly,
the correlation probably did not reach
statistical significance because of the
small study sample size, but such a link
is likely to exist.
and participants’ well-being. In general
support of my predictions, I found that
rumination was a partial mediator
between narrative resolution and wellbeing. This finding suggests that
narrative resolution’s effect on well-being
is also due to its effect on rumination;
therefore, narrative resolution and
rumination interact, resulting in a
victim’s well-being level. For example,
higher levels of narrative resolution lead
to lower levels of rumination, which, in
turn, are predictive of higher levels of
well-being in intimate partner violence
victims. Past research supports this
finding as individuals who repress
trauma (i.e., fail to resolve it) have a
higher likelihood of obsessing over the
trauma and failing to incorporate the
trauma into their overall life stories
(Fivush, 2004; Pennebaker & Seagal,
1999).
Given that these
narrative
processes were all significantly correlated
with well-being, it was important to
examine which process was the most
important predictor of well-being.
Therefore, I also conducted a linear
regression analysis between narrative
processes (i.e., narrative complexity,
resolution, and growth) and well-being,
in which I found narrative resolution to
be the most accurate predictor of
participants’ well-being levels. This result
made intuitive sense as victims’
understanding of their abuse, integration
of their abuse into a coherent life story to
avoid negative effects, and positive
perception of their current life situations
and decisions are instrumental in having
positive psychological functioning and
overall life satisfaction. However,
narrative complexity and growth are
important, too, as victims’ ability to find
redemption (narrative growth) after
considering
multiple
perspectives
(narrative complexity) when reflecting
on the abuse leads to less regret (i.e.,
higher levels of resolution) and increased
well-being (Ryff, 1989, in King, 2001).
Overall, my results suggest that
storytelling is a powerful force in an
individual’s well-being development.
Different narrative processes pinpoint
intimate partner violence victims’
overall, daily psychological health and
functioning. Within this context,
narrative resolution, or victims’ abilities
to focus on the positives in their current
situations, is the most powerful predictor
of their well-being levels.
Communion:
Intimate
Relationship
Attitudes and Positive Relations with
Others
In light of research suggesting
that rumination on negative events has
negative effects on well-being (see
Segerstrom, Tsao, Alden, & Craske, 2000;
Thompson, et al., 2005), I examined
whether rumination played a role in the
relationship between narrative resolution
Generally, narrative processes
correlated to and were predictive of both
relationship
outcome
measures.
Narrative resolution was the only
narrative process significantly related to
60
victims’ attitudes toward intimate
relationships.
However,
narrative
complexity, resolution, and growth were
all strongly correlated with participants’
positive relationships with others;
additionally, participants’ satisfaction
with
the
punishment
of
their
perpetrators and also willingness to
engage in intimate relationships were
strongly correlated with their general
relationships with others. Rumination
mediator analyses were also telling, as
narrative resolution was also the most
accurate
predictor
of
intimate
relationship attitudes.
example, even though past research has
found positive change post-trauma to
relate to relationship enhancement (see
Linley & Joseph, 2003), this does not
mean attitudes of victims not currently
in relationships about potential romantic
relationship involvement will be any
more positive per se. Secondly, past
research has supported a link between
well-being and a trauma victim’s ability
to understand multiple perspectives in
relationships (Bauer, McAdams, & Pals,
2006); thus, a victim’s overall sense of
well-being (e.g., hope, lack of fear in
daily life) may relate to a victim’s specific
attitudes towards intimate relationships.
In fact, in this study, although well-being
and intimate relationship attitudes were
not significantly correlated (r=.35, p>.05)
probably due a small sample size, a
substantial relationship does exist
Intimate Relationship Attitudes
In terms of the effects of narrative
processes on IPV victims’ intimate
relationship
attitudes,
narrative
resolution was the only narrative process
that significantly correlated with these
attitudes. This result suggests that
victims’ level of acceptance about their
abuse is predictive of their intimate
relationship fear, confidence, and
motivation. For example, if a victim has
resolved the abuse trauma to a greater
extent, she will be more confident and
willing to be an intimate partner.
Lastly, contrary to my predictions,
victims’ levels of contentment with their
perpetrators’ punishments and their
intimate relationship willingness were
not significantly correlated with their
intimate relationship attitudes. Even if a
victim is content with her previous
perpetrator’s punishment level, this does
not necessarily mean that she lacks fear
or has confidence as an intimate partner;
her satisfaction with the perpetrator’s
punishment is just one sentiment
pertaining to her abusive experience and
does not necessarily represent how she
feels
overall
(e.g.,
punishment
satisfaction
was
not
significantly
correlated with well-being (r=.34, p>.05);
however, a substantial relationship does
exist, which could potentially be
significant in larger studies).
However, although I predicted
that higher levels of narrative complexity
and growth would also be related to
more positive relationship attitudes,
these correlations were surprisingly not
significant. A possible explanation for the
lack of a strong relationship between
narrative growth and such relationship
attitudes may have to do with the
difference
between
individuals’
sentiments towards relationships and
actual engagement in relationships; for
61
I also conducted a linear
regression analysis to ascertain which
narrative process was the most accurate
predictor of participants’ intimate
relationship attitudes. I found that
narrative resolution was the most
predictive of participants’ attitudes
concerning romantic involvement with
potential partners. This result makes
intuitive sense as victims’ abilities to
make peace with and overcome their
past abusive experiences should be
foretelling of how open they are to
engage in a new intimate relationship.
Such a strong relationship between these
measures suggests that a victim’s level of
acceptance of the abuse, redemption,
and integration of multiple emotions and
perspectives into her narrative are
indicative of her general trust, comfort,
and intimacy level in relationships. These
results relate to past research that has
shown
post-trauma
growth
and
happiness to be linked to relationship
development, understanding, and wellbeing (Linley & Joseph, 2003; Bauer,
McAdams, & Sakaeda, 2005a).
Secondly, I also found that
punishment satisfaction and intimate
relationship willingness were strongly
correlated with the outcome measure of
positive relations with others. These
findings are interesting in that, for
example, a victim’s contentment with her
perpetrator’s level of punishment is then
indicative of her overall contentment in
her general relationships with other
individuals. This may suggest that a
victim’s feelings of validation and that
justice has been achieved relate to her
ability to then gain satisfaction,
validation, and trust in her general
interpersonal relationships. Furthermore,
the strong correlation between intimate
relationship willingness and positive
relations with others suggests that
victims’ motivation to engage in an
intimate relationship is very indicative of
their overall enthusiasm and comfort in
their relationships besides purely
intimate ones. This is in conjunction
with past research that has found that
communion themes in narratives (e.g.,
love) are positively related to the
development of and need for warm, close
relationships
(McAdams,
Hoffman,
Mansfield, & Day, 1996).
Lastly, I conducted a regression
analysis to see if rumination was a
mediator between narrative resolution
and intimate relationship attitudes, in
which I found rumination to not be a
mediator between these two measures.
Surprisingly, rumination was not even
significantly predictive of intimate
relationship attitudes when I conducted
the mediator analyses, indicating that
narrative resolution’s effect on these
relationship attitudes is purely due to a
participant’s ability to make peace with
her trauma, and not her level of
rumination about the trauma. This result
is important as it solidifies the
importance of narrative resolution’s
predictability and relationship to
outcome
measures,
particularly
resolution’s role in more positive
intimate relationship attitudes.
Positive Relations with Others
In support of my predictions, all
narrative processes (i.e., narrative
complexity, resolution, and growth) were
strongly correlated with the position
relations with others outcome measure.
62
In conducting a linear regression
analysis to see which narrative process
was the most indicative of the positive
relations with others outcome measure, I
found narrative complexity to be the
most accurate predictor of this outcome
measure. This result was unexpected as I
thought narrative resolution or growth
would be more predictive of individuals’
general relationship engagement than
narrative complexity. However, this
finding can be explained by past research
that has shown complexity to buffer the
effects
of
resolution,
indicating
complexity’s pertinent role as a narrative
processing tool in lieu of the other
processes (see Mansfield et al., 2009).
Therefore, the ability of individuals to
see
multiple
perspectives
(e.g.,
perpetrator’s or children’s perspective)
and to integrate numerous emotions
about their trauma into a coherent life
story is predictive of how healthy their
general relationships are and how they
conduct
themselves
in
these
relationships.
This study also has important
implications for the conduct of therapy
with intimate partner violence victims.
Therapists
should
emphasize
storytelling, particularly how different
negative experiences fit into an overall
life narrative. This conceptualization of
events is significant as it helps prevent
traumas from overtaking life narratives,
and, instead, helps individuals realize
that their abusive experiences do not
encompass their entire identity.
Significance of the Study
Limitations and Future Directions
The present study’s focus on
narrative processing in victims of
intimate partner violence is an important
addition to the literature on coping with
trauma. Although past research has
examined the role of narrative processing
in various types of trauma victims,
research to date lacks in its examination
of the relationship between narrative
processing, well-being, and capacity for
relationships in this specific type of IPV
trauma. Finding numerous significant
relationships
between
narrative
processes, rumination, and outcome
measures is significant in further
One of the main limitations in my
study is the issue of causality. Although I
conducted linear regression analyses
examining which narrative processes
were the most predictive of outcome
measures, I did not consider whether
such outcome measures of well-being
and capacities for relationships were
predictive of individuals’ narrative
processing. For example, a victim’s level
of well-being could influence how she
tells her story; higher levels of well-being
could lead to enhanced resolution or
growth in one’s story, whereas lower
well-being levels could predict less
showing the potent importance of
storytelling, especially since intimate
partner violence is culturally silenced to
a great extent in our society (see Fivush,
2004). Individuals vary in their
integration of past experiences into their
life stories, and, in effect, their abilities to
recover
from
traumas
such
as
relationship violence; however, as my
research indicates, different ways of
processing such traumatic experiences
are linked to and predictive of different
outcomes.
63
emotional processing of traumas in one’s
life story. Future research should address
these limitations by not only looking at
the effect of narrative processing on wellbeing and other proposed outcome
measures, but also seeing if such
outcome measures are indeed predictive
of proposed predictor variables (see
McLean, Pasupathi, & Pals, 2007).
However, the most effective way to
address this limitation would be for
researchers to conduct a longitudinal
study in which they assessed narratives
at time 1 and then assessed changes in
well-being from time 1 to time 2.
genders in the way respective genders
narrate and interpret events, and how
such processing is related to well-being
levels. Conducting the study with a larger
sample from more diverse venues would
also be interesting; although I conducted
my study at two advocacy non-profits for
relationship violence victims and
obtained a quite diverse sample, it would
also be helpful to obtain participants
from college campuses to examine
differences in victim experiences in a
close university setting versus in the
larger community. Also, it would be
interesting to study the effects of therapy
length and age on intimate partner
violence victims’ stories and overall wellbeing. Some past research has shown, for
example, the beneficial effects of therapy
on individuals’ storytelling abilities,
understanding
of
traumas,
and
psychological well-being (Pals, 2006).
Furthermore, past research has found a
positive relationship between age, life
story development, understanding of
traumas, and overall self-view (Pasupathi
& Mansour, 2006; McAdams et al., 2006).
These future research endeavors would
give us a better understanding of other
forces that are important in shaping
victims’ narratives, particularly how
individual differences emerge and how
such individuals cope with traumas in
their
lives.
Another main limitation in my
study is its small sample size and the fact
that all of the participants were females.
The small sample size prevented my
results from being more significant and
also prevented me from using variables I
initially wanted to include in my study.
The all female sample is also limiting, as
past research indicates that women are
more likely to view themselves as victims
than are men (Sanderson & McKeough,
2005), so my study may not be
generalized across genders.
Future research should address
these limitations by duplicating this
study with a larger, more diverse sample.
This way, researchers could see whether
there are marked differences across
Bauer, J. J., McAdams, D. P., & Sakaeda,
A. R. (2005b). Crystallization of desire
and crystallization of discontent in
narratives of life-changing decisions.
Journal of Personality, 73, 1181-1213.
Works Cited
Bauer, J. J., McAdams, D. P., & Sakaeda,
A. R. (2005a). Interpreting the good life:
Growth memories in the lives of mature,
happy people. Journal of Personality and
Social Psychology, 88, 203-217.
64
King, L. (2001). The hard road to a good
life: The happy, mature person. Journal of
Humanistic Psychology, 41, 51-73.
Bauer, J. J., McAdams, D. P., & Pals, J. L.
(2006).
Narrative
identity
and
eudaimonic well-being. Journal of
Happiness Studies, 9, 81-104.
King, L. A., & Hicks, J. A. (2006).
Narrating the self in the past and the
future: Implications for maturity.
Research in Human Development, 3, 121138.
Berry, J. W., Worthington, E. L.,
O’Connor, L. E., Parrott, L., & Wade, N.
G. (2005). Forgivingness, vengeful
rumination, and affective traits. Journal
of Personality, 73, 183-225.
Linley, P. A., & Joseph, S. (2003). Trauma
and personal growth. Psychologist, 16,
135.
Fivush, R. (2004). The silenced self:
Constructing self from memories spoken
and unspoken. In D. R. Beike, J. M.
Lampinen, & D. A. Behrend (Eds.), The
self and memory: Studies in self and
identity (pp. 75-93). New York:
Psychology.
Mansfield, C. D. (2009). Storying the
transgressive self: Narrative processing as
a mediator of implicit self-theory and
proneness
to
shame
and
guilt.
Unpublished master’s thesis, Western
Washington University.
Frazier, P.A. (2003). Perceived control
and distress following sexual assault: A
longitudinal test of a new model. Journal
of Personality and Social Psychology, 84,
1257-1269.
Mansfield, C. D., McLean, K. C., &
Lilgendahl, J. P. (2009). Does narrative
processing matter and for whom?: Links
between narrative processing, wisdom,
and well-being in narratives of traumas
and
transgressions.
Unpublished
manuscript,
Western
Washington
University and Haverford College.
Frazier, P. A., Mortensen, H., & Steward,
J. (2005). Coping strategies as mediators
of the relations among perceived control
and distress in sexual assault survivors.
Journal of Counseling Psychology, 52, 267278.
Mansupathi, M., & Mansour, E. (2006).
Adult age differences in autobiographical
reasoning in narratives. Developmental
Psychology, 42, 798-808.
Frazier, P., Tashiro, T., Berman, M.,
Steger, M., & Long, J. (2004). Correlates
of levels and patterns of positive life
changes following sexual assault. Journal
of Counseling and Clinical Psychology, 72,
19-30.
McAdams, D. P. (1995). The life story
interview. Retrieved September 10, 2009,
from Northwestern University, School of
Education and Social Policy’s Web site:
http://www.sesp.northwestern.edu/foley
/instruments/interview/
Helson, R. (1992). Women’s difficult
times and the rewriting of the life story.
Psychology of Women Quarterly, 16, 331347.
McAdams, D. P. (2001). Psychology of life
stories. Review of General Psychology, 5,
100-122.
65
McAdams, D. P. (2008). Personal
narratives and the life story. In O. John,
R. Robins, and L. Pervin (Eds.),
Handbook of personality: Theory and
research (3rd Ed., pp. 241-261). New York:
Guilford Press.
McLean, K. C., Pasupathi, M., & Pals, J. L.
(2007). Selves creating stories creating
selves: A process model of narrative self
development
in
adolescence
and
adulthood. Personality and Social
Psychology Review, 11, 262-278.
McAdams, D. P., & Bowman, P. J. (2001).
Narrative
life’s
turning
points:
Redemption and contamination. In D. P.
McAdams, R. Josselson, and A. Lieblich
(Eds.), Turns in the Road: Narrative
Studies of Lives in Transition (1st Ed., pp.
3-34). Washington, D.C.: American
Psychological Association.
Orth, U. (2004). Does perpetrator
punishment satisfy victims’ feelings of
revenge? Aggressive Behavior, 30, 62-70.
Pals, J. L. (2006). Narrative identity:
Processing of difficult life experiences:
Pathways of personality development
and positive self-transformation in
adulthood. Journal of Personality, 74,
1079-1109. .
McAdams, D. P., Hoffman, B. J.,
Mansfield, E. D., & Day, R. (1996).
Themes of agency and communion in
significant
autobiographical
scenes.
Journal of Personality, 64, 339-377.
Pennebaker, J. W., & Seagal, J. D. (1999).
Forming a story: The health benefits of
narrative. Journal of Clinical Psychology,
55, 1243-1254.
McAdams, D. P., Reynolds, J., Lewis, M.,
Patten, A. H., & Bowman, P. J. (2001).
When bad things turn good and good
things
turn
bad:
Sequences
of
redemption and contamination in life
narrative
and
their
relation
to
psychological adaptation in midlife
adults and in students. Personality and
Social Psychology Bulletin, 27, 474-485.
Pennebaker, J. W., Mayne, T. J., &
Francis, M. E. (1997). Linguistic
predictors of adaptive bereavement.
Journal of Personality and Social
Psychology, 72, 863-871.
Ryff, C. D., Lee, Y. H., Essex, M. J., &
Schmutte, P. S. (1994). My children and
me: Midlife evaluations of grown
children and of self. Psychology and
Aging, 9, 195-205.
McAdams, D. P., Bauer, J. J., Sakaeda, A.
R., Anyidoho, N. A., Machado, M. A.,
Magrino-Failla, K., et al. (2006).
Continuity and change in the life story: A
longitudinal study of autobiographical
memories in emerging adulthood.
Journal of Personality, 74, 1371-1400.
Sanderson, A., & McKeough, A. (2005). A
narrative analysis of behaviourally
troubled
adolescents’
life
stories.
Narrative Inquiry, 15, 127-160.
Segerstrom, S. C., Tsao, J. C. I., Alden, L.
E., & Craske, M. G. (2000). Worry and
rumination: Repetitive thought as a
concomitant and predictor of negative
mood. Cognitive Therapy and Research,
24, 671-688.
McLean, K. C. (2008). The emergence of
narrative identity. Social and Personality
Psychology Compass, 2, 1685-1702.
66
Singer, J. A. (2005). Personality and
psychotherapy: Treating the whole
person. New York: Guilford Press
checklist (HSCL-10) and Pakistan anxiety
and depression questionnaire (PADQ), in
a rural self-motivated population in
Pakistan. BMC Psychiatry, 8.
Snell, W. E., Jr., Schicke, M., & Arbeiter,
T.
(2002).
The
Multidimensional
Relationship
Questionnaire:
Psychological dispositions associated
with intimate relations. In W. E. Snell, Jr.
(Ed.), New Directions in The Psychology
of Intimate Relations: Research and
Theory. Cape Girardeau, MO: Snell
Publications.
Thompson, L. Y., Snyder, C. R., Hoffman,
L., Michael, S. T., Rasmussen, H. N.,
Billings, L S., Roberts, D. E. (2005).
Dispositional forgiveness of self, others,
and situations. Journal of Personality, 73,
313-359.
Weissman, M. M., & Bothwell, S. (1976).
Assessment of social adjustment by
patient self-report. Archives of General
Psychiatry,
33,
1111-1115.
Syed, H. R., Zachrisson, H. D., Dalgard,
O. S., Dalen, I., & Ahlberg, N. (2008).
Concordance between Hopkins symptom
Appendix
Table 1
Descriptive Statistics for Narrative Processes, Well-being, and Relationship Outcome
Measures
M
SD
Range
Scale
Narrative Complexity
3.22
1.21
1.00-5.00
1-5
Narrative Resolution
3.47
1.19
2.00-5.00
1-5
Narrative Processes
67
Narrative Growth
3.25
1.24
1.00-5.00
1-5
2.96
1.07
1.00-5.00
1-5
Intimate Relationship Willingness 3.03
1.23
1.00-5.00
1-5
Additional Narrative Components
Punishment Satisfaction
Well-being
Symptoms Checklist
2.73
0.83
1.20-3.90
1-4
Psychological Well-being
4.09
1.14
2.00-5.89
1-6
Positive Relations with Others
4.22
0.86
1.75-5.75
1-6
Intimate Relationship Attitudes
2.79
1.01
1.00-4.87
1-5
Relationship Measures
68
Table 2
Interview Protocol Sections and Specific Questions
Interview section
I. Role model
Question types
How this individual is a role model
II. Intimate Partner Violence Description Basic information about IPV
Effect on values, self-concept, and world view
How feelings about IPV have changed over time
III. Treatment
Path to treatment decision
Path to escaping abusive relationship
Effect on life overall and view of self
IV. Punishment of Perpetrator
Has perpetrator been punished
If punishment led to satisfaction/relief in victim
V. Life beliefs
Fundamental beliefs and values about life/world
Overall philosophy of life
VI. Future Chapters
Next life chapter
Ability to engage
relationships
69
in
future
intimate
VII. Reflection
Thoughts and feelings during interview
Additional comments about interview process
______________________________________________________________________________
IPV=Intimate Partner Violence
70
.46**
2. Narrative Resolution
.48**
.41**
4. Punishment Satisfaction
5. Relationship Willingness
.58**
-
1. Narrative Complexity
3. Narrative Growth
1
Variables
.47**
.42**
.64**
-
.46**
2
71
.67**
.36
-
.64**
.58**
3
Correlations: Narrative Processes, Well-being, and Relationships
Table 3
.38*
-
.36
.42*
.48**
4
-
.38*
.67**
.47**
.41**
5
-
-
-
-
-
6
-
-
-
-
-
7
-
-
-
-
-
8
-
-
-
-
-
9
.00
.58**
7. Intimate Relationship Attitudes
8. Positive Relations with Others
p<.05*, p<.01**
9. Rumination
-
.52**
6. Well-being
-.31
.49**
.42**
.59**
72
-.50**
.53**
.16
.51**
-.58**
.58**
.13
.34
-.44*
.55**
.34
.43*
-.64**
.51**
.35
-
-.69**
.06
-
-
-.44*
-
-
-
-.47**
-
-
-
Table 4
Summary of Linear Regression Analysis for Narrative Processes and Well-being (N=32)
Variable
B
SE B

Narrative Growth
.06
.15
.07
Narrative Resolution
.32
.14
.41*
Narrative Complexity
.29
.22
.13
______________________________________________________________________________
Note: R2 =. 43, p<.05*
Table 5
Summary of Linear Regression Analysis for Narrative Processes and Intimate Relationship
Attitudes (N=32)
Variable
SE B
B
73

Narrative Growth
-.06
.19
-.07
Narrative Resolution
.48
.19
.57*
Narrative Complexity
-.22
-.18
.17
______________________________________________________________________________
Note: R2 =. 23, p<.05*
Table 6
Summary of Linear Regression Analysis for Narrative Processes and Positive Relations with
Others (N=32)
Variable
B
SE B

Narrative Growth
.13
.14
.18
Narrative Resolution
.14
.14
.19
Narrative Complexity
.38*
.27
.13
______________________________________________________________________________
Note: R2 =. 41, p<.05*
Appendix A
74
NARRATIVE CODING MANUAL
Complexity
Narrative complexity refers to the depth of thought in narratives and individuals’ abilities
to see different perspectives and outcomes (Mansfield, 2009), and is reflected in plots,
themes, characters, and settings that are intertwined into an overall life narrative (Singer,
2005).
Scoring: Listed below are different sub-scales to consider while deciding on an overall score
for complexity. A participant will be given one overall score for complexity, ranging from a
1-5.
•
Multiple Perspective Taking
a. Seeing/experiencing the IPV experience from someone else’s point of view
b. Reporting how event has affected other people besides oneself
c. Examples:
i. Seeing experience from children’s point of view (“The abuse was very
difficult for my children.”; “I had to get out of the relationship for my
children’s sake.”)
ii. Seeing experience from perpetrator’s point of view (“I don’t think he
realized what the impact of his actions was/is.”; “It seemed like he
felt like it was my fault only.”)
iii. Seeing friend/parent perspective
•
Emotion
a. Participant identified various emotions (at least two different, contrasting
emotions)
b. The coder must identify the first emotion and then compare it to other
emotions, thereby ensuring that at least 2 different emotions are present
c. Crying is not sufficient for presence of an emotion—emotions must be
explicitly stated
d. The variety of emotions applies to the narrative as a whole. The participant
may identify two emotions in the same sentence or an emotion may be
stated and then later a different emotion about the same subject/event may
be stated. It is important that there is a very clear differentiation between
75
emotions (i.e., “I was sad but now I am still sad, but to a lesser degree”
would not count as two different emotions)
e. Examples:
i. “I felt sad at the time… but now I am happy to be out of the
relationship.”
ii. “I am so angry that I let it go on for so long… I felt helpless at the
time, though. I’m pleased the abuse is over at least.”
iii. “I am glad we did the interview today…I get sad thinking about the
event, though.”
•
Analytical Reasoning
a. The participant communicates clear cause and effect by identifying causes
or effects of behavior, psychological states, self-examination, and weighing
multiple options in a situation to determine the most appropriate course of
action.
b. Examples:
i. “I wanted to change my negative outlook about my IPV, so I realized
I needed to strive to think positively and change my negative
thoughts.”
ii. “I was depressed, lonely, and downtrodden during the abuse, but I
made an effort to live a purposeful, fruitful life even though this
event had scarred me.”
•
Meaning/Implications
a. Participant has found meaning from the event, has learned from the event,
or has found the event self-transforming
b. Change in the way participant views self or behaves/acts
c. Participant may change the way she thinks about others or how she treats
others
d. Discussion of far-reaching implications of IPV event
e. Discussion about past IPV impacts participant’s present self or overall life
story
f. Examples:
i. “The event happened for a reason and it has impacted how I view
myself and those individuals surrounding me.”
ii. “I have learned that I am not alone—other women are victims of IPV
as well. I think of my life in that way and the others who suffer from
this abuse, too—I have learned we are all fighting this together.”
Complexity Scoring:
76
(1) Absence of complexity: ~0 of the above items
a. The woman sees herself as the total victim and focuses on the emotional
wreckage she personally is experiencing
b. The woman has overall one emotion present (e.g., sadness, anger)
c. The woman conveys explicitly or implicitly that she is not ready to change
her negative self outlook, behaviors, or psychological/mental state or is
unaware of the cause/effect relationship between these phenomena
d. The woman does not believe she has learned anything meaningful from the
IPV experience or that the experience impacts her or her life story
(2) Minimum complexity: ~1 of the above items
a. Woman has presence of approximately 1 of the above criteria
b. Examples:
i. Woman explicitly conveys various emotions in her narrative (i.e.,
sadness, happiness, regret, hope). General presence of only woman’s
perspective of events. Woman is still unable/not willing to see how
negative outlook impacts her well-being and does not really see
meaningfulness or the learning value of her IPV experience.
ii. Woman has gained meaning from the event and believes the event
has made her a stronger person. General presence of only woman’s
perspective of events. Woman is still generally unable/not willing to
reason how her negative outlook inhibits her from fully changing her
behavior into actions that are more productive. Overall, only one
emotion present.
c. Overall, a general lack of diverse emotions in narrative and presence of only
woman’s perspective of events. Woman still lacks reasoning in the causal
nature of her mindset and behavior and has yet to see the meaningfulness
of the event.
(3) Moderate complexity: ~2 of the above items
a. Woman has presence of approximately 2 of the above criteria
b. Examples:
i. The woman has presence of both sadness and happiness in her
narrative and sees how her previous negative outlook was causing
her sadness, which then caused her to be proactive in striving for
happiness and contentment. It is still not clear that the woman has
gained meaning or has learned from her IPV experience and she still
tells the story in a way that only implicates her.
ii. The woman conveys clear analytical reasoning as she states how she
recognized how her negative outlook was bringing her down and
then proactively tried to change her negative ways/thoughts. The
woman is also cognizant of how the IPV experience has affected her
children as she notes the emotional toll of the abuse on her children;
the woman sees her children as one reason for escaping the abusive
relationship. However, the woman has not gained positive meaning
77
from the abuse and does not clearly convey multiple, differing
emotions (e.g., the narrative is very one-dimensional—blunted
affect)
c. Overall, the woman is starting to gain meaning from the IPV experience
and shows some diversity of emotion, even though one emotion is
dominant. Woman has some reasoning and insight into the IPV experience,
but still spends a lot of the time focusing on how the event has mainly
affected her.
(4) Advanced complexity: ~3 of the above items
a. Woman has presence of approximately 3 of the above criteria
b. Examples:
i. The woman identifies various emotions in her narrative, has gained
meaning from the IPV event, and has clear reasoning as to how her
negative view on life prompted a positive behavioral/mental change
in herself. The woman’s narrative, however, is still one-dimensional,
as she tends to focus on how the IPV experience has affected her
only.
ii. The woman identifies various emotions in her narrative, has clear
reasoning as to how her negative view on life prompted a positive
behavioral/mental change in herself, and encompasses multiple
perspectives (children’s, perpetrator’s, etc.). The woman has not
gained a clear meaning from the IPV experience and does not clearly
define what she has learned from it.
c. Overall, the woman has multiple emotions present in the narrative and
most likely notes how the event has affected her children, perpetrator, etc.
The woman sees the cause and effect nature of her thought processes and
behaviors/outlook and has a general sense of having gained positive
meaning from the IPV experience.
(5) Highest complexity: ~4 of the above items
a. The woman expresses having gained a comprehensive understanding of the
event and has developed positive meaning from the event
b. The woman clearly identifies causes and effects of her thoughts and outlook
on life and wants to change such sentiments
c. Definite presence of multiple, contrasting emotions that the woman
integrates into narrative
d. Presence of multiple perspective taking (e.g., how this event affects her
children; woman feeling horrible for children; woman’s perpetrator has
done some bad things, but she doesn’t believe that makes the perpetrator a
terrible person, etc.)
78
Resolution
Narrative resolution occurs when an individual makes peace with difficult life events and
experiences within context of life story (Mansfield, 2009).
Scoring: Listed below are different sub-scales to consider while deciding on an overall score
for resolution. A participant will be given one overall score for resolution, ranging from a 15.
•
Closed Influence vs. Continuing Influence
a. Closed influence: This means that the IPV event is still influential in the
participant’s life, but the event is not affecting the participant in negative
ways (e.g., disrupting daily routine, letting event take over thought
processes)
b. Continuing influence: This means that the participant is still overly
concerned with event details and ruminates over such details and has a
negative affect due to the event
c. Examples:
i. Closed: “I think about the event still, but I don’t get sad and start
crying about it anymore.”; “I don’t let my past experience with abuse
rule my life anymore.”
ii. Continuing: Participant is audibly distraught by event; participant
goes into unnecessarily long detail about the event; “I still feel like I
can’t go places or do things that I used to like to do… I live in fear.”;
“I don’t feel like I have a new outlook on life due to the abuse…I still
feel angry, upset, and degraded.”
•
Positivity of the Solution
a. The participant describes the solution (e.g., if applicable, getting out of the
relationship, receiving treatment) in a positive manner and views it as a
‘positive course of action’
b. If no solution has been found (e.g., participant is still in abusive
relationship, participant is trying to get out of relationship), participant can
still gain credit for this section if she is feeling positively/ok at the moment
c. Examples:
i. “I am so thankful I got out of the abusive relationship and joined a
support group. I feel like I am such a strong woman.”
79
ii. “Although I am still in the relationship, I am feeling good about my
plan of action—that is, to get a separation and take care of my kids.”
•
Present Counteracts Past
a. Participant’s current situation or current events in participant’s life
counteract, or offset, the past abusive event
b. If abusive is current: a change in the participant’s outlook even though
abuse is still on-going; participant has taken initial steps to address abuse
and make the situation better
c. There is a contrast of emotions in participant’s narrative, which implies that
thoughts, feelings, etc. are different now, which further demonstrates
resolution and emotional positivity
d. Examples:
i. “That was the worst period of my life.”
ii. “I was so angry and sad about the abuse for the longest time…now
I’ve realized that everything happens for a reason and life is going to
work out all right for me.”
iii. “I am trying to get out of the relationship at the moment…hopefully I
can get a separation from my spouse…I am just glad that I joined a
support group to start the healing process.”
•
Overt statements that event is resolved/unresolved
Resolution Scoring:
(1) No resolution
a. The IPV issue is completely unresolved and the woman is still immersed in
it.
b. Overt statement of no resolution or, if overt statement of no resolution is
accompanied by presence of resolution (i.e., by meeting the criteria for one
of the other resolution items), coder must make the decision of whether to
move the participant to 2.
(2) Beginning resolution
a. The woman is in the beginning/initial stage of resolving the IPV experience.
b. The woman has begun the resolution process but still has lots more work to
do before she will have the IPV experience behind her.
c. Small amount of closure and a stronger lack of resolution.
(3) Mixed/neutral resolution
a. Woman shows moderate amount of resolution.
b. The woman may also say nothing about resolution at all—narrative neutral.
80
c. The woman seems to be making genuine progress in her resolution of the
event but still states things that make it apparent that the situation is still
not resolved.
d. The woman will appear neither more nor less resolved but in the middle.
(4) Advanced resolution
a. The woman is on her way to resolving the issue, but the issue is still not
closed.
b. The woman makes a lot of statements that suggest the situation is almost
resolved, but doubt still lingers in the coder as to how much peace or
closure the woman is actually experiencing.
c. The woman’s actions must show that the event is nearly resolved (words are
not sufficient).
(5) Complete resolution
a. Woman communicates complete/overt evidence of resolution.
b. The woman will typically see the event as part of her past (the event can
still be influential, though).
c. The woman is no longer processing the event and explicitly or implicitly
conveys acceptance of the event and that it is closed.
d. The woman can still feel sad/angry/etc. but her feelings are a normal,
negative affect that is to be expected.
e. The woman’s actions must show that the event is resolved (words are not
sufficient).
Growth
Growth is a positive change in thinking that occurs when reflecting on a difficult event,
leading to maturity, strength, and health (Mansfield, 2009).
Scoring: Listed below are different sub-scales to consider while deciding on an overall score
for growth. A participant will be given one overall score for growth, ranging from a 1-5.
•
Life Goals and Best Possible Self: Personal and Communal
a. When setting life goals, an individual seeks to become the best version of
him or herself, which is referred to as one’s ‘best possible self’ (King &
Hicks, 2006). In narratives, the amount, ease, vividness, and emotional
depth with which individuals think about their best possible selves is
pertinent to growth.
81
b. Examples:
i. “Next in my life story, I am going to get an apartment, a job, etc….I
am going to be an independent woman.”
ii. “I do envision myself having intimate relationships in the
future…maybe not right now…but I think someone is out there for
me.”
iii. “I am hopeful about the future…there is a plan for me and I know it’s
good.”
•
Redemption Sequences
a. Redemption sequences/redemption imagery occur when a negative life
experience is followed by or transformed into a positive life experience
scene.
b. Thematic categories/examples:
i. Redemption imagery
ii. Enhanced agency
iii. Enhanced communion
iv. Ultimate concerns
c. Redemption imagery: movement from demonstrably negative (can be a
negative feeling or a negative event) to demonstrably positive scene (e.g.,
positive emotion, positive cognitive state (self understanding, self insight,
etc.), and events leading to positive feelings in most people (e.g., healing,
growth, close relationships, learning, reconciliation, etc.).
i. The positive state does not need to be positive to the same degree
that the negative state is negative.
ii. If conceptualized as Aà B, A may cause B or a may just precede B in
time
1. A à B examples:
a. Sacrifice: Person endures negative A to get to a benefit
of B (Woman leaves abusive husband and is homeless
à finds support in shelter)
b. Recovery: Physical/psychological negative state is
followed by healing state (Depression à regaining of
positive outlook on life)
c. Growth:
Negative
experience
leads
to
psychological/interpersonal growth/benefit (Woman
endures and eventually ends unhappy relationship
with perpetrator à woman becomes more confident,
happier, and experiences self-understanding)
d. Learning: Negative event leads person to gain new
knowledge, wisdom, skills, etc. (Woman almost dies
from perpetrator’s extreme abuse à woman learns to
no longer fear death)
82
e. Improvement/Other: Covers areas not included in the
other categories but are still under the umbrella of bad
situation containing negative affect becoming a better
situation containing positive affect (very bad marriage
à very good relationship; lonely/depressed à
conversion to Christianity, feels ecstatic)
d. Common Redemption Examples:
i. Enhanced agency
1. Transformation from negative to positive leads to
enhancement of woman’s personal power or agency (e.g., selfconfidence, efficacy, personal resolve, or insight into personal
identity)
ii. Enhanced communion
1. Transformation from negative to positive leads to
enhancement of woman’s personal relationships of love,
friendship, family ties, etc.
iii. Ultimate concerns
1. Transformation from negative to positive contains
involvement in or confrontation with existential issues or
ultimate concerns. For example, this event brings the woman
face-to-face with death, God, or other religious/spiritual
dimensions of life.
Growth Scoring:
(1) No growth
a. Woman does not show any signs of growth from the IPV experience.
b. The negative event (i.e., IPV) has not led to any positive outcome.
(2) Minimum growth
a. The woman is in the beginning/initial stage of growing from the IPV
experience.
b. The woman has begun to show signs of growth from the negative event
but still needs to grow a lot more from the IPV experience.
c. Minimal presence of redemption sequences in narratives and minimal
positive outlook on the future.
(3) Moderate growth
a. Woman shows modest amount of growth.
b. The woman seems to have grown somewhat from the event but still
states things that make it apparent that she is still in the process of
growing and learning things from the IPV experience.
83
(4) Advanced growth
a. The woman is well on her way to fully growing from the event, but the
issue is still not closed.
b. The woman makes a lot of statements that suggest that she is in the
advanced stages of growth from the event, but doubt still lingers in the
coder as to how much growth or closure the woman is actually
experiencing.
c. Overall inclusion of redemption sequences and statements that lead the
coder to believe the woman is envisioning her best possible self.
(5) Complete growth
a. Woman communicates complete/overt evidence of growth.
b. The woman will typically see the event as part of her past self (the event
can still be influential, though) and she has grown from this event in
many ways.
c. The woman exhibits a strong commitment to her best possible self and
sees her future in a clear, positive way.
d. An overall very strong presence of redemption sequences.
Perpetrator Punishment Response
Participant’s level of satisfaction with the consequences for the perpetrator:
1
2
3
4
5
Punishment Satisfaction Scoring:
(1) No satisfaction
• The participant is not satisfied at all with her perpetrator’s punishment/lack
of punishment.
(2) Minimum satisfaction
• The participant is largely dissatisfied with perpetrator’s level of punishment
and is not happy with the result.
(3) Moderate satisfaction
• The participant has some satisfaction with her perpetrator’s degree of
punishment but the coder senses she is still not very satisfied.
84
(4) Strong satisfaction
• The participant is overall satisfied with her perpetrator’s punishment and
does not convey that she wants any large changes to his/her punishment.
(5) Complete satisfaction
• The participant is totally satisfied and content with her perpetrator’s
punishment and does not see any changes that need to be made.
Intimate Relationship Response
Participant’s level of willingness to enter into intimate relationship:
1
2
3
4
5
Intimate Relationship Willingness Scoring:
(1) No willingness
• The participant is not willing at all to engage in an intimate relationship.
(2) Minimum willingness
• The participant is largely unwilling to engage in an intimate relationship.
(3) Moderate willingness
• The participant has some willingness to engage in an intimate relationship
but the coder senses she is still not very willing.
(4) Strong willingness
• The participant is overall willing to engage in an intimate relationship.
(5) Complete willingness
The participant is totally willing to engage in an intimate relationship. The
coder senses no hesitation in participant’s willingness.
The participant is currently in a successful intimate relationship (not abusive).
•
85
The Dynamics of Doubt
Max Counter, Class of 2010
Introduction
points out, the term ‘doubt’ comes from
the Latin dubito, derived from the Old
Latin dubo, from duo or ‘two.’ MacGregor
therefore notes that doubt “is not to be
equated with belief or unbelief but rather
with a vacillation between the two
opposites: unbelief and belief” (2424). As
such, the word doubt, “although often
regarded as the opposite of belief,
signifies primarily vacillation, perplexity,
irresolution” (Geddes 2424). Therefore,
doubt is not taken as an expression of
disbelief but of uncertainty.
In the polemics surrounding
religion in contemporary society, few
voices ring more loudly than Richard
Dawkins’. An evolutionary biologist at
Oxford
University,
Dawkins
unapologetically attacks religion and
blames it for a host of societal ills. In his
best-selling book, The God Delusion,
Dawkins constructs a seven part
spectrum of religious belief. He notes
that the spectrum is continuous, but it is
encapsulated by two poles of ‘certainty’:
The first pole is “1. Strong theist. 100
percent probability of God. In the words
of C.G Jung, ‘I do not believe, I know.” At
the other end of the spectrum we find “7.
Strong atheist. ‘I know there is no God,
with the same conviction as Jung ‘knows’
there is one’” (Dawkins 73). While these
two positions lie at opposite extremes of
the spectrum, they nonetheless share the
common of element of certainty. Each
side is utterly and unquestioningly
convinced of the validity of its own
propositions.
I first attempt to detail the
dynamic potentialities of doubt within
religious life. In this respect, doubt can
play
two
equally
powerful,
yet
diametrically opposed roles: it has the
ability to be a wellspring of creativity or a
source of crippling stagnancy. Secondly,
I seek to understand the sociological
factors present by which doubt emerges
within a specific time and space,
particularly modern, ‘secular’ society. In
so doing, I seek to move the focus away
from individual experience of doubt to
illumine possible societal contexts which
are conducive to engendering doubt.
Ultimately, a broad survey of doubt—
whether focusing on the lived experience
of individuals or larger sociological
trends—demonstrates that doubt is by
no means a unitary phenomenon.
It is this sort of certainty with
which this paper does not concern itself.
Rather, throughout this investigation I
seek to further illumine the implications
of doubt in the lives of religious persons.
More often than not, the term doubt is
confounded with disbelief. Yet it is vital
to distinguish between these two terms.
As philosopher Geddes MacGregor
With respect to individual
experience of doubt, the character of
86
such doubt is dictated by several factors.
It is important to distinguish the object
towards which doubt is directed. The
range is expansive: we may doubt the
existence of ultimate metaphysical
reality; we may doubt the religious
activities and symbols we employ in
order to stand in relation to that reality;
or we may even doubt our own ability as
agents to be effectively in accord with
transcendence. Religious individuals
have directed doubt towards these
various objects, and in so doing found
either the ability to respond creatively or
faced the possibility of a stifling
skepticism. Protestant theologians Paul
Tillich and M. Holmes Hartshorne prove
especially
insightful
in
further
understanding the first possibility of
doubt’s creative potential. The Greek
school of Skepticism elucidated by
Sextus Empiricus further elucidates the
second option, namely the stagnating
capacity of doubt.
Perhaps the truest thing that can be said
about doubt is that its only essential
element is uncertainty.
The Creative and Crippling
Potentialities of Doubt
Insofar as we question, look upon
the world critically, and challenge the
certainty of any given proposition, we act
as doubters. ‘The unexamined life is not
worth living,’ proclaimed Socrates.
Philosopher Robert M. Baird explains
that in so doing, “Socrates meant that to
be a rational human being is to have a
kind of creative doubt about one’s
fundamental value beliefs, a creative
doubt that promotes, within reason, a
continual reappraisal of such beliefs”
(Baird 172). In this instance, doubt
doesn’t engender a destructive state of
hopeless despair in which one becomes
disenchanted with what he or she deeply
believes and values. Rather, doubt acts a
sort of refining mechanism—a means by
which one tests and appraises what he or
she holds to be true. In proceeding as
such—especially with respect to religious
belief—Baird argues that one’s doubt
offers the possibility of significant
creative potential.
Apart from doubt’s creative
and/or crippling potential, we can see
how— specifically through St. Augustine
as well as the Zen notion of the Great
Doubt—doubt in and of itself can be
integral to religious experience. Further,
contemporary psychological studies
illumine the mental effects doubt has in
the lives of religious persons. As is
evident, the focus on doubt will be
construed through many lenses in the
following pages. It is my hope that by
looking at doubt in the lives of religious
men and women—albeit in varying
times, places, and religious traditions—
we can come to see more clearly the
dynamic nature of uncertainty in
religious life. What we doubt, and how
we respond to such doubt, vary radically.
However, Baird cautions that
doubt’s creative potential is limited to
certain contexts. As a foil to Socrates’ call
for an open-minded inquisitiveness
towards our beliefs, Baird highlights
Spanish author Miguel de Unamuno’s
renowned work, San Manuel Bueno,
martir (St. Emmanuel the Good, Martyr).
In Unamuno’s short story, we come to
know the parish priest San Manuel, who
is deeply beloved by the townspeople of
Valverde de Lucerna. We also see a
87
profound tension: just as San Manuel
serves as a model of religious inspiration
and devotion for his flock, he personally
harbors deep seated doubts about God.
Yet he doesn’t disclose his secret, and
instead chooses to take it to his grave
rather than shake the faith of the people
who love and adore him so much.
Throughout the book, San Manuel
undergoes extreme duress in wrestling
with his doubt. Yet he ultimately
conceals his doubt so as not to arouse a
devastating threat to the faith of the
townspeople.1
whose dogmatic and inconsistent views
are her abiding source of joy and
comfort” (174). Given this premise, Baird
presents four foci around which doubt
can manifest its creative potential, even
though he acknowledges “that there are
other settings in which such activity
would be inappropriate” (174). He
identifies these four foci as 1)
acknowledging human limitations; 2)
keeping fundamental beliefs alive; 3)
challenging the adequacy of symbols;
and 4) challenging the quest for
certainty. In these four areas Baird seeks
to show the domains in which doubt
functions as a wellspring of vitality as
opposed to inhibiting uncertainty.
Baird explains that in this instance
we see “an age-old conflict. Socrates
versus
Emmanuel.
Truth
versus
happiness. And the reason it is an ageold conflict is precisely because the
competing claims of the two values—
truth and contentment—are both so
compelling” (173). Socrates calls his
students to doubt so that they may better
know truth while San Manuel conceals
doubt so as to not rupture the
contentment of his flock. Consequently,
Baird posits that doubt ought to be
approached differently, depending on
whose beliefs are subject. Presumably,
embracing doubt will have different
consequences for curiously committed
individuals seeking truth as opposed to
“a ninety-five-year-old grandmother
With respect to the first domain,
‘Acknowledging human limitations’ Baird
focuses specifically on Paul Tillich’s
conception of faith. In Dynamics of Faith
Tillich argues that faith is the state of
being ‘ultimately concerned’: “Faith is a
total and centered act of the personal
self, the act of unconditional, infinite,
and ultimate concern” (Tillich 9). Faith is
the means by which we become engaged
with whatever it is that we consider to be
the Absolute. Baird argues that doubt
can potentially play a creative role when
directed towards our ‘ultimate concern’
to the extent that we use doubt as means
to acknowledge our limitations. That is,
as doubters we critically ask ourselves if
we are truly standing in relation to the
ultimate. Baird explains,“The very fact
that faith can be misplaced, that is, that
one can take as ultimate that which is
not ultimate at all, is a reflection of the
logical possibility that a person may be
mistaken” (175). Baird’s presumption is
that through doubt, we recognize our
possibility for error, and thus enact our
1
A 2004 study published in Sociology of Religion
entitled “Religious Doubt and Health: Exploring
the Potential Dark Side of Religion” found that
religious doubt had psychologically deleterious
effects for religious Americans. The study also
found that religious doubt has a “more
deleterious effect on the health and well-being of
individuals who occupy formal roles in the
church, than on rank and file members who do
not have a formal church role” (36).
88
creative capacities in order to correct and
avoid error, or misplacement of faith.
‘atman’ (Hecht 104). The Buddha
doubted the traditional rendering of God
and gods commonly expounded at the
time. Yet in doubting God he didn’t
negate that there was ultimate meaning
to reality. Rather, through his doubt he
articulated a different means by which
people come to engage with such
meaning. For the Buddha, dharma not
Brahman would be of ultimate concern.
Baird concludes that “creative doubt is a
constructive
expression
of
the
recognition that one is not God, and
therefore, just may be mistaken” (175). In
recognizing that we may be mistaken, we
are called to be actively refining our
faith, and this is where doubt’s creative
potential arises.
Historian of Religion Wilfred
Cantwell Smith explains that each caste
in Indian society had its own dharma or
“pattern of religio-social behavior”
(Smith 13). He additionally explains that
dharma was only important to the extent
that it facilitated the “truly major
endeavor” of transcending the mundane
world to “attain a numinous Reality
(“Bhraman”) beyond it” (14). Smith
continues that what “the Buddha did was
to reverse the order of these levels. For
him Brahman and the Gods, while not
negated, became part of this world; while
dharma was elevated to finality, to
absolute transcendence” (14). That is, the
Buddha maintained that dharma , right
living in this world, should be of ultimate
concern, not attaining an otherworldly
Brahman status. Smith concludes that for
the Buddha, “the only final truth is
goodness, that a human life well lived
reflects,
exemplifies,
transcendent
reality. The point is not to transcend the
world but to live well in it” (14). Here we
see Baird’s first facet of doubt’s creative
potential manifest. The Buddha felt that
an ultimate orientation towards ‘atman’
and Brahman was a human error. He
doubted
the
dominant
religious
conceptions of his time and in so doing
articulated a new means by which men
The Buddha’s response to the
dominant Vedic religious traditions of
his time illustrates this specific facet of
doubt’s creative potential. As historian
Jennifer
Michael
Hecht
explains,
“Siddhartha Gautama created a way of
living that actively addressed the
seeming rupture between the human
world and the nonhuman universe, and
he did so while profoundly doubting God
or gods” (Hecht 111). In particular,
Siddhartha expressed a great uncertainty
towards the Vedic notion of True Self, or
‘atman.’ Hecht continues, “The religion
of the Vedas held that human beings
who had reached the stage of the
Brahman could come to know their true
self, their atman, and thereby reach bliss
and release” (104). Yet throughout his
religious endeavors, Siddhartha was
never able to realize such a notion. In
fact, quite the opposite occurred.
Upon becoming enlightened he
sought to fundamentally redirect our
‘ultimate’ focus: “Instead of using [Vedic]
techniques to find our atman, our truest
self, we are to use these techniques, the
Buddha instructed, to come to
understand Anatman, the doctrine of
noself—the Sanskrit negation ‘an-‘
attached to the great Hindu ideal,
89
and
women
transcendence.
could
partake
then, may serve as a check against
idolatry itself. (177)
of
Baird’s second focus maintains
that doubt also has the power to
revitalize old beliefs and practices. He
specifically claims that doubt plays a role
in ‘keeping fundamental beliefs alive’:
“Most
basic
beliefs
and
value
commitments are initially inherited from
parents, peers, and society at large. If
these beliefs and commitments are not
challenged by creative doubt, they tend
to become simply verbal professions
having little vitality” (175). Baird’s point
here is simple: passively accepting the
veracity of a statement without ever
questioning its integrity can potentially
lead to stagnant, seemingly dogmatic
engagement. Doubt can be creative if it
stimulates a vivacious, inquisitive spirit
with respect to inherited religious beliefs
and commitments.
Protestant theologian M. Holmes
Hartshorne echoes this sentiment,
noting that not just symbols, but
religious practice itself, can fall prey to
idolatrous tendencies. In The Faith to
Doubt he contends, “Without doubt,
faith easily becomes idolatrous, and
religious activities and beliefs usurp the
place of God…like all else human the
church’s doctrines are subject to error
and bias, and only radical doubt can
prevent their employment as symbols”
(Hartshorne 102-103).
Within the Protestant tradition,
Martin Luther’s ‘radical doubt’ played a
fundamentally creative role. In fact,
Hartshorne argues that the central
Protestant notion of justification by faith
is the product of Luther’s doubt
creatively directed toward idolatrous
tendencies. Hartshorne notes that Luther
through “bitter experience…discovered
the endemic impurity of all human
motives” (94). Luther saw these impure
motives acutely present in the Church,
specifically
through
the
church’s
“separating men from God by fostering
an idolatrous dependence upon their
own religious zeal” (94). That is, Luther
felt that the Church’s stress on good
works and the importance of the
sacraments (not to mention the sale of
indulgences) too heavily redirected focus
away from God and back onto humans’
own productive efforts. In expressing his
uncertainty in the efficacy of such
practices, Luther helped revitalize the
notion of justification by faith. “For
The third facet of doubt’s creative
potential, argues Baird, is its ‘challenge
to the adequacy of symbols.’ In most
religious traditions there is heavy use of
symbols as a means of expressing
conceptions of and relations to the
ultimate. As Baird points out, the
confounding of symbols and the object
to which such symbols point is the basis
of idolatry. And it is with respect to
idolatry that we see doubt’s creative
capacity:
[T]he one who never doubts the
adequacy of his symbols is
worshipping his symbols. To worship
symbols is to worship the pointers to
God rather than God; and to worship
something other than God is, by
definition, idolatry. Creative doubt,
90
heartlessness; he himself did not
know what a day might bring forth
but only that the hour for serving
God was always present. The Cross of
Christ points to the ultimate and
radical insecurity of life in the service
of God. (102)
Luther,” Hartshorne remarks, “the
discovery that God’s grace was accessible
to simple trust of the heart, was the
source of hope and healing” (94). We
here see doubt’s creative capacity being
used to question the adequacy of
symbols and practices. As Hartshorne
concludes, “The power of doubt to clarify
our thinking, to judge our idolatries, to
renew our honesty and our respect for
truth, is an aspect of what Luther
experienced as the grace of divine
forgiveness” (100). Thus, through
doubting the teachings and practices of
the Church, Luther set in motion the
process by which new and vibrant modes
of being religious would emerge.
In calling into question the legal
certainty of the Pharisees, Jesus sought to
orient people towards God in a radically
new manner. This new orientation
towards God, argues Hartshorne, hinges
on doubt rather than certainty. He
asserts, “The presence of God is utterly
ambiguous….The
sign
which
the
Pharisees sought was the certainty which
legalism and literalism promise, and this
is idolatry. The sign that was given
them—i.e., the crucified Messiah—they
could not accept” (87-88). In this
instance, doubting the quest for certainty
creates a new mode of addressing
ultimate concern. Instead of putting full
trust in the Law, there is a new call to
become an example of Christ. As
Hartshorne points out, doubt about
God’s intentions and whether he could
accept God’s call were hallmarks of
Christ’s life.
The fourth facet of doubt’s
creative potential, Baird argues, lies in
“challenging the quest for certainty.” He
states, “The quest for certainty (and the
avoidance of doubt) often manifests itself
in the effort to find some authority to
whom
burdensome
freedom
and
responsibility can be surrendered” (177).
Baird goes on to point out that, “To flee
from this burden, however, is to flee
from the very freedom that is a defining
characteristic of the fully human” (178).
Construed as such, doubt is an essential
facet of creativity; by abdicating one’s
freedom and responsibility, one loses the
ability to think critically and come up
with new ideas.
Thus far, an analysis of Baird’s
four foci of doubt’s creative potential
(acknowledging
human
limitation,
keeping fundamental beliefs alive,
challenging the adequacy of symbols,
and challenging the quest for certainty)
has more or less followed a rather simple
schema: someone expresses uncertainty
with respect to the integrity or veracity
of a religious ideal or notion. From this
questioning, practices or ideas may be
refined or entirely novel means by which
one can be religious are created.
Turning
once
again
to
Hartshorne, we can see how doubting
the quest for certainty has played a
creative role in the Christian tradition.
He writes,
The confidence of the Pharisees in
their knowledge of God’s will is
condemned by Jesus as pride and
91
Yet to Baird’s four foci, I believe it
is possible to add a fifth facet of doubt’s
creative capacity. Looking to St.
Augustine as well as the Great Doubt in
Zen, we can see how doubt helps create
the
circumstances
of
religious
conversion. ‘Conversion’ in this sense
does not necessarily mean changing
allegiance from one tradition to another,
although this may certainly be the case.
Rather, I take conversion to mean
profoundly a transformative experience
whereby one comes to a higher sense of
connectedness towards or understanding
of that which he or she takes to be the
ultimate. Philosopher Mark O. Webb
argues that one manner in which this
type of experience unfolds is via the
‘doubt-resolution-belief’ sequence. The
sequence begins with a feeling of anxiety,
and ends with a sense of resolution.
Webb explains that “part of the initial
anxiety is a doubt about some
proposition or set of propositions, and
part of the final peace is confidence in a
new belief about some proposition or set
of propositions” (83). Doubt engenders a
sense of anxiety as well as an impetus for
resolving such anxiety. And it is in this
resolution that such ‘conversion’ is
manifest.
Augustine “did not feel he was a
Christian until he could give up all sex,
all food beyond his barest needs, and all
worldly enterprise” (195). However, he
found himself continually unable to meet
these standards and fully dedicate
himself to God. In his Confessions he
writes:
I was deeply disturbed in spirit, angry
with indignation and distress that I
was not entering into my pact and
covenant with you, my God…in the
agony of hesitation I made many
physical gestures of the kind men
make when they want to achieve
something and lack the strength.
(Augustine 146-147).
In this instance, Augustine’s doubt isn’t
directed towards God, but rather towards
his own abilities to love God fully. He is
trapped by his earthly desires, and no
matter how hard he tires, he does not
feel he has the strength to break their
ever-more suffocating hold. Hecht
mentions, “We are now on the other side
of doubt…hearing not from someone
whose doubt is all about getting to the
bottom of what’s real, but rather from
someone whose doubt is all about
actively trying to commit oneself to
belief and, momentarily at least, failing”
(197).
Augustine
is
particularly
illustrative of Webb’s ‘doubt-resolutionbelief’ sequence. Before becoming a
Christian, Augustine had varying levels
of commitment to a range of Greek,
Hebrew, Eastern, and Roman religious
traditions (Hecht 198). Throughout this
time, Augustine exhibited sensual and
prideful tendencies. These tendencies
prompted a profound sense of self-doubt
when Augustine attempted to convert to
Christianity.
Hecht
argues
that
Augustine’s profound sense of
failure engendered by his self-doubt
helps to create an accordingly profound
religious experience. In the midst of such
turmoil, the voice of a small girl chanting
‘pick up and read’ causes Augustine to
pick up the Bible and open to Romans
13:13-14 where Paul says, “Put on the Lord
Jesus Christ and make no provision for
the flesh in its lusts.” Upon reading this
92
Death.2 It is the inner spiritual doubt
that motivates the student’s search
for the fundamental meaning of his
existence, and finally revives him as a
new man of real freedom (345).
sentence, Augustine writes in his
Confessions, “I neither wished nor
needed to read further. At once, with the
last words of this sentence, it was as if a
light of relief from all anxiety flooded
into my heart. All the shadows of doubt
were dispelled” (153). In this instance,
doubt plays an integral role in setting the
stage
for
Augustine’s
conversion
experience. Mark O.Webb comments
that the longer doubt remains, it tends to
become more intense. As such, “the
subject
must
regain
equilibrium
somehow in order to continue to
function in the world, so a radical
readjustment of the whole belief system
takes place under the pressure of
persistent doubt” (84). It is not clear that
Augustine
underwent
a
‘radical
readjustment’ of everything he believed,
yet there was nonetheless a shift. He had
presumably read Paul’s words in Romans
before, but it was under the sense of
duress caused by his sense of self doubt
that Paul’s words seemed to offer a
rather bereaved Augustine a new found
solace and understanding.
In this instance, doubt is creative in so
far as it seeks to transcend one’s false
‘dualistic consciousness,’ point to
‘fundamental meaning’ of existence, and
revive a being of ‘real freedom.’ In
turning to the koans presented in
Shibayama’s Zen Comments on the
Mumonkan, we can see how this
theoretical rendering of the Great Doubt
unfolds in the lived experience of Zen
Buddhists.
The Mumonkan is a collection of
Zen koans, teaching, and commentary
complied in 13th century Japan.
Shibayama oversaw the translation of
this text into English in the 1980’s, and in
his comments, or teisho, on the koans
contained within this text, the
constituent elements of the Great Doubt
find further illumination. The 45th koan
of the Mumonkan reads, “Our Patriarch
Master Hoen of Tozan said, ‘Sakyamuni
[Buddha] and Maitreya [The Buddha to
come] are but his servants. Now tell me,
who is he?’” (307). In his teisho,
Shibayama comments, “Who is this ‘he’?
Arouse the Great Doubt in yourself and
see!”(308). He then goes on to state:
The Great Doubt found in the Zen
tradition also illustrates the manner in
which
doubt
can
help
create
transformative
experience.
Zenkei
Shibayama, former Zen Master of the
Nanzenji Monastery in Kyoto, explains
that the Great Doubt,
Inquire,
inquire,
and
inquire
exhaustively until the whole universe
is just one lump of Great Doubt. At
this extremity, if you do not stop but
is not our ordinary intellectual doubt,
but the fundamental doubt, or quest,
of man that drives him to the last
extremity
of
his
dualistic
discriminating
consciousness
in
order to break through it in the Great
2
Shibayama writes that the Great Death means to “be
dead through and through—to transcend both life and
death” (345). To experience the Great Death is to
have fully transcended duality and to be in accord
with the world of Reality (112).
93
go on doubting and inquiring further,
then, when the opportunity is ripe,
the time will come when the Great
Doubt will naturally be broken
through by itself, and you will
directly come upon ‘him.’ (308)
discriminating consciousness. In pushing
through the Great Doubt we break
through the separation that is the self;
we break through the endless dualities
that are engendered by the self. In turn,
Buddha Nature arises. As Shibayama
explains, “All sentient beings have
Buddha Nature. This is the fundamental
truth of nondualism and equality (24).
We see doubt’s creative capacity as a
facet of transcending the discriminating
nature of the self and experiencing
nondualistic Buddha Nature.
How then are we to understand the
relation between the Great Doubt and
‘him’? Shibayama tells us, “The ancient
Masters were never tired of pointing out
that ‘he’ is not someone standing over
against ‘I’…that he is one’s original True
Self” (308). It is this True Self, one’s
‘Buddha nature’ so to speak (which
transcends all dualistic distinctions
between ‘he’ and ‘I’) that is realized and
intimately known by pushing through
the Great Doubt.
Thus far, I have attempted to
show the manner in which doubt can
serve a creative purpose within the life of
religious persons. This doubt has been
manifest in various ways, including
doubting whether our faithful efforts are
truly directed toward the ultimate;
doubting the adequacy of religious
symbols and practice; as well as via the
self-doubt of Augustine and the doubting
of ‘self’ in Zen. Given the wide-ranging
foci of doubt, what are the conditions
necessary that doubt might function in a
‘creative’ manner within religion? There
appear to be two constituent criteria: 1)
The ultimate is not the object of doubt,
but rather our relation to the ultimate;
and 2) such doubt must not be
‘skeptically dogmatic’. Hence, criterion
one establishes doubt’s focus and
criterion two details doubt’s character.
Shibayama
refers
to
this
realization as ‘satori,’ the fundamental
experience in Zen. In order to experience
satori, “One has to cast his ordinary self
away and be reborn as a new self in a
different dimension…to transcend the
dualism of yes and no, subject and
object” (25-26). Satori is experienced via
the Great Doubt when one doubts in
order ultimately to transcend the self.
There is an interplay between one’s own
self effort, and the arising of Satori in
which ‘self’ is transcended. As Shunryu
Suzuki Rōshi explains, “We must make
some effort, but we must forget ourselves
in the effort we make” (Carter 118). It is
this forgetting of the self towards which
the Great Doubt orients us. From the
standpoint of ‘self’ we employ a
discriminating consciousness which
propagates false dualities. In proceeding
to doubt with more profundity the
constituent ideas that make up the
notion of ‘self’ we begin to doubt all else
that is a product of the self’s
Hartshorne directly addresses the
first criterion, placing it in a specifically
theistic frame. He states that a man is
“given the courage to doubt creatively
only as he experiences in doubt itself the
self-destroying, self-creating power of
truth—not some particular truth, not
some particular god, but truth as the
94
power of being, the power of God” (101).
Creative doubt opens up an avenue for
God’s truth. But this doubt is not
directed towards God, but to the ‘truths’
and ‘gods’— the finite and idolatrous
facets of religion—which we as finite
beings construct. As these finite things
are projections of ourselves, when we
doubt them we experience the ‘selfdestroying’ detailed by Hartshorne; we
subsequently engage in ‘self-creating’ by
recognizing our true relation to God.
Doubt’s creative potential—as manifest
in Baird’s four foci—hinges on its object.
Implicit in Baird’s four foci is that doubt
isn’t directed towards whether or not
there is an ‘ultimate’ about which to be
concerned; rather we question the
veracity of our notions and the manner
in which we stand (or fail to stand) in
accord with the ultimate.
of equal validity. Thus, “[u]nable to
decide which was more convincing in all
of the inquiries they undertook, they
were led to epochē, or to the ‘suspension
of judgment’” (32). This suspension of
judgment is the cornerstone of the goal
of the Skeptical attitude, ataraxia, a state
of ‘tranquility, or freedom from
disturbance” (32). The Skeptics held that
belief, especially with respect to opinions
about good and evil, was a primary
source of human suffering. Yet, by:
[S]uspending judgment about the
nature of good and evil, the Skeptic
avoids this chaotic pursuit of fragile
and
uncertain
belief:
‘neither
avoiding nor pursuing anything with
intensity.’ In this way he is freed from
imaginary fears, and so ‘his feelings
are moderate.’ (Lom 33)
Avoiding intensity and moderating
feelings constitute the tranquility of
ataraxia which the Skeptics believed
could only come about through the
suspension of judgment. In this sense,
Skepticism epitomizes dogmatic doubt
because the pursuit of truth is necessarily
bracketed. Under this attitude, one does
not question his or her doubting attitude
towards belief.
The second criterion dictates that
doubt will not take on a ‘dogmatic’
character. This dogmatic doubt is best
exemplified by the ancient Greek school
of Phyronnian Skepticism. The Skeptical
attitude, as articulated by the 3rd century
A.D. scholar Sextus Empiricus, has three
constituent parts: isostheneia, epochē ,
and ataraxia. Political philosopher Petr
Lom explains, “The Skeptic finds that for
every proposition he examines in any
area of inquiry, there is always another
contradicting it, and he is never able to
come up with a criterion to resolve these
differences” (33). This stage represents
the first part of Skepticism, isostheneia,
in which “an ‘equipollence,’ or seemingly
equal persuasiveness of all claims
considered” (32). Lom explains that
Skeptics used philosophy to evaluate the
truth claims of a given proposition, but
continually found a contradictory claim
This attitude proves destructive
for doubt’s creative capacity. Hartshorne
specifically addresses this skeptical
attitude, taking into consideration its
destructive capacity:
[D]oubt is also a sign of
defensiveness, expressing our distrust
of genuine passion. Men also doubt
in order that they may relieve the
finality of utter commitment…where
doubt refuses the discipline of reason
and fails to honor the demand for
95
truth, which give it its integrity and
power, it is transmuted into its
opposite: dogma. (96-97)
transcendent, convinced that ultimate
truth is inevitably beyond our grasp. Or,
as Hartshorne so deftly asserts, doubt
may be the means by which we choose to
reject truth, ‘in order that we may relieve
the finality of utter commitment.’ Given
these
parameters,
how
might
encountering doubt in a creative, as
opposed
to
crippling,
manner
psychologically affect religious persons?
In turning to psychological studies of
religious doubt, the implications of
experiencing doubt in one of these two
general capacities become manifest.
Hartshorne is decrying doubt when it is
used as an excuse to not take a stand.
This is precisely the base upon which the
ataraxia derived from epochē in Greek
Skepticism rests. In Hartshorne’s view,
when doubt or uncertainty is taken as
‘ultimate,’ then any genuine search for
truth is thwarted. Doubt becomes
unquestionable and hence dogmatic.
Tillich specifically refers to this stance as
‘skeptical doubt.’ He explains, “The
skeptical doubt is an attitude toward all
the beliefs of man, from sense
experiences to religious creeds….It is an
attitude of actually rejecting any
certainty” (22). Baird’s fourth focus
contends that doubt helps us evade the
quest for certainty and in so doing avoid
the stifling of creativity that comes with
the ‘certainty’ of religious dogma. Yet
when doubt itself cannot be questioned,
it transforms into its own dogma; in this
sense, doubt loses its creative capacity
within religion because it becomes a
stagnant agnosticism.
Psychologist
Neal
Krause
examined doubt in older, religious
Americans and his research illustrates
the varying psychological effects caused
by either creative or crippling doubt.
Krause’s 2006 study, “Religious Doubt
and
Psychological
Well-Being:
A
Longitudinal Investigation,” surveyed
Americans “who were Black or White,
noninstitutionalized, English speaking,
and at least 66 years of age” (291).
Additionally, all those surveyed were
self-professed members of the Christian
tradition. The study sought to “see if
religious doubt is associated with change
in three measures of psychological wellbeing over time: Life-satisfaction, selfesteem,
and
optimism”
(287).
Additionally,
Krause
investigated
whether the educational attainment of
older Americans influenced the manner
in which they experienced doubt. Across
the three spectra of life-satisfaction, selfesteem, and optimism, Krause found that
doubt about one’s religion had a
deleterious psychological effect (298).
However, Krause found that the
deleterious effects of doubt in these three
areas were significantly mediated given
There appear to be two general
potentialities of doubt within the lives of
religious persons. On one hand, doubt is
the means by which we exercise a critical
attitude towards our religious ideas,
beliefs, and capacities. In this sense,
doubt can serve a creative function; it
compels us to reanalyze, reconfigure, and
reconstruct the means by which we
faithfully engage our ultimate concern.
On the other hand, doubt can have a
crippling capacity. When doubt is not
dealt with critically, we may find
ourselves unable to feel and relate to the
96
one’s degree of educational attainment.
That is, doubt had a larger psychological
impact on the life satisfaction, selfesteem, and optimism of older
Americans who had fewer years of
education. Doubt produced less severe
psychological effects in Americans with
more years of formal education (298).
and additionally feeling that doubt is
wrong and unforgivable— are why doubt
diminishes more severely the lifesatisfaction, self-esteem, and optimism of
older Americans with less educational
attainment.
It is worth mentioning that that
these psychological survey data reflect
the sentiments of a certain group of
people operating within a specific time,
place, and religious tradition. It would be
implausible to suggest that Krause’s
findings concerning the relationship
between doubt’s psychological effects
and educational attainment could be
extrapolated to comment on religious
persons in differing historical, social, and
religious contexts. That being said, the
data provided by this study give
grounding to the general dichotomy
between creative and crippling doubt
detailed earlier. With respect to the older
Americans with higher degrees of
educational attainment in Krause’s study,
doubt engenders a creative response.
Critical thinking skills were used to
create new understandings and outlooks
which would mitigate doubt’s effects on
psychological well-being. On the other
hand, those who were not able to
confront their doubt critically were
forced to repress or deny it. In this way,
doubt has a paralyzing effect. It becomes
non-resolvable, constantly lurking and
not offering any opportunities for escape.
Responding creatively to doubt mitigates
its psychological effects; it is when doubt
is dogmatic (ever present and nonresolvable)
that
the
greatest
psychological stress is caused.
Krause theorizes that, “Given the
lifelong influence of schooling and
occupational experiences, people with
higher levels of educational attainment
are more likely to apply their skills in
wrestling with, and resolving, religious
doubt” (290). This theory suggests that
the
greater
intellectual
resources
(specifically critical thinking skills)
afforded to individuals with higher
education give them greater capacities to
reason with and confront the challenges
presented by doubt. Conversely, Krause
posits that,
In contrast to those with a good
education, older people with less
schooling are likely to lack the skills
needed to grapple with and resolve
doubt. As a result, the options open
to them are more limited and less
attractive. If doubt cannot be
resolved, then one option is to
ignore, dismiss, repress, or deny it.
(290)
Data gathered from the survey also
revealed that older adults with less
education “are more likely to feel that
having doubts about religion is
wrong…and they are less likely to forgive
themselves when they encounter doubt
about their faith” (298-299). Krause’s
longitudinal survey data suggest that
these facets—feeling that one can only
deal with doubt via repression or denial,
97
Doubt and Secular Society
were ‘Fairly certain’ and 4% were ‘Not
too certain/Not at all certain/Unsure
how certain.’ These data suggest that
theological
doubt
is
relatively
uncommon amongst Americans, as
roughly 88% of people express high
degrees of certainty concerning God’s
existence.3
What might be the sociological
conditions that ground doubt in a
‘modern’ secular society like the one in
which the individuals in Krause’s study
find themselves? In the following pages I
turn the discussion away from the role of
doubt’s potentially creative and/or
incapacitating
potential
and
its
subsequent psychological effects. Rather
I seek to illustrate the manner in which
secular society offers grounding for
doubt. I further seek to show the manner
in which a key element of contemporary
society, modern science, institutionalizes
doubt within its internal functioning.
However,
religious
studies
professor Jack Miles, in conducting
research for his Pulitzer Prize winning
book God: A Biography, met many
Americans who did not fit into this mold
of widespread certainty. While talking to
Americans about God he continually
came across a defiant attitude which
questioned whether doubt and religion
were incompatible: “If I may doubt the
practice of medicine from the operating
table, if I may doubt the political system
from the voting booth, if I may doubt the
institution of marriage from the conjugal
bed, why may I not doubt religion from
the pew?” (2). Miles notes that this
attitude of ‘doubting religion from the
pew’ was most often manifest in
“newcomers” to the expression of doubt
as well as “newcomers” to the pew.
Generally
speaking,
we
might
characterize these individuals as being in
a liminal or intermediary state: they may
not have been ‘religious persons’ before,
and still may not ‘believe’ in the truth
claims of a religious tradition, yet they
nonetheless decide to insert themselves
in a religious community. How might
In its 2008 ‘U.S. Religious
Landscape Survey,’ the Pew Forum on
Religious and Public Life sought to gain a
better understanding of American
religiousness. Over 35,000 American
adults were surveyed in the study, which
sought to measure various facets of their
religious lives including ‘Frequency of
Prayer,’ ‘Importance of Religion’ and
‘Frequency of Attendance.’ One section
in particular measured the degree of
certainty concerning ‘Belief in God or
Universal Spirit by Religious Tradition.’
The survey included Americans from a
variety of religious backgrounds as well
as people ‘Unaffiliated’ with a religious
tradition. Those surveyed were asked to
respond to the following series of
questions: “Do you believe in God or a
universal spirit? [IF YES, ASK:] How
certain are you about this belief? Are you
absolutely certain, fairly certain, not too
certain, or not at all certain?” Overall,
with respect to ‘Total Population’s Belief
in God’ the survey found that 71% of
Americans were ‘Absolutely certain,’ 17%
3
All information regarding the Pew Forum’s
findings, including survey techniques and statistical
analysis can be found at http://www.pewforum.org/.
Data and statistical analysis specifically concerning
‘Belief in God’ can be found at
http://www.pewforum.org/Topics/Beliefs-andPractices/Belief-in-God/
98
one understand the phenomenon of
doubters nonetheless attending religious
services?
range of options and alternatives
concerning religion are apparent.
Doubt, in this instance, engenders
an attitude whereby individuals engage
in an odyssey of sorts. Individuals may
find themselves unsure as to which
option (religious or otherwise) will
provide a sense of meaning or
fulfillment. In A Secular Age Charles
Taylor highlights the lived experience of
‘secularity,’ and this conceptualization
helps us further ground the phenomenon
of doubt in contemporary society. In
detailing the secular, Taylor notes that
he wants specifically to “focus attention
on the different kinds of lived experience
involved in understanding [one’s] life in
one way or the other, on what it’s like to
live as a believer or an unbeliever” (5). He
offers a schema whereby we understand
our lives in terms of ‘fullness,’ ‘exile,’ and
a ‘stabilized middle condition’ between
the two. (5-6). He notes that religious
and non-religious persons have different
means of attaining this state of ‘fullness.’
For the former, it is often found ‘outside’
of the self, specifically in relation to an
external religious tradition and/or
religious community. For the latter, “the
predicament is quite different. The
power to reach fullness is within” (8).
That is, this sense of ‘fullness’ can be
achieved through our own natural
capacities and social constructs; one is
not
dependent
on
a
higher,
transcendental force.
Miles suggests that we focus on
endemic American individualism. He
argues that religion has been a
counterforce to American individualism,
but “religion has not been the only
corrective available. Innumerable secular
forms of association have also tried to
deliver the psychological and moral
counterbalance
that
American
individualism requires” (2). Miles
contends that the phenomenon of the
‘doubter in the pew’ reflects that
“Americans have not so much recovered
their faith in religion as lost their faith in
the alternatives” (3). As such, these
people find themselves in a sort of
intermediary
state—they
are
not
convinced of the validity of religion nor
do they find a sense of fullness in secular
alternatives. Miles concludes:
What appeals to them in the first
instance may be the social and
esthetic refuge provided by religion,
but they arrive with open minds
regarding belief. This openness is the
defiance I noticed in my book-tour
conversations. It is the defiance of
the doubter in the pew. (3)
While Miles’ evidence is largely
anecdotal—we don’t have statistics about
the prevalence of ‘doubters in the pew’(I
would imagine that the phenomenon is
exceedingly
rare)—it
nonetheless
encapsulates a situation indicative of a
larger sociological trend. While most
Americans may never find themselves as
a ‘doubter in the pew,’ they nonetheless
live in a time and place in which the
Thus, we see the ‘doubter in the
pew’ caught between two or more
competing alternatives in his or her
search for fullness. And it is this idea of
competing alternatives that, according to
Taylor, truly characterizes living in a
99
secular age. In characterizing the secular,
Taylor chooses not to focus on
institutional separation between church
and state (what he refers to as ‘secularity
1’) or on declining belief in God/church
attendance (secularity 2) (3). According
to Taylor these two notions focus on
institutional and quantifiable aspects of
secularism, and do not focus primarily on
lived experience as it relates to the search
for ‘fullness.’ As such he proposes
‘secularity 3,’ “which came to be along
with the possibility of exclusive
humanism, which thus for the first time
widened the range of possible options,
ending the era of ‘naïve’ religious faith”
(19). ‘Naïve religious faith’ in its most
extreme context implies a religious
atmosphere in which it is literally
impossible to conceive of alternatives to
the dominant religious tradition. Taylor
argues that this state effectively
characterized much of Medieval Europe.
gives rise to an ‘Index of doubt.’ This
index means that religious persons
live in a condition where we cannot
help but be aware that there are a
number of different construals, views
which
intelligent,
reasonably
undeluded people, of good will, can
and do disagree on. We cannot help
looking over our shoulder from time
to time, looking sideways, living our
faith also in a condition of doubt and
uncertainty. (Taylor 11)
The secular condition is optimized by an
ever-present doubt stemming from the
apparent validity of an array of
competing options for fullness, both
religious and secular. The possibility of
‘naïve’ certainty is all but usurped. As
such, the Pew Forum’s findings
concerning the certainty or near
certainty of Americans’ belief in God
must be understood in a context of
alternatives: “No one today can but be
conscious that living according to the
precepts of a determined faith is one
choice among other possibilities”
(Giddens 181). While Americans may
profess to be certain, they are almost
certainly aware that their notions
necessarily compete with other options,
of which disbelief is a prime contender.
Giddens further explains that “doubt
filters into most aspects of day-to-day
life, as least as background phenomenon.
So far as lay actors are concerned, its
most important consequence is the
requirement to steer between the
conflicting claims of rival types of
abstract system” (181). The prevalence of
so many alternatives— be they religious
or secular ‘abstract systems’— is key in
helping to fully understand the
Less severe gradations of this
naiveté
still
produced
a
near
impossibility of alternatives. That is,
although there may have been a milieu of
various traditions, “these alternatives
were scarcely substitutes for the
overarching authoritative reach of the
dominant religious system” (Giddens
194). However, the secular age is
characterized by competition between
feasible alternatives. That is, no one is
able to live ‘naively’ as if their religious
tradition were the only conceivably
viable option. Additionally, disbelief has
come to be a widely accepted option.
Accordingly, the lived experience of
secularity consists of having to choose
between options (whether they be
‘religious’ or ‘exclusively humanist’) in
one’s quest for ‘fullness.’ Secularity 3
100
sociological situation from which doubt
arises in a secular, modern state.
immutable truths. This notion of
‘securely founded knowledge’ grounds
Baconian science. Educational historian
Julie A. Reuben illustrates that until the
1870’s Baconian science was the
dominant
paradigm
in
American
universities. Under this system, "science
was based on careful, unbiased
observations of nature” and the
subsequent classification of these
observations into categories (Rueben 38).
Reuben notes that the collection and
classification of this data served to
uncover natural law. Then, once "a true
law was identified, it would become part
of
an
unchanging
body
of
knowledge....Scientists assumed that a
finite number of natural laws would
eventually account for all phenomena"
(39). Yet the predominance of immutable
bodies of knowledge would be short
lived, specifically undercut by the
implications of Darwinian evolution.
Besides the institutionalization of
‘alternatives’ within modern society,
doubt also finds grounding within the
realm of modern science. In one sense,
modern science can be construed as an
‘alternative’ interpretation of reality in
contrast
with
the
worldview
encapsulated by various religious
traditions. Perhaps this tension is most
acutely seen in the often polemical
conflict between biblical literalists and
scientists concerning Intelligent Design
and
Darwinian
Evolution.
The
competition between these two notions
could act as a wellspring of doubt for
those who are unsure as to which is more
valid; curiously enough, one might argue
the opposite, seeing as the polemical
nature of this debate often causes those
on either side to become more ardently
attached to their already chosen position.
Yet regardless of the ability of hot-button
scientific issues such as these to inspire
doubt, there is another manner in which
modern science helps foster a doubting
atmosphere. Through a historical
analysis of the development of modern
science, we can see the manner in which
doubt has become an intrinsic part of
this institution.
Reuben
argues
that
in
demonstrating the evolution of natural
entities
over
time
via
natural
selection, Darwin helped usher in a new
era of 'Progressive' science: "According to
this view, science could not be
considered simply the discovery of
regularity in nature because nature was
not necessarily regular" (Rueben 45).
Thus, many scientists rejected the system
whereby immutable natural objects were
organized into categories of immutable
natural law, and rather embraced a
process
which
hinged
on
the
verification/falsification of hypothesis
and an open attitude of inquiry (Rueben
48). While some specific facets of
progressive science were later dropped,
new ‘objectivist’ forms of science still
held on to this attitude of open inquiry
In the time of the Enlightenment,
“The original progenitors of modern
science
and
philosophy
believed
themselves to be preparing the way for
securely founded knowledge of the social
and natural worlds” (Giddens 21). This
idea of ‘securely founded knowledge’
essentially held that through science and
philosophy, the natural world could be
rendered intelligible in terms of
101
(Rueben 193). Indeed, the attitude of
open inquiry about the natural world,
hinging on notions of verification and
falsification, is an enduring hallmark of
science, especially as it is practiced
today: “Science depends, not on the
inductive accumulation of proofs, but on
the methodological principle of doubt.
No matter how cherished…a given
scientific tenet might be, it is open to
revision—or might have to be discarded
altogether” (Giddens 21). As such, doubt
is an intrinsic facet of the scientific
process as the certainty of any
proposition is never guaranteed.
religiously. For in both camps, doubt has
creative potential.
Doubting the Ultimate
The examination of doubt up until
this point has been articulated through a
range of lenses. Greek philosophers, St.
Augustine, Zen monks, 20th century
liberal
Protestant
theologians,
contemporary
psychologists,
and
theorists of modernity/secularism have
all contributed perspectives by which we
may come to further understand the
complicated intricacies of doubt in the
lives of religious persons. In examining
these perspectives, I have largely focused
on a certain type of doubt—that is, doubt
directed toward finite entities such as the
self and the self’s ability to stand in
relation to the ultimate. The first
criterion of ‘creative’ doubt mentioned
earlier stipulated ‘1) The ultimate is not
the object of doubt, but rather our
relation to the ultimate.’ That being said,
individuals throughout history have
doubted the ultimate, and it is to the
implications of this doubt that I now
turn.
I do not intend to show that the
prevalence of modern science necessarily
acts as a wellspring of doubt for religious
persons, although this may be the case in
some instances. Rather, the intrinsic
function of doubt in modern science
shows a heavy parallel between science
and the viewpoints espoused earlier by
theologians such as Hartshorne and
Tillich. Certainly not all religious persons
would
accept
these
theologians’
understanding of doubt; yet each
nonetheless shows that doubt can have
significant creative potential in the lives
of many religious individuals. In the
same way, scientific creativity hinges on
the notion that any proposition is subject
to critical analysis. It is through the
process of doubting by which theories
are refined or entirely new ones are
created. As mentioned, doubt may also
play a stifling role in religious life. Yet in
the scientific realm, stagnation primarily
occurs when people have quit doubting.
As such, we can see how doubt plays a
prominent role in the activities of
persons acting both secularly or
Uncertainty as to whether or not
there is a meaningful ultimate towards
which humans may stand in accord is
perhaps the most profound focus of
doubt. As Hartshorne hauntingly relates,
“Many a sincere Christian has knelt to
pray, only to watch his prayer fall limp to
earth, destroyed by the lurking suspicion
that it is but the weak voice of hope
sounding hollowly in the immense
emptiness of nature” (92). This ‘lurking
suspicion’ is not limited to the Christian
tradition alone; it is a suspicion that men
and women from any number of contexts
102
have had to contend with in questioning
whether or not the universe is empty of
an ultimate meaningfulness to which
they can relate. What responses have
men and women employed with respect
to the possibility of such emptiness?
symbols, practices, etc.), doubt may
engender either a creative or an
inhibiting response. This possibility of a
dual response is at play concerning
questions of ‘ultimate doubt’ as well,
although this twofold response is more
adequately rendered in terms of ‘trust’
and ‘distrust.’ Turning to H. Richard
Niebuhr we can more clearly see the
wellspring from which ‘ultimate doubt’
issues and how we might characterize
responses to such doubt in terms of
‘trust’ and ‘distrust.’
Some people have indeed found
ultimate
meaning
in
emptiness.
Nagarjuna, an early Indian Buddhist
philosopher articulated how this might
be the case, via his explanation of
śūnyatā and svabhāva . The Sanskrit
word śūnyatā , generally defined as
‘emptiness,’
“derives
from
the
recognition…that implicit in thinking
and speaking resides a tendency to create
an illusion (of self-sufficient realities)
that is itself the cause of that suffering”
(Streng, 8856). The notion that there are
completely autonomous, self-sufficient
entities which possess “own-being”
(svabhāva) is antithetical to śūnyatā. We
construct conventional truths, creating
discriminations between one object and
another. To these objects we attribute
svabhāva, yet fail to recognize that
everything is made of constituent parts.
Ultimate truth maintains that “all things
are empty of own-being, or are
dependently arisen” (Mitchell, 144).
Thus, for Nagarjuna and the many
people who have responded faithfully to
his thought, śūnyatā, or emptiness, is
ultimate truth and meaning.
H. Richard Niebuhr posits that we
are radically contingent beings, and the
manner in which we are ‘contingent’
provides the foundation from which
‘ultimate doubt’ proceeds. In The
Responsible Self, Niebuhr maintains that:
The action by which I am, is not one
by which I was thrown into existence
at some past time to maintain myself
thereafter by my own power….I live
but I do not have the power to live.
And further, I may die at any
moment but I am powerless to die. It
was not in my power, nor in my
parents’ power, to elect my self into
existence. Though they willed a child
or consented to it they did not will
me.” (114)
We did not will our particular selves into
being, nor do we know what happens to
us after we die: “We can choose among
many alternatives; but the power to
choose self-existence or self-extinction is
not ours” (115). In this sense, we are
contingent upon the ‘radical action’
which prompted our existence: Niebuhr
thus contends that as embodied selves,
we are here (and sustained in life, and
possibly after death) by some radically
To be sure, not all have found
meaning in emptiness. For many, the
deepest expressions of doubt stem from
uncertainty as to whether or not the
universe is indeed empty of all meaning.
With respect to the finite-oriented doubt
mentioned earlier (that is, doubt directed
towards ourselves and/or our religious
103
other force. We know that we are here,
but we do not know what action or force
rendered
our
existence.
Niebuhr
explains, “The radical action by which I
am and by which I am present with this
body, this mind, this emotional
equipment, this religion, is not
identifiable with any of the finite actions
that constitute the particular elements in
physical, mental, personal existence”
(112). The constraints of our ‘physical,’
‘mental’ and ‘personal’ existence do not
let us know what force engendered our
existence. We do not know if this force is
ultimately meaningful or empty. Hence,
Niebuhr maintains that uncertainty
towards the ultimate is a fundamental
facet of our existence.
they may relieve the finality of utter
commitment” (96-97). This type of doubt
exemplifies an attitude of distrust. Doubt
is taken as an excuse to abandon the
search for truth; through doubting
whether or not there is a meaningful
‘ultimate,’
the
‘finality of
utter
commitment’ is avoided. This sentiment
is paralleled in the Buddhist notion of
Vickicchā. Vickicchā is a form of doubt
hinging on “‘the not wishing to
understand or think.’ It has doubting as a
characteristic, as nature wavering, as
manifest undecision or not accepting
fully, as foundation, the paying of unwise
attention to doubt” (Nyanatiloka, 164).
Vickicchā
essentially
reflects
the
Skeptical doubt detailed by Sextus
Empiricus. When our uncertainty is
directed towards the ultimate itself,
‘skeptical doubt,’ (‘Vickicchā’) is the
means by which we reject the search for
ultimate truth.
Yet just because we cannot
ascertain certainty nor fully understand
the ‘radical action’ by which we are does
not mean we cannot relate to such action
nor proceed to understand it. Niebuhr
explains that ‘faith’ is the means by
which we engage in such relation: “Our
primordial interpretation of the radical
action by which we are is made in faith
as trust or distrust. Between these two
there seems to be no middle term. The
inscrutable power by which we are is
either for us or against us” (119, emphasis
added). In so far as this power is
inscrutable, we lack certainty. Thus, in
some sense we are forced to doubt, or
profess an attitude of uncertainty,
towards this ultimate, radically other
power.
Conversely, Paul Tillich illustrates
the manner in which doubt about the
ultimate may still reflect an attitude of
‘trust.’ He writes:
An act of faith is an act of a finite
being who is grasped by and turned
to the infinite….faith is uncertain in
so far as the infinite to which it is
related is received by a finite being.
This element of uncertainty in faith
cannot be removed, it must be
accepted. And the element of faith
which accepts this is courage. (18)
Given the finite confines of the self,
certainty with respect to the ultimate is
beyond our grasp. Tillich contends that
although we may lack certainty, this does
not mean we cannot seek to engage the
ultimate.
We
may
doubt—be
That being said, how might doubt
be faithfully expressed as either ‘trust’ or
‘distrust’ of this ‘inscrutable power’? I
have previously quoted Hartshorne as
saying, “Men also doubt in order that
104
uncertain— as to whether or not there is
an ultimate significance to which we can
meaningfully relate. Choosing to respond
‘trustfully,’ given the fact that we doubt,
is an expression of courage. We do not
know if we will find ultimate meaning or
only emptiness. This is doubt. Yet Tillich
contends that in choosing to proceed
with this element of uncertainty, we
indeed act courageously.
both Hartshorne and Tillich contend,
doubt is a fundamental facet of faith in
the ultimate. To expect certainty, to be
free from doubt, is to ignore the
conditions of our present state: in
Tillich’s terms, we are finite entities
trying to stand in relation to the infinite;
in Niebuhr’s rendering, we are beings
‘contingent’ on a radically other force
about which we can only be uncertain.
But because we doubt does not mean we
cannot seek to understand.
Doubt thus plays an integral role
in relation to ultimate truth. It may be
the means by which we reject the search
for truth and express an attitude of
‘distrust’ towards the ultimate; at the
same time, doubt necessitates that we act
courageously should we decide to pursue
an engaged relationship with ultimate
truth. In this latter sense, doubt is a
fundamental facet of ‘trust’ or faith in the
ultimate. We may then understand
doubt having two sets of dual natures,
depending on its focus. Concerning the
finite orientations of doubt, we see both
creative and crippling potential; in
doubt’s orientation toward the ‘ultimate’
we see the possibility of ‘trust’ and
‘distrust.’
Conclusion: A Personal Reflection
I would like to conclude this
journey through uncertainty with…a
commencement. That is, I wish to share
some words from the commencement
baccalaureate sermon delivered by M.
Holmes Hartshorne on June, 1st 1975. In
the Colgate Chapel, filled with young
men and women like myself about to
graduate, Hartshorne remarked:
It takes courage to doubt—to face
unflinchingly
the
frightening
questions posed by a world in
dissolution or forced on you by the
reluctant awareness that your life is
pitifully superficial and your ultimate
concerns little else than your own
fears and frustrations. If your
graduation
is
in
fact
a
commencement, this will in good
measure be because Colgate has
helped you to think critically,
radically, in spite of great inner
resistance. Such radical doubt is no
mere technique: it is the fruit of an
attitude, a communal attitude of men
and women who somehow have
Doubt’s stifling capacity is
intrinsically linked to this attitude of
‘distrust.’ From an attitude of ‘distrust’
we do not seek to find out if ultimate
truth exists in the first place, much less
try to meaningfully engage with it. The
Skepticism of Sextus Empiricus, the
Buddhist notion of Vickicchā—both
represent an attitude of distrust because
doubt is taken to be ultimately
meaningful and non-questionable. For
individuals living life religiously, seeking
ultimate truth, such doubt may prove
incapacitating. On the other hand, as
105
together found the courage to risk
peace of mind for the sake of truth
and conscience. (2-3)
hours in the library, poring over texts
and reaching for ideas in what feels like a
boundless abyss of knowledge I cannot
even being to truly understand has made
me perpetually question my ability to put
forth a meaningful body of work. More
often than I would like to admit,
finishing this thesis has been my
‘ultimate concern.’ But it is of course not
ultimate, and all of the anxiety I have felt
in writing it reflects, as Hartshorne says,
that my ultimate concerns are ‘little else
than my own fears and frustrations.’ And
for as much time as I spent ‘ultimately
concerning’ myself with putting these
ideas down on paper, I am sure that I will
not be able to recall most of them in a
short time from now. Be that as it may, I
will always know what it is to doubt. I
know the thrill of doubt creating
previously unimagined thoughts and
insights. I also know the despair of giving
up in the face of uncertainty. In both of
these respects I am by no means unique.
I merely have had the privilege to wrestle
with thoughts on doubt as articulated by
some
truly
profound
people.
As I continued to write about the doubt
of men and women, it became
increasingly apparent to me that I was
indeed living my thesis. I have felt
profoundly uncertain as to whether or
not the ideas contained within this work
are accurate reflections. My doubt
towards what I have written has indeed
engendered creative response: In some
cases I simply scrapped ideas about
which I felt uncertain and sought more
refined ways of articulating novel
thoughts. I hope to have had some
success in this endeavor. More often than
not, however, I have been unable to
wrestle with my doubt. There are ideas
I’ve put forth in this paper about which I
feel deeply uncertain. Yet instead of
confronting my doubt about them, I have
let them lie as imperfectly articulated as
they are. In this sense my doubt has
crippled my search for truth.
In pursuing this project, my self
doubt has been profound. Hours upon
Clark, Mary T. Augustine. Georgetown
University Press: Washington, D.C., 1994.
Bibliography
Dawkins, Richard. The God Delusion.
Mariner Books: New York, 2008.
Augustine. The Confessions. Trans.
Henry Chadwick. Oxford University
Press: Oxford, 1991.
Denbeaux, Fred. The Art of Christian
Doubt. Association Press: New York, 1960
Baird, Robert M. “The Creative Role of
doubt in Religion.” Journal of Religion
and Health. 19.3 (1980): 172-179.
Giddens, Anthony. Modernity and Self
Identity: Self and Society in the Late
Modern Age. Stanford University Press:
Stanford, 1991.
Carter, Jonh Ross, ed. Of Human Bondage
and Divine Grace. Open Court: La Salle,
Illinois, 1992.
106
Hartshorne, M. Holmes. The Faith to
Doubt: A Protestant Response to
Criticisms of Religion. Prentice Hall:
Englewood Cliffs, N.J., 1963.
Miles, Jack. “Faith Is an Option: Religion
Makes a Comeback (Belief to Follow).”
New York Times Magazine. 7 December
1997.
http://www.nytimes.com/1997/12/07/ma
gazine/faith-is-an-option-religionmakes-a-comeback-belief-tofollow.html?scp=49&sq="Jack%20Miles"
&st=cse
--Baccalaureate Sermon. Jun 1,
1975 Colgate University.
Hecht, Jennifer Michael. Doubt: A
History.
HarperSanFrancisco:
San
Francisco, 2003.
Mitchell,
Donald
W.
Buddhism:
Introducing the Buddhist Experience. New
York: Oxford University Press, 2008.
Helfaer, Philip M. The Psychology of
Religious Doubt. Beacon Press: Boston,
1972.
Niebuhr, H. Richard. The Responsible
Self: An Essay in Christian Moral
Philosophy. Westminister John Knox
Press: Louisville, Kentucky, 1999.
Hempton,
David.
Evangelical
Disenchantment: 9 Portraits of Faith and
Doubt. Yale University Press: New
Haven, 2008.
-- Radical Monotheism and
Western Culture. Harper &
Brothers Publishers: New York,
1960.
Jones, Lindsay, ed. Encyclopedia of
Religion Detroit: Macmillan Reference
USA, 2005.
Nyananatiloka. Buddhist Dictionary:
Manual of Buddhist Terms and Doctrines.
Frewin & Co., LTD.: Colombo, 1950.
--MacGregor, Geddes. “Doubt and
Belief” p. 2423-2429.
--Streng, Frederick J. “Śūnyam and
Śūnyatā”p.8855-8860.
Pollack, Victor L. “Science Education: I:
The Spirit of Science.” Journal of Science
Education and Technology. 2.4 (1993):
513-518.
Krause, Neal and Keith M. Wulff.
“Religious Doubt and Health: Exploring
the Potential Dark Side or Religion.”
Sociology of Religion 65.1 (2004): 35-56.
Reuben, Julie A. The Making of the
Modern University. The University of
Chicago Press: Chicago, 1996.
Krause, Neal. “Religious Doubt and
Psychological
Well-Being:
A
Longitudinal Investigation.” Review of
Religious Research 47.3 (2006): 287-302.
Rescher, Nicholas. “The Legitimacy of
Doubt.” The Review of Metaphysics.
13.2(1959): 226-234.
Lom, Petr. The Limits of Doubt: The
Moral and Political Implications of
Skepticism. State University. State
University of New York Press: Albany,
2001.
Smith, W. Cantwell. “Religious Atheism?
Early Buddhist and Recent American.”
The Australian Bulletin of Comparative
Religion
107
Shibayama, Zenkei. Zen Comments on
the Mumonkan. Trans. Sumiko Kudo.
Harper & Row: New York, 1974.
Shore, Jeff. Being Without Self: Zen for
the Modern World. Asoka Publishing:
Rotterdam, 2008.
Schellenberg, J.L. The Wisdom to Doubt:
A Justification of Religious Skepticism.
Cornell University Press: Ithaca, 2007.
Sinnot-Armstrong,
Walter,
ed.
Phyronnian
Skepticism.
Oxford
University Press: Oxford, 2004.
Swain, Charles W. “Doubt in Defense of
Faith.” Journal of the American Academy
of Religion. 36.2 (1968): 109-117.
Tillich, Paul. Dynamics of Faith.
Perennial Classics: New York, 2001.
Webb, Mark O. “Religious Experience as
Doubt Resolution.” International Journal
for Philosophy of Religion 18.5 (1985): 8186.
108
Occupational Diversity of Immigrants in the United
States
Colleen Tubridy, Class of 2010
This paper employs IPUMS United States Census data to study the occupational distribution of
immigrants between 1850 and 2000. Examining occupational diversity is crucial to understanding
how immigrants adjust to a host economy and affect growth. I use occupations as a way of
measuring assimilation. I analyze whether or not the spread of immigrants across occupations
looks like the spread of natives over time. To do this, I create a unique measure of diversity based
on the fractionalization index. I construct occupational diversity indices for each immigrant group
and decade. Results from a case study examination indicate that relative occupational diversity has
increased over time as immigrants have a greater ability to enter a wide array of occupations. I use
empirical models to show that a positive relationship exists between relative occupational
diversity and the immigrant to native ratio.
attainment of immigrants is an inherent
part of their socioeconomic adjustment.
Initially, immigrants may have an
occupational disadvantage because they
lack knowledge about how to find
employment, or their pre-migration skills
may not be fully transferable. But with a
longer stay in the U.S. and more
investment in U.S. specific skills, their
socioeconomic status can be expected to
improve and resemble more closely that
of natives (Toussaint-Comeau 2004).
Researchers have proposed several
theories to explain why immigrants
select and change jobs. These provide
insight into the extent to which
immigrants may or may not assimilate in
their occupational profiles. These
theories
include
human
capital
accumulation,
the
host
country’s
immigration policy regarding the type of
skills required for entry, and the
importance of ethnic concentration in
shaping the occupational distribution of
immigrants, to name a few.
Introduction
Information on the occupational
distribution of immigrants and how it
changes over time in a host economy is
central
to
understanding
how
immigrants affect economic growth and
how they adjust to a host country both in
economic and social terms (Green 1999).
The assimilation process, or the
convergence between immigrants and
natives, is thus a very important
phenomenon to study. This is especially
true when looked at through the lens of
the United States, the “melting pot” of
our world. Much of the prior research
that has been done to examine
immigrants in the labor market has
considered wages or unemployment to
make inferences about the assimilation
of immigrants, while very few have
looked at occupations. Occupation,
however, is just as important a measure
of the skill set of an individual that is
used to convey the socioeconomic status
of the worker. The occupational
109
Most people, when studying
assimilation, ask whether immigrants are
taking jobs more similar to natives over
time or if they are catching up to the
performance of natives in regards to
skills or wages. Unlike this assimilation
literature focusing on the catch-up
question, I am asking a different question
from another angle. I am looking at
whether the spread, or distribution, of
immigrants across occupations looks
more like the spread of natives across
time. My emphasis is on whether or not
immigrants have been stuck in a low
number of occupations over time or if
they have been able to break free from
certain classifications. Rather than
studying if immigrants are in “better”
occupations, I seek to find if they are in a
variety of occupations to make sure they
have not been pigeon-holed in particular
categories. To explore this idea, I employ
the concept of occupational diversity.
Occupational diversity can be defined as
the probability that two people, drawn at
random, are of different occupations.
The question of occupational diversity is
thus extremely important, especially
when viewed as one type of assimilation.
A focus on occupational distribution
contributes to the importance of
understanding the immigrant experience
upon arrival to the U.S. Here, I examine
how immigrants assimilate, at least
occupationally, as their group’s residency
in the U.S. increases. I use a blend of
quantitative and qualitative analysis to
get to the root of this question. Through
a mini-case study approach, I am able to
delve into the immigration history and
experience of several of the largest
foreign-born populations in an attempt
to explain changes in occupational
diversity. To do this, I compare natives,
or U.S. born, to those who are foreignborn. I consider a variety of immigrant
groups, including Mexicans, Irish,
Germans, Puerto Ricans, Cubans,
Chinese, Filipinos, and Indians. I look at
trends in occupational diversity over
time, from 1850 to 2000, for each group. I
analyze how they are different from each
other and from natives. Thus, I attempt
to take a new approach to the
assimilation question. I seek to find if
occupational diversity has increased over
time and if so, why that has been the
case.
The paper is organized as follows.
The next section describes the
motivation behind the paper and the
literature related to the topic. Section 3
describes the data, while section 4
explains the measurement index used.
Section 5 presents the results and
interpretation of the mini-case studies.
Section 6 introduces the empirical model
and section 7 provides the regression
results and analysis. Section 8 concludes
the study.
Motivation
In the past, there have been
several studies done about occupational
assimilation and mobility of immigrants.
Amuedo-Dorantes and De La Rica (2006)
analyze
the
employment
and
occupational assimilation of recent
immigrant waves to the Spanish labor
market as their residencies lengthen and
find
evidence
that
occupational
assimilation varies by gender, origin, and
educational attainment. They examine
immigrants’ employment assimilation as
captured by changes in the employment
110
probability differential between similar
immigrants and natives as immigrants’
stay in Spain lengthens. Their omitted
variable, or comparison group, is natives.
They rank occupations on the basis of
their average earnings. 40% of recent
immigrants were concentrated in the 5
worst paid occupations compared to 30%
of non-recent immigrants and 12% of
natives. The fact that non-recent
immigrants are better off occupationwise than their recent counterparts
suggests upward occupational mobility
as migrants’ residencies lengthen. In
terms of educational attainment, the
assimilation of less educated immigrants
seems to take place at a slower pace.
Thus, assimilation may vary according to
human capital. They found that there
were significant differences in the
occupational attainment and mobility of
immigrants by origin.
She cites the Cohort Quality Model
(Borjas 1995) which suggests that there
are differences between groups, or
cohorts, of immigrants who enter the
U.S. in particular time periods. A higher
skill set, or a more transferable one,
changes the occupational path for each
cohort in different times.
Previous research has also shown
the importance of ethnic concentration
in shaping the occupational distribution
of immigrants. Patel and Vella (2007)
study the occupational allocation of
immigrants.
They
examine
the
relationship between the occupational
choice of recently arrived immigrants
with those of established immigrants of
the same country. They use a simple
model of the search behavior of
immigrants. They find evidence of a
network system in the immigrant labor
market. Evidence shows that new arrivals
are locating in the same occupations as
their established countrymen. Thus, they
highlight the importance of immigrant
networks in developing occupation
niches, or channeling immigrants of
certain ethnic backgrounds into specific
occupations. Occupational choice of
immigrants is not based purely on
comparative skill advantages.
Toussaint-Comeau
(2004)
investigates
whether
Hispanic
immigrants assimilate in occupational
status with natives and the factors that
determine occupational status. She
models
occupational
status
and
convergence of Hispanics relative to U.S.
born non-Hispanics as a function of
human
capital
and
demographic
exogenous variables, U.S. experience
(assimilation effects) and periods of
migration (cohort effects). Her results
show that the length of time spent in the
U.S. narrows the occupational gap and
that the level of human capital affects the
rate of occupational mobility and
determines whether convergence occurs
in occupational status. Immigrants who
arrived to the U.S. in different periods
may have different propensities to
assimilate due to human capital levels.
Data
I was able to draw all of my data
from the IPUMS-USA website. IPUMSUSA (Integrated Public Use Microdata
Series) is dedicated to collecting and
distributing United States census data.
Using variables from this data base, I was
able to create a new dataset that is
unique to my purposes. I gathered the
following variables for each decade
111
between 1850 and 2000 to examine the
full spectrum during which the bulk of
migration to the U.S. occurred. However,
there is no census data for the decade
1890. My variable, Age, limits the data to
those who are between 16 and 65, or
those who are of age to be considered in
the labor force. In addition, my variable
for Sex limits the data to males because
occupational patterns of females are
likely to be very different and would thus
skew my results. The variable Perwt
(person weight) indicates how many
people in the U.S. population are
represented by a given person in the
IPUMS sample. The variable GQ (group
quarter) classifies households under the
1970
definition
and
additional
households under the 1990 definition.
The variable Laborforce limits the data to
those who are in the labor force.
other barriers in regards to occupational
options. On the other hand, perhaps
immigrants from countries with colonial
status may have an easier time entering
into a wide array of occupations.
I compute the occupational
diversity of employed non-agricultural
civilian workers for each ethnic group. I
use the variable OCC1950, which applies
the 1950 Census Bureau occupational
classification system to occupational
data, in order to increase the
comparability of occupations over time.
This variable reports the primary
occupation of census participants.
Occupations are classified by job title
within
the
following
categories:
Professional and Technical; Managers,
Officials, and Proprietors; Clerical and
Kindred; Sales workers; Craftsmen;
Operatives; Service Workers; Laborers.
In order to provide a representation of
how these occupations are coded, I
included a list of all of the job titles
within three of these categories in
Appendix A. For example, it can be seen
that two people working in an
accounting office, one as an accountant
and one as a secretary, would be coded
differently. In this way, the variable
provides
data
that
shows
how
occupational diversity is changing in
regards to job title and hierarchy, rather
than solely industry.
I also include the variable BPL
(birthplace) for the following countries:
Mexico, Ireland, Germany, Puerto Rico,
Cuba, China, Philippines, India, and the
U.S. states. I chose these 8 countries in
order to perform a more in-depth
analysis of each group’s experience. I
wanted to examine a wide range of
immigrant
groups
that
were
representative of different regions and
sizes, whose migration waves started at
different points and have spanned across
the majority of the 150 year period that I
am looking at. In addition, I was
interested to see if certain immigrant
groups
had
different
levels
of
occupational diversity due to factors
such as race, language, or having been a
colony of the United States. Non-white
immigrants or those from a country
where English is not the first language
spoken might face discrimination or
Measurement Index
In order to accomplish my task, I
needed to create a unique measure of
diversity. To do this, I based my model
on the fractionalization index. Most
studies about the impact of ethnic
diversity on economic growth use a
112
measure of ethnic diversity called the
ELF (Ethnolinguistic Fractionalization),
which is the probability that two people
drawn at random are of different ethnic
groups. The fractionalization index’s
theoretical maximum is reached (at the
value of 1) when each person belongs to a
different group. Thus, values represent a
range of complete homogeneity (an
index of 0) to complete heterogeneity (an
index of 1) (Alesina et al 2003). I use the
same formula, applied to different
underlying data, to compute my own
measure of fractionalization:
group in each decade compared to
natives, or the Div(Immigrant Group)/
Div(Natives). A value of 1 implies equal
diversity, while values below 1 imply that
immigrants are clustered in fewer
occupations than natives are. The results
can be seen through the following minicase study analysis.
Case Study Results and Analysis
In general, I found that the
relative
occupational
diversity
of
immigrant groups increased over time
from 1850 to 2000. This trend can be
clearly seen as depicted in Figure 1. With
just a few immigrants, occupational
diversity is low because they are only
represented in a few jobs. As the number
of immigrants increases, however, they
become more spread out. Thus, it
appears as though all of the immigrant
groups have
spread
out across
occupations, like natives. However,
despite an overall upward trend in
diversity, many groups experience
declines at certain points. Influxes of
immigrants usually come with the same
level of skills as a group, thus going into
the same occupational area. This
concentration is most often due to the
fact that they are recruited for certain
targeted purposes or they are restricted
from entering the U.S. due to
exclusionary policies. All of the above
serve to produce an effect of decreased
occupational diversity for short amounts
of time. As time goes on, immigrants
entering the U.S. are more skilled for the
most part; this serves to reinforce the
upward trend of occupational diversity
since they are able to enter a wide array
of occupations.
N
FRACTj = 1 – ∑ s2ij,
i=1
where sij is the share of each occupation i
(i = 1…N) in birthplace j. Using this index,
my new data set allows for the
computation of an alternative measure of
heterogeneity. I therefore used this same
idea to construct occupational diversity
indices for each group and decade. This
created occupational diversity for
immigrants from 8 countries, as well as
for natives, across 16 decades. Thus,
fractionalization can be viewed here as a
measure of occupational diversity. One
issue when measuring this type of
fractionalization, however, is that the
number of occupations varies every year.
The problem with this is that if the
number of occupations increases over
time, it will automatically increase
occupational diversity, thus eliminating
the
ability
to
observe
whether
occupational
diversity
is
actually
increasing due to other factors. To try to
address this issue, I calculate the relative
occupational diversity of each immigrant
113
The one issue with these results is
that while it is exciting that all of the
immigrant groups appear to have broken
free from their respective occupational
classifications, it is slightly odd that they
are all clustered around 1. This would
essentially mean that they are just as
occupationally diverse as natives. I
attempted to control for the general
increase in the number of occupations
over time by using the OCC1950 variable
and looking at the relative diversity of
immigrant groups. However, I still may
have failed to control for the fact that
there are just more occupations over
time.
Natives
In Figure 2, we see that the
probability of randomly drawing 2
natives
that
were
in
different
occupations was about 88% in 1850.
Occupational diversity amongst natives
increased over time with natives being
represented in 144 occupations in 1850 to
188 occupations in 2000. Thus, the
probability that natives were in different
occupations in 2000 was about 95%.
Occupational diversity amongst natives
has been largely constant over time and
can thus be used for comparison
purposes. Due to this lack of variation,
the fractionalization graphs for each
country essentially look the same as their
relative
fractionalization
graphs.
.4
.6
.8
1
Occupational Diversity: Immigrants to Natives
1850
1900
1950
Census year
Puerto Ricans
Cubans
Germans
Filipinos
Figure 1
114
Mexicans
Irish
Chinese
Indians
2000
.8
.84
occ fract
.88
.92
.96
1
Native Occupational Fractionalization
1850
1900
1950
2000
Census year
Figure 2
Therefore, I only included both graphs
for Mexicans in order to illustrate this
trend.
(Takaki 1998, 241). In that year, the
United States also signed the Bracero
Treaty, which reopened the floodgates
for legal immigration of Mexican
laborers. Between the period of 1942 and
1964, millions of Mexicans were imported
into the U.S. as "braceros" under the
Bracero Program to work temporarily on
contract to United States growers and
ranchers (“Mexican Immigrant Labor
History”). At the end of World War II,
Mexican workers were ousted from their
jobs by returning servicemen and those
who were coming out of wartime
industries (“Mexican Immigrant Labor
History”).
Mexicans
After the Treaty of Guadalupe
Hidalgo ended the war with Mexico in
1848, about 55,000 Mexican workers
immigrated to the United States over the
next 30 years to become field hands in
regions that had, until very recently,
belonged to Mexico (Takaki 1998, 239).
The presence of Mexican workers in the
American labor scene was also reinforced
with the construction of the railroad
between Mexico and the U.S. That
presence grew between 1880 and 1900.
When the United States entered World
War II in 1942, labor was drawn from all
areas of U.S. industry and poured into
those which supported the war efforts.
Thus, there was high demand for
Mexican laborers in a variety of areas
In
Figure
3,
Mexican
fractionalization
shows
that
the
probability of randomly drawing 2
Mexicans that were in different
occupations was about 43% in 1850.
Occupational diversity within the group
increased over time with Mexicans being
represented in 10 occupations in 1850 to
115
.4
.6
occ fract
.8
1
Mexican Occupational Fractionalization
1850
1900
1950
2000
Census year
Figure 3
175 occupations in 2000. Thus, the
probability that Mexicans were in
different occupations in 2000 was about
93%. Figure 4 shows that Mexicans were
half as occupationally diverse as natives
in 1850, but were only 3% less diverse in
2000.
On the graphs, we
occupational diversity in 1850
most likely due to the fact
workers were field hands at
see low
which was
that most
that time.
.5
.6
relative occ diversity
.7
.8
.9
1
Occupational Diversity: Mexicans to Natives
1850
1900
1950
Census year
Figure 4
116
2000
We see a rise but then a decrease in
occupational diversity in the period from
1880 to 1900, most likely due to the fact
that workers were being concentrated in
the building of railroads. We see
increased diversity in the 40s as
Mexicans were recruited for a wide range
of occupations in both the industrial and
agricultural sectors during the WWII
period. Diversity continues on an upward
trend after the war as they were forced to
scatter into various occupations when
soldiers returned to their jobs after the
war.
way of life. Fortunately for them, the
expansion of the American economy
created heavy demands for labor. At this
time, thousands of miles of rail were
being laid and so Irish laborers became a
part of the construction gangs that did
this grueling work (Takaki 1998, 114).
Thus, the first generations were in largely
unskilled and semiskilled occupations,
but their children found themselves
working in increasingly skilled trades. By
1900, when Irish Americans made up
about a thirteenth of the male labor
force, they were almost a third of the
plumbers,
steamfitters,
and
boilermakers. Industry working Irish
soon found themselves lifted up into
manager positions as more and more
common laborers arrived from southern
and eastern Europe (“Immigration: The
Journey to America”).
Irish
The main reason that the Irish
immigrated to the United States was to
escape the widespread Potato Famine
that plagued the country beginning in
1845 (Takaki 1998, 112). Over the next
seven years, more than one million Irish
came to the United States. This strained
the resources of New York City and
Boston as they struggled to manage the
large influx of immigrants. Hence,
competition for employment with the
existing population was fierce and
employment discrimination against the
Irish became common (Takaki 1998, 113).
These peasants had arrived without
resources or capital to start farms or
businesses. In addition, very few of them
had ever accumulated the resources to
make any meaningful choices about their
Irish fractionalization indicates
that the probability of randomly drawing
2 Irishmen that were in different
occupations was about 69% in 1850.
Occupational diversity within the group
increased over time with the Irish being
represented in 81 occupations in 1850 to
91 occupations in 2000. Thus, the
probability that Irish were in different
occupations in 2000 was about 92%.
Figure 5 shows that the Irish were at 78%
of the occupational diversity level of
natives in 1850, but were up to 96% of
the native diversity level in 2000.
117
.8
relative occ diversity
.85
.9
.95
1
Occupational Diversity: Irish to Natives
1850
1900
1950
2000
Census year
Figure 5
On the graph, we see relatively
low occupational diversity in 1850 which
was most likely due to their lack of skills
and the discrimination that they faced
when they landed in America. Thus, they
were concentrated in certain sectors of
the labor market. Over time, however,
we see a largely consistent upward trend
in occupational diversity. Thus, it
appears that after their initial troubles,
the Irish faced very few problems in
terms of entering a wide array of
occupations. Perhaps this was because
the Irish may have been aided by their
white race and capability of speaking the
English language.
Germans
Many German individuals had
already been in the United States for
quite some time before the midnineteenth century. However, nearly one
million German immigrants entered the
United States in the 1850s for different
reasons
than
earlier
generations.
Increasing industrialization and the use
of machines to perform tasks previously
done by manual labor forced many
Germans from their respective family
businesses
(“History
of
GermanAmerican Relations”). In these later
phases
of
German
immigration,
newcomers joined established settlers.
Germans in America had a strong
influence on the labor movement in the
United States. Thus, labor union
membership
enabled
German
immigrants to improve their working
118
.9
relative occ diversity
.94
.98
1.02
Occupational Diversity: Germans to Natives
1850
1900
1950
2000
Census year
Figure 6
On the graph, we see largely
consistent occupational diversity over
time for Germans. Immigrants who came
in 1850 appear to have almost
immediately assimilated to natives and
had similar, if not higher, levels of
occupational diversity. This is probably
due to the fact that they not only had
skills when they got to the U.S., but they
also had an established network already
in the U.S. to turn to for help. Thus, they
have remained in a diverse array of
occupations over time.
conditions and integrate into American
society as a whole. Originally strong in
such occupations as baking, carpentry,
and brewing, they were also laborers,
farmers, musicians, and merchants
(“History
of
German-American
Relations”).
German
fractionalization
indicates that the probability of
randomly drawing 2 Germans that were
in different occupations was about 89%
in 1850. Occupational diversity within
the group increased over time with
Germans being represented in 73
occupations in 1850 to 163 occupations in
2000. Thus, the probability that Germans
were in different occupations in 2000
was about 95%. Figure 6 shows that
Germans were actually about 1% more
occupationally diverse relative to natives
in 1850 and about 1% less diverse in 2000.
Puerto Ricans
Puerto Ricans have migrated to
the United States since the late
nineteenth
century.
The
Spanish
American War officially ended on
December 10, 1898. The Treaty of Paris
stated that Spain was to cede Puerto
119
looking to Puerto Rico for cheap labor,
and they sent agents there to recruit
workers. Abundant jobs and active
recruitment increased the average yearly
migration of Puerto Ricans from 1,800
between 1930 and 1940 to 31,000 from
1946 to 1950, and to 45,000 from 1951 to
1960 (“Historical Background”). The
history of Puerto Rican American
assimilation has been one of great
success mixed with serious problems.
Many Puerto Rican mainlanders hold
high-paying white collar jobs. Outside of
New York City, Puerto Ricans often boast
higher college graduation rates and
higher per capita incomes than their
counterparts in other Latino groups
(“Historical Background”).
Rico, Guam and the Philippines to the
United States. Puerto Rican migration to
the United States really began, however,
during World War I when the United
States government needed both soldiers
and workers (Takaki 1998, 292). On
March 17, 1917, Congress passed the Jones
Act making the Puerto Rican people
United States citizens. As citizens, Puerto
Rican men were drafted into the United
States army to fight in the war. In
addition, due to the scarcity of workers
in the United States to construct ships
and armaments, the federal government
actively encouraged and recruited Puerto
Ricans for that purpose. During 1917 and
1918, the government transported
thousands of men from Puerto Rico to
industrial
complexes
(“Historical
Background”). Again at the end of World
Puerto Rican
indicates that the
fractionalization
probability of
.9
relative occ diversity
.92
.94
.96
.98
1
Occupational Diversity: Puerto Ricans to Natives
1900
1920
1940
1960
Census year
1980
2000
Figure 7
randomly drawing 2 Puerto Ricans that
War II, United States companies began
120
were in different occupations was about
86% in 1910. Occupational diversity
within the group increased over time
with Puerto Ricans being represented in
12 occupations in 1910 to 153 occupations
in 2000. Thus, the probability that Puerto
Ricans were in different occupations in
2000 was about 95%. Figure 7 shows that
Puerto Ricans were only 9% less
occupationally diverse relative to natives
in 1910 and about 1% less diverse in 2000.
this was considered the first wave of
Filipino immigration (“Countries and
Their Cultures”). Also during this time,
plantation workers were recruited to
work in Hawaii on sugar plantations. In
1934, the U.S. made the Philippines a
commonwealth and limited immigration
to the U.S. to 50 Filipinos a year. The
largest immigration wave arrived after
the passage of the 1965 Immigration Act
which eliminated country specific
quotas.
Changes
in
American
immigration law in 1965 significantly
altered the type and number of
immigrants coming to the United States.
Unlike pre-war immigrants who largely
worked as unskilled laborers in West
Coast and Hawaiian agriculture, the
third wave was composed of larger
numbers of well-educated Filipinos
between the ages of 20 and 40 who came
looking for better career opportunities
than they could find in the Philippines.
This highly skilled third wave population
had a command of the English language,
allowing them to enter a wide range of
professions (“Countries and Their
Cultures”).
Before 1900 and the influx of
Puerto Rican immigrants after the
Spanish American war, the only
occupation that Puerto Ricans were in
was laborers. Thus, they had no
occupational diversity. However, on the
graph, we see a large jump in diversity
after this event as immigrants began to
flow in. Occupational diversity drops
during WWI though because Puerto
Ricans were drafted into the war since
they were now U.S. citizens. Active
recruitment of targeted Puerto Rican
labor occurred in the period from 1940 to
1960 and so there was a huge influx of
immigrants, thus drastically lowering
occupational diversity during this time.
After 1960, however, occupational
diversity has been on an upward trend as
Puerto Ricans have successfully entered a
variety of occupations.
Filipino fractionalization indicates
that the probability of randomly drawing
2 Filipinos that were in different
occupations was about 49% in 1910.
Occupational diversity within the group
increased over time with the Filipinos
being represented in 5 occupations in
1910 to 144 occupations in 2000. Thus,
the probability that Filipinos were in
different occupations in 2000 was about
96%. Figure 8 shows that Filipinos were
only about half as occupationally diverse
as natives in 1910, but were actually .7%
more diverse than natives in 2000.
Filipinos
In 1898, the United States
defeated Spain in the Spanish American
War and the Philippines became a U.S.
territory, so the U.S. began to recruit
Filipinos into the Navy. Between 1903
and 1934, 500,000 Filipino students came
to the U.S. to study in American schools;
121
.5
.6
relative occ diversity
.7
.8
.9
1
Occupational Diversity: Filipinos to Natives
1900
1920
1940
1960
Census year
1980
2000
Figure 8
Before 1910 and the influx of
Filipino immigrants after the Spanish
American war, the only occupation that
Filipinos were in was farmers. Thus, they
had no occupational diversity. However,
on the graph, we see a large jump in
diversity after this time as Filipinos
flooded into America as students
especially in the second and third
decades of the 20th century, and began to
enter into an increasingly diverse array of
occupations thereafter. When the U.S.
government limited the immigration of
Filipinos in the 1930s, we see
occupational diversity leveling out for
the time being. Diversity continues on
an upward trend especially after the
1960s when a large wave of highly skilled
immigrants was allowed into the U.S.
and thus went into a wide scope of
occupations.
Cubans
In the latter part of the 1800s and
early 1900s, many Cubans moved freely
between Cuba and the U.S. The two
countries traded sugar, coffee, and
tobacco. Cigar factories from Cuba began
to move to Florida (“Countries and Their
Cultures”). There have also been at least
four
distinct
waves
of
Cuban
immigration to the United States since
1959. The first of these recent migrations
began immediately after Castro's victory
and continued until the U.S. government
imposed a blockade of Cuba at the time
of the Cuban missile crisis (“Countries
and Their Cultures”). The second major
migration started in 1965 and continued
through 1973. Cuba and the United States
agreed that Cubans with relatives
residing in the United States would be
transported from Cuba. The third
122
migration, known as the Mariel Boat Lift,
occurred in 1980 after Fidel Castro
permitted Cubans residing in the United
States to visit relatives in Cuba
(“Countries and Their Cultures”).
Economic conditions have worsened
since the fall of Cuba's principal
economic supporter, the Soviet Union,
and so in recent times many Cubans have
left Cuba in makeshift boats for Florida.
Over the years, just as the migration
“push factors” have changed, so has the
composition of the migrant population.
While the earliest migrants were drawn
from
the
highly
educated
and
conservative middle and upper classes,
more recent migrants have been poorer
and less educated. However, for the most
part, the Cuban American community
today is well assimilated in the United
States. Cuban Americans enjoy greater
economic security than other Hispanic
groups and the community as a whole is
known to be highly educated (“Countries
and Their Cultures”).
Cuban fractionalization indicates
that the probability of randomly drawing
2 Cubans that were in different
occupations was about 50% in 1850.
Occupational diversity within the group
increased over time with the Cubans
being represented in 2 occupations in
1850 to 139 occupations in 2000. Thus,
the probability that Cubans were in
different occupations in 2000 was about
95%. Figure 9 shows that Cubans were
only at 56% of the occupational diversity
level of natives in 1850, but were up to
99% of the native diversity level in 2000.
On the graph, we see low
occupational diversity in the period from
1850 to 1900 which was most likely due to
the fact that most Cubans were
.5
.6
relative occ diversity
.7
.8
.9
1
Occupational Diversity: Cubans to Natives
1850
1900
1950
Census year
Figure 9
123
2000
concentrated in the trading business
during this time. After the Spanish
American war when Cuba became a
protectorate of the U.S., the first real
wave of Cubans migrated to the U.S.,
thus bringing Cubans from a diverse
array of occupations into the U.S. We
can see a slight decline in 1960 which is
aligned with the election of Castro in
Cuba and the influx of immigrants that
occurred as a result. Most Cuban
immigrants have been highly educated
and thus appear to have spread out
across occupations quite nicely relative
to natives.
passage of the Chinese Exclusion Act of
1882, the first race-based immigration
law (Takaki 1998, 131). Thousands of
Chinese left their homes to become
contract laborers in the American West.
They were recruited to extract minerals
and construct a vast railroad network
during the Gold Rush period (Takaki
1998, 129). Throughout most of the
second period (the period of exclusion;
1882-1965), only diplomats, merchants,
and students and their dependents were
allowed to travel between the United
States and China (“Countries and Their
Cultures”). The Civil Rights movement in
the 1960s finally ushered in a new era,
the third period in Chinese American
immigration history. Chinese Americans
were liberated from a structure of racial
oppression. Under new immigration
laws, thousands of Chinese came to the
United States each year to reunite with
their families.
Chinese
In many respects, the motivations
for Chinese to come to the United States
are similar to those of most immigrants;
some came to the United States to seek
better economic opportunities, while
others were compelled to leave China
either as contract laborers or refugees.
However, their collective experience as a
racial minority, since they first arrived in
the mid-nineteenth century, differs
significantly
from
the
European
immigrant groups and other racial
minorities. Chinese were singled out for
discrimination through laws enacted by
the states in which they had settled.
Chinese immigration can be roughly
divided into three periods: 1849-1882,
1882-1965, and 1965 to the present. The
first period, also known as the first wave,
began shortly after the Gold Rush in
California and ended abruptly with the
Chinese
fractionalization
indicates that the probability of
randomly drawing 2 Chinese that were in
different occupations was about 40% in
1850. Occupational diversity within the
group increased over time with the
Chinese being represented in 2
occupations in 1850 to 138 occupations in
2000. Thus, the probability that Chinese
were in different occupations in 2000
was about 94%. Figure 10 shows that
Chinese were only at 46% of the
occupational diversity level of natives in
1850, but were up to 98% of the native
diversity level in 2000.
124
.4
relative occ diversity
.6
.8
1
Occupational Diversity: Chinese to Natives
1850
1900
1950
2000
Census year
Figure 10
On the graph, we see low
occupational diversity in 1850 because
most of the Chinese were concentrated
in occupations dealing with the Gold
Rush in California. We see occupational
diversity rise, but then fall beginning in
1880 due to the Chinese Exclusion Act
which only allowed certain occupational
groups of Chinese into the United States.
Occupational diversity then remains
fairly constant with a continued increase
especially after the 1960s due to the
Immigration Act which brought new
Chinese immigrants into the U.S. from a
diverse array of occupations.
in agriculture and construction (Takaki
1998, 281). However, several pieces of
legislation were introduced in the United
States such as the Congressional
Exclusion Law of 1923, which attempted
either to restrict the entry of Indians or
to deny them residence and citizenship
rights in America. Another significant
wave thus followed in the 1950s which
mainly
included
students
and
professionals. From 1965 onward, a
second significant wave of Indian
immigration began, spurred by a change
in U.S. immigration law that lifted prior
quotas and restrictions and allowed
significant numbers of Asians to
immigrate (Takaki 1998, 281). With the
technology boom of the 1990s, the largest
influx of Indians arrived between 1995
and 2000. The occupational profile
presented
by the
Asian Indian
community today is one of increasing
Indians
The first Asian Indians arrived in
America as early as the late nineteenth
century. The majority of Indians worked
125
here as private household workers. Thus,
they had no occupational diversity.
However, on the graph, we see a large
jump in diversity after 1900 as many
Indians came to America to work in
certain sectors of the labor market. We
diversity. Although a large number of
Asian Indians are professionals, others
own small businesses or are employed as
semi- or non-skilled workers.
Indian fractionalization indicates
.6
.7
relative occ diversity
.8
.9
1
Occupational Diversity: Indians to Natives
1900
1920
1940
1960
Census year
Figure 11
that the probability of randomly drawing
2 Indians that were in different
occupations was about 55% in 1900.
Occupational diversity within the group
increased over time with the Indians
being represented in 3 occupations in
1900 to 141 occupations in 2000. Thus,
the probability that Indians were in
different occupations in 2000 was about
91%. Figure 11 shows that Indians were
only at 59% of the occupational diversity
level of natives in 1900, but were up to
95% of the native diversity level in 2000.
The first immigrants from India
were recorded in 1870 and they were all
126
1980
2000
then see a significant drop in diversity in
the 1920s which was most likely due to
the Exclusion Act of 1923 which
prevented immigrants of different skill
levels from entering the U.S. We see a
slight decline in diversity in the 1990s,
when
a
significant
number
of
technologically-oriented
Indians
immigrated in reaction to the tech boom.
Overall, however, the Indian American
community
is
highly
diverse,
representing both skilled and unskilled
fields.
B2(∆ln(Immigrant/Native ratio)c,t) + Year
FEt + Country FEc + uc,t
Both regressions look at the
relationship
between
relative
occupational diversity and a ratio
variable measuring the number of
immigrants relative to the number of
natives, within the U.S. population, for
each country and decade. In equation (1),
I use the contemporary natural log of this
ratio, while in equation (2) I use the
lagged natural log of the ratio in addition
to examining the change in the natural
log of the ratio. The change ratio
represents the change in the immigrant
ratio from the previous decade in order
to capture large immigrant influxes. In
both regressions I control for year fixed
effects to control for occupational
changes over time as well as country
fixed effects to control for permanent
immigrant group differences. To do
country fixed effects, I omit Germany
because it has the most constant
occupational diversity across time.
Empirical Model
Through my case studies, I have
seen a general theme regarding the
apparent effect of large immigrant
influxes
on
relative
occupational
diversity as well as the idea that
immigrant
groups
are
inherently
different. The persisting problem with
my approach is that there are changes in
measurement over time. These issues can
be addressed in a model looking at the
effect of the immigrant to native ratio on
relative occupational diversity after
controlling for year and country fixed
effects. This would give me the ability to
observe if occupational diversity is
increasing due to other factors. The
regression analysis of this study is thus
based on the following empirical models
of relative occupational diversity:
Regression Results and Analysis
The following table shows the
regression results for equations (1) and
(2).
Table 1
ln(ratio)
=
B0
+
(1) RelativeDiversityc,t
B1(ln(Immigrant/Native ratio)c,t) + Year
FEt + Country FEc + uc,t
lagged ln(ratio)
(2) RelativeDiversityc,t = B0 + B1(lagged
+
ln(Immigrant/Native
ratio)c,t)
dy1 (1850)
.061***
(.009)
.035***
(.009)
.010
(.021)
∆ln(ratio)
dy2 (1860)
127
-.025
(.091)
dy3 (1870)
dy4 (1880)
dy5 (1900)
dy6 (1910)
dy7 (1920)
dy8 (1930)
dy9 (1940)
dy10 (1950)
dy11 (1960)
dy12 (1970)
dy13 (1980)
dy14 (1990)
dy15 (2000)
dbpl2 (Puerto Rico)
dbpl3 (Mexico)
dbpl4 (Cuba)
dbpl5 (Ireland)
dbpl7 (China)
dbpl8 (Philippines)
dbpl9 (India)
Number
of 110
Observations
-.045
(.085)
.035
(.087)
.031
(.086)
.194**
(.084)
.279***
(.084)
.275***
(.084)
.364***
(.084)
.357***
(.084)
.334***
(.083)
.337***
(.083)
.309***
(.083)
.291***
(.082)
.272***
(.082)
-.010
(.064)
-.095*
(.054)
.064
(.063)
-.009
(.052)
-.015
(.056)
-.010
(.067)
.040
(.072)
95
.075
(.074)
-.018
(.073)
R-squared
.6551 .5459
Note: *** Significant at 1% level; **
Significant at 5% level; * Significant at
10% level
.081
(.070)
.188***
(.071)
.186***
(.070)
.263***
(.073)
.263***
(.071)
.255***
(.069)
.265***
(.068)
.253***
(.067)
.243***
(.067)
.225***
(.068)
-.002
(.055)
-.097**
(.046)
.019
(.055)
-.011
(.045)
-.035
(.049)
.007
(.059)
.007
(.064)
These results highlight the effect
of the number of immigrants on relative
occupational diversity. It appears that
the ratio of immigrants to natives has a
statistically significant relation to the
occupational spread of immigrants. The
results show that as the number of
immigrants in the population increases
relative to the number of natives, the
relative occupational diversity increases.
Both of the coefficients on ln(ratio) and
lagged ln(ratio) are highly significant and
positive. The high r-squared values for
both regressions reaffirm the strength of
these results. For equation (1), the
coefficient on ln(ratio) shows that if the
immigrant to native ratio increases by
1%, then the relative occupational
diversity increases by .06 percentage
points. For equation (2), the coefficient
on lagged ln(ratio) illustrates that if the
immigrant to native ratio increases by
1%, then the relative occupational
diversity increases by .03 percentage
points. These findings support my
individual country case studies because
as the number of immigrants increases
over time, convergence with natives
actually does begin to occur as
immigrants’
occupational
diversity
trends upward. I have shown that this
convergence is statistically significant
and that relative occupational diversity is
indeed increasing. This can be seen
graphically as illustrated in Figures 12
and 13. A clearly positive linear
relationship can be seen between the
ratio and occupational diversity here.
128
Relationship between ln(ratio) and Rel Occ Div
-1
relative occupational diversity
-.5
0
(controlling for year FE and country FE)
-12
-10
-8
-6
-4
-2
ln(ratio)
Figure 13
Relationship between Lagged ln(ratio) and Rel Occ Div
-1
relative occupational diversity
-.8
-.6
-.4
-.2
0
(controlling for year FE and country FE)
-12
-10
-8
-6
lagged ln(ratio)
Figure 12
129
-4
-2
Relationship between Change in ln(ratio) and Rel Occ Div
-.6
relative occupational diversity
-.4
-.2
0
.2
(controlling for year FE and country FE)
-2
0
2
4
change in ln(ratio)
Figure 14
omitted group) in occupational diversity.
This considerable 9 percentage point gap
suggests that there is an innate country
difference that is affecting relative
occupational diversity. In regards to year
fixed effects, as time goes on the
coefficient values on the year dummies
increase and they become increasingly
significant. So, by including year fixed
effects I am able to control for higher
occupational diversity in later years.
In equation (2), the coefficient on
the change in ln(ratio) is insignificant.
Therefore, the change ratio is unrelated
to relative occupational diversity.
Through this regression, it can be seen
that there is no evidence for a
relationship between immigrant influxes
and relative occupational diversity. This
can be seen graphically as illustrated in
Figure 14. Thus, it appears that once I
control for year and country fixed effects,
immigrant influxes actually do not
matter.
Conclusion
This paper documents the
occupational diversity of immigrants in
the United States in the period between
1850 and 2000. The study looks at
immigrants from the following countries:
Mexico, Ireland, Germany, Puerto Rico,
Cuba, China, Philippines, and India. A
case
study
examination
of
the
In regards to the country fixed
effects which control for the size of the
immigrant population, the coefficient on
the Mexico dummy is significant in both
regressions. Mexican immigrants are
thus the only group to exhibit a
significant difference from Germans (the
130
occupational diversity indices for each
group and decade reveals that the
relative
occupational
diversity
of
immigrants has increased over time. By
looking at the immigration experience of
each group both quantitatively and
qualitatively, we can attempt to explain
the reasons for various trends in
occupational diversity across time. These
results do not explicitly show any one
significant overarching difference in
occupational diversity of immigrants that
can be explained by individual factors
such as race, language, or colonial status.
Both the Irish and Germans, who are
white, had fairly consistent high
occupational diversity across time.
Puerto Ricans were similar in this
respect, which was possibly due to their
initial colonial status. However, no
conclusive evidence can be drawn in this
regard. Overall, the occupational spread
of immigrants came to look more like
that of natives over time. In fact, most of
the immigrant groups had about the
same level of occupational diversity in
2000 and were able to break free from
certain
occupational
classifications.
Statistically,
relative
occupational
diversity has indeed increased. A clear
positive relationship exists between
relative occupational diversity and the
immigrant to native ratio. The Mexican
immigrant group is the only one that
appears to have an innate difference in
occupational diversity.
A recent New York Times article,
“Work Force Fueled by Highly Skilled
Immigrants”, released in April 2010
actually discusses the new analysis of
census data and highlights its findings
regarding immigrant occupations. The
data shows that “about two-thirds of all
immigrants in the country are nearly
evenly distributed across the job and
income spectrum” (Preston 2010). This
data therefore directly supports my
findings. The United States is receiving a
more
diverse
and
economically
important flow of immigrants than many
people realize. The fastest economic
growth over the past two decades has
actually been in cities that have received
“large influxes of immigrants with a mix
of
occupations”
(Preston
2010).
Occupational diversity is thus proven to
be highly significant in regards to
economic health. This data serves to
offer a new perspective on the
immigration debate as it refutes the
common perception that the U.S. is
overwhelmed with solely low wage
foreign-born workers. Thus this data,
which my results complement, can be
used to inform Congress as they seek to
reform immigration policy moving into
the future.
131
Appendix A
Professional
and
Technical
Occupations:
Accountants and auditors
Actors and actresses
Airplane pilots and navigators
Architects
Artists and art teachers
Athletes
Authors
Chemists
Chiropractors
Clergymen
College presidents and deans
Professors and instructors
Dancers and dancing teachers
Dentists
Designers
Dietitians and nutritionists
Editors and reporters
Engineers, aeronautical
Engineers, chemical
Engineers, civil
Engineers, electrical
Engineers, industrial
Engineers, mechanical
Engineers, metallurgical, metallurgists
Engineers, mining
Entertainers (n.e.c.)
Foresters and conservationists
Funeral directors and embalmers
Lawyers and judges
Librarians
Musicians and music teachers
Nurses, professional
Agricultural scientists
Biological scientists
Geologists and geophysicists
Mathematicians
Physicists
Clerical and Kindred Occupations:
Attendants and assistants, library
Attendants, physician's and dentist's office
Baggagemen, transportation
Bank tellers
Bookkeepers
Cashiers
Collectors, bill and account
Dispatchers and starters, vehicle
Express messengers and railway mail clerks
Mail carriers
Messengers and office boys
Office machine operators
Shipping and receiving clerks
Stenographers, typists, and secretaries
Telegraph messengers
Telegraph operators
Telephone operators
Ticket, station, and express agents
Clerical and kindred workers (n.e.c.)
Service Worker Occupations:
Attendants, hospital and other institution
Attendants, professional and personal
service (n.e.c.)
Attendants, recreation and amusement
Barbers, beauticians, and manicurists
Bartenders
Bootblacks
Boarding and lodging house keepers
Charwomen and cleaners
Cooks, except private household
Counter and fountain workers
Elevator operators
Firemen, fire protection
Guards, watchmen, and doorkeepers
Housekeepers and stewards, except private
household
Janitors and sextons
132
Optometrists
Osteopaths
Personnel and labor relations workers
Pharmacists
Photographers
Physicians and surgeons
Radio operators
Recreation and group workers
Religious workers
Social and welfare workers
Economists
Psychologists
Statisticians and actuaries
Miscellaneous social scientists
Sports instructors and officials Surveyors
Teachers (n.e.c.)
Technicians, medical and dental
Technicians, testing
Technicians (n.e.c.)
Therapists and healers (n.e.c.)
Veterinarians
Marshals and constables
Midwives
Policemen and detectives
Porters
Practical nurses
Sheriffs and bailiffs
Ushers, recreation and amusement
Waiters and waitresses
Watchmen (crossing) and bridge tenders
Service workers, except private household
(n.e.c.)
and Mobility over Time.” Journal of Labor
Economics 17:1 (49-79).
Bibliography
Alesina,
Alberto,
Arnaud
Devleeschauwer, William Easterly, Sergio
Kurlat, and Romain Wacziarg (2003).
“Fractionalization.” Journal of Economic
Growth 8:2 (155-94).
“Historical
Background.”
Latino
Education
Network
Service,
<http://palante.org/History.htm>.
Amuedo-Dorantes, Catalina and Sara de
la
Rica
(2006).
“Labor
Market
Assimilation of Recent Immigrants in
Spain.” IZA Discussion Paper, No. 2104.
“History of German-American Relations.”
U.S. Diplomatic Mission to Germany,
2008.<http://usa.usembassy.de/garelatio
ns8300.htm>.
“Countries
and
Their
Cultures:
Multicultural America.” Advameg Inc.,
2009.
<http://www.everyculture.com/multi/>.
“Immigration: The Journey to America.”
<http://library.thinkquest.org/20619/Iris
h.html>.
Green,
David
(1999).
“Immigrant
Occupational Attainment: Assimilation
“Mexican Immigrant Labor History.”
Public Broadcasting Service (PBS), 2010.
133
<http://www.pbs.org/kpbs/theborder/his
tory/timeline/17.html>.
Patel, Krishna and Francis Vella (2007).
“Immigrant
Networks
and
Their
Implications for Occupational Choice
and Wages.” IZA Discussion Paper, No.
3217.
Preston, Julia (2010). “Work Force Fueled
by Highly Skilled Immigrants.” The New
York Times. April 15, 2010.
Ruggles, Steven, J. Trent Alexander,
Katie Genadek, Ronald Goeken, Matthew
B. Schroeder, and Matthew Sobek.
Integrated Public Use Microdata Series:
Version
5.0
[Machine-readable
database]. Minneapolis: University of
Minnesota, 2010.
Takaki, Ronald (1998). A Larger Memory:
A History of Our Diversity, With Voices.
Boston: Little, Brown and Company.
Toussaint-Comeau, Maude (2004). “The
Occupational Assimilation of Hispanics
in the U.S.: Evidence from Panel Data.”
WP 2004-15, Federal Reserve Bank of
Chicago.
134
Southern Racism and Reality: A Case Study
Dan Lasker, Class of 2010
The notion pervasive in the popular media of the racist Southerner is a stereotype with serious
implications. Despite recent evidence showing that slavery and even plantations were plentiful in
the North into the early 19th Century, people commonly ascribe a higher level of racist
predisposition to individuals living in the South. This paper looks at the period between 1950 and
1970 to see if cities from the former Confederacy responded differently to the Civil Rights
Movement as compared to cities from the former Union along two different segregation indices.
The results show that while Southern cities had significantly higher levels of racial isolation, there
was no significant difference in the way Southern segregation changed over the course of the Civil
Rights Movement relative to the North.
has any roots in truth is an important
undertaking if we are serious about
moving past race.
Introduction
Nearly 150 years after the end of
the American Civil War, calls for
secession1 and celebration of the
Confederacy2 still find their way into our
common
culture.
While
racial
integration has become the legal norm
and tolerance socially expected, there is
still a common acceptance by most that
we are in no ways a “post-racial” society.
We can all pull up images of modern day
racist figures, and in popular culture, this
caricature is often of a person from the
South. While the Southern “good ol’ boy”
clinging to the Confederate flag, which
has been poked fun at by everything
from the Dukes of Hazzard to Saturday
Night Live, is a crude overgeneralization,
it is a pervasive stereotype in our society.
Uncovering whether or not this meme
This paper uses the Civil Rights
Movement as a case study to see if the
Southern cities responded differently
than the North to the period of great
social change in the 50s and 60s. Given
the magnitude of the changes that the
era brought about, it is easy to conceive
of an area with more deep-rooted racial
tension responding less dramatically
than one in which more open minded
people cling to the cause. Alternatively, it
could be argued that the rise in legal
measures to spread equality across races
affected cities that had previously higher
levels
of
racial
tension
more.
Conveniently, the key events of the Civil
Rights Movement (see Table 1) primarily
fall directly in between 2 Census years,
1950 and 1970, so by comparing
outcomes in 1970 with those in 1950 this
paper should be able to get a picture of
the short term impact of the Civil Rights
Movement
across
1
Linkins, Jason. “Glenn Beck: Secession or Suicide.”
The Huffington Post. 16 April 2009. 2 May 2010.
<http://www.huffingtonpost.com/2009/04/16/glennbeck-secession-or-s_n_187779.html>.
2
Erdman, Shelby Lin. “Virginia governor declares
April as Confederate History Month.” CNN Politics.
7 April 2010. 2 May
2010.<http://politicalticker.blogs.cnn.com/2010/04/0
7/virginia-governor-declares-april-as-confederatehistory-month>.
135
Table 1: Civil Rights Movement Timeline
1954
* U.S. Supreme Court declares school segregation unconstitutional in Brown
v. Board ruling.
1955
* Rosa Parks refuses to move to the back of a Montgomery, Alabama, bus as
required by city ordinance; boycott follows and bus segregation ordinance is
declared unconstitutional.
* Federal Interstate Commerce Commission bans segregation on interstate
trains and buses.
1956
* Coalition of Southern congressmen calls for massive resistance to Supreme
Court desegregation rulings.
1957
* Arkansas Gov. Orval Rubus uses National Guard to block nine black
students from attending a Little Rock High School; following a court order,
President Eisenhower sends in federal troops to ensure compliance.
* Four black college students begin sit-ins at lunch counter of a Greensboro,
NC, restaurant where black patrons are not served.
1960
* Congress approves a watered-down voting rights act after a filibuster by
Southern senators.
1961
* Freedom Rides begin from Washington, D.C., into Southern states.
* President Kennedy sends federal troops to the University of Mississippi to
quell riots so that James Meredith, the school's first black student, can attend.
* The Supreme Court rules that segregation is unconstitutional in all
1962
transportation facilities.
* The Department of Defense orders full integration of military reserve units,
the National Guard excluded.
1963
* Civil rights leader Medgar Evers is killed by a sniper's bullet.
* Race riots prompt modified martial law in Cambridge, Maryland.
* Dr. Martin Luther King Jr. delivers "I Have a Dream" speech to hundreds of
thousands at the March on Washington.
* Congress passes Civil Rights Act declaring discrimination based on race
illegal after 75-day long filibuster.
1964 * Three civil rights workers disappear in Mississippi after being stopped for
speeding; found buried six weeks later.
* Riots in Harlem and Philadelphia.
136
* March from Selma to Montgomery, Alabama, to demand protection for
voting rights; two civil rights workers slain earlier in the year in Selma.
1965 * Malcolm X assassinated.
* Riot in Watts, Los Angeles.
* New voting rights act signed.
1966
* Edward Brooke, R-Massachusetts, elected first black U.S. senator in 85
years.
* Riots in Detroit, Newark, New Jersey.
* Thurgood Marshall first black to be named to the Supreme Court.
1967
* Carl Stokes (Cleveland) and Richard G. Hatcher (Gary, Indiana) elected first
black mayors of major U.S. cities.
* Martin Luther King Jr. assassinated in Memphis, Tennessee; James Earl Ray
later convicted and sentenced to 99 years in prison.
1968
* Poor People's March on Washington -- planned by King before his death -goes on.
regions. After difference-in-difference
estimation, there is no significant
difference between the South’s response
to the Civil Rights Movement relative to
the North in terms of segregation
outcomes, which suggests that a regional
difference in short-term response to the
Civil Rights Movement is not justified by
the data.
segregation and how segregation is
changing as the makeup of our country
changes. More relevant for my time
period is their unsupported assertion
that social changes do not affect
segregation patterns. By looking at a time
period of some of the greatest social
change in history, I can see if their
assertion is correct.
Literature Review
Rhode and Strumpf (2003) show
that models based on the assumption
that residential choice depends solely on
local public goods (Tiebout models) are
not consistent with long-run segregation
trends. This is important for my topic
because it demonstrates that local public
finance figures are not sufficient
explainers of segregation outcomes as
people who think that a lack of Tiebout
sorting is to blame for segregation might
claim. Therefore, if the South shows
signs of having a different segregation
differential between 1950 and 1970 it
While the segregation literature is
extensive, papers dealing with historical
case studies such as my own are harder
to find. Nevertheless, a thorough
literature
review
is
useful
to
contextualize my method and results,
even if my objective is less studied in the
rest of the literature.
Logan, Stults, and Farley (2004)
primarily focus on segregation between
1980 and 2000, providing an extensive
amount of background on the roots of
137
cannot easily be swept away by an
argument that a Tiebout model is
actually the simplest way of describing
what is going on.
results are
explainable.
more
believable
and
Tolnay and Beck (2002) puts the
move of blacks from South to North
during the twentieth century into
context and works to portray the blacks
who moved northward in a more
accurate manner. While a large
percentage of the black migrants were
former sharecroppers pushed out by the
demise of the plantation system, Tolnay
makes the point that these migrants were
also most likely to be the go-getters who
tended to be relatively more skilled than
their peers they were leaving behind.
This notion lends credence to the
difference-in-difference
estimation
method I later employ, as I am able to
pull out regional effects from the data, all
else equal, to get a broad picture of the
differences in segregation between the
North and South.
Bayer (2002) makes an interesting
case for why race itself may not be the
sole driver of segregation, but rather that
properly measured income, education,
and immigration status data can explain
away a lot of the perceived racially driven
segregation. While Bayer’s point has
more to do with the limitations of large
data pools (which unfortunately I am
stuck with since I’m working with data
from 60 years ago), he raises some
interesting points about the complexity
of how segregation emerges that could
lend credence to my zero result. His
controls primarily rely on fine-tuned
Microdata that have only become
available recently, but his intuition is
important in contextualizing my results.
Card, Mas, and Rothstein (2007)
builds off of Schelling (1971) in making
the case that extreme segregation that
marks many major US metropolitan
areas is not the result of discriminatory
institutions and practices, like real estate
steering, but actually individual choices
based in homophily, preference for like
individuals. They show that once the
black population eclipses a tipping point
around the 13% mark, whites move out in
great force, which will lead to increased
segregation.
Their
intuition
that
segregation is the result of some tacit
force beyond the control of institutions
seems to suggest that my interaction
coefficient
should
in
fact
be
insignificantly different from zero. In the
context of this and Bayer (2002) my null
Collins and Margo (2004) uses a
similar historical approach to the one I
want to employ to see how the race riots
of the 50s and 60s influenced property
value outcomes. They use an interesting
model
in
which
their
primary
independent variable is a “riot index”
that they regress on property values to
find, in their eyes, that race riots drove
down property values for blacks. While
again, their results are largely predicated
on a highly specialized data set and a
complex riot index, they show the
viability of a case study approach and the
magnitude of the effect of the social
upheaval during the 50s and 60s.
Massey and Denton (1993) uses a
series of studies with data mainly from
1970-1980 to conclude that blacks faced a
138
near-apartheid situation of persistent
high segregation, which was barely
responsive to improvements in their
socioeconomic standing. Unlike Hispanic
and Asian minorities, who experienced
substantially lower levels of residential
separation than did blacks, blacks
remained highly segregated even in
metropolitan areas where they had
relatively high income and education.
Reductions in black-white segregation
were found mainly in areas with small
black populations, suggesting that
population and black population (which
is implicit to my indices) are both going
to play a key role in segregation
outcomes (Massey and Gross 1991).
and more against segregation, blacks'
demand for housing in all-white or
overwhelmingly white neighborhoods is
weak, which helps to maintain
residential segregation. According to the
analysis he and his fellow coauthors
undertook, in Atlanta, Boston, Detroit,
and Los Angeles, 65% of blacks would
not move into an attractive, affordable
home in an all-white neighborhood. The
paper goes on to say that blacks typically
explain their decision not in terms of
racial solidarity, but rather of the cool
reception and hostility they expect from
white would-be neighbors.
Similarly, Ellen (2000) found that
blacks who moved avoided census tracts
in which less than 1 resident in 10 was
black in favor of similar census tracts
with greater representations of blacks.
The absence of blacks from a
neighborhood or a low representation of
blacks is, apparently, an important signal
to blacks not to seek new homes in that
neighborhood. This factor could slow
progress in regions with historically high
levels of segregation.
Schuman et al. (1998) pulled data
from the University of Chicago’s National
Opinion Research to portray the
changing of racial attitudes around my
time of interest. When national samples
of whites were first asked by the Center
in 1963 whether they agreed or disagreed
with the statement that "White people
have a right to keep blacks out of their
neighborhoods if they want to, and
blacks should respect that right," 60%
agreed. When the question was asked in
1980 and 1996, the percentages of whites
who agreed with that principle of
segregation dropped to 34% and 13%,
respectively.
But segregation still
persists, and this paper examines
whether it persisted throughout the time
period of social unrest differently in
different parts of the country.
In the most seminal work dealing
with historical segregation, Cutler,
Glaeser, and Vigdor (1999) build upon
the finding in Cutler and Glaeser (1997)
that segregation is bad for nearly every
outcome variable for blacks and provide
a thorough examination of segregation in
the 20th Century. Even though historical
contextualization was only a part of their
lengthy paper that dealt more extensively
with the ghetto, their data is the most
cohesive set of historical segregation data
available. The paper motivated my
resolve to begin with 1950, instead of
trying
to
do
any
additional
Farley et al. (1997) countered that
paper’s assertion that segregation is
improving dramatically by looking at the
situation from the other side. He makes
the case that while whites may be more
139
contextualization, as pre-World War II
data for the South showed signs of being
skewed by ward-based segregation
practices. That is, government officials
would draw ward lines to segregate their
respective communities to garner favor
among
constituents.
They
also
convincingly showed that dissimilarity
and isolation indices are the most
consistent measures of segregation and
the easiest to come by, so they are the
dependent variables in my model.
1890, only cities with over 25,000
inhabitants had ward reports, so cities
with less than 25,000 inhabitants were
omitted. From 1900 onward, the
population cutoff was 50,000. All cities
sampled had both dissimilarity and
isolation indices available. Tract
reporting began with a select group of
large cities in 1940 and 1950, with Census
tracts being defined as geographic units
containing 3000 to 5000 individuals.
Because segregation measures are most
meaningful when the black population is
sizable, only cities or MSAs with at least
1,000 blacks were sampled. The number
of cities in the sample increased greatly
between 1950 and 1960 because tracts
became ubiquitous and population
growth brought a lot of MSAs over the
selection threshold. While this disparity
in data is not ideal for a case study like
mine, I am confident that my differencein-difference approach should resolve
most of the selection issues.
As I mentioned, this paper
primarily looks at the difference between
the two time periods that envelope the
Civil Rights movement, which is most
commonly appropriated the dates of
1954-1971. From a historical point of view
this is an interesting period to study
because any regional differences might
be explained by one region’s higher level
of acceptance of the civil rights message
or the changing times in general. Cutler,
Glaeser, and Vigdor (1999) did some
similar peering into regional effects over
time as controls for their larger
regressions, but they did not clearly
define their regions and brought in no
interaction term to capture the
difference in outcomes between the
regions. They did however show that
there clearly were regional effects in 1970
and 1990.
Segregation is measured by both a
dissimilarity index and an isolation index
as is done in Cutler, Glaeser, and Vigdor
(1999). The Dissimilarity Index was
proposed by Duncan and Duncan (1955)
and refined in Tauber and Tauber (1965)
and represents whether or not blacks
disproportionately reside in certain areas
of a city relative to whites:
Data Description:
The data set used is a modified
version of the data used by Cutler,
Glaeser, and Vigdor (1999). Their overall
data was drawn from 1890 to 1990 Census
data. To be included in their sample, a
city or MSA had to have ward or tractlevel data available in a given year. In
where blacki represents the number of
blacks in area i, blacktotal the total
number of blacks in a city, nonblacki the
number of non-blacks in area i, and
nonblacktotal the total number of nonblacks in a city. The index ranges from 0
to 1, and can thought of as the percent of
140
for a neighborhood with like peers,
which may occur outside of race
(McPherson, Smith-Lovin and Cook
2001). Just because blacks and non-blacks
do not live close to one another, it does
not mean that they do not interact with
one another (Blau 1977), so coming up
with a way to quantify interaction is
important. Using the model from Bell
(1954) as Cutler, Glaeser, and Vigdor
(1999) do, an isolation index can be
constructed as:
the black population that would need to
move to bring about an even distribution
of blacks and non-blacks. That is, a
dissimilarity index of 0.74 (the mean in
Table 2) means that 74 percent of blacks
would need to relocate to bring about an
even distribution. This is a
considerably high index (Massey and
Denton 1993) and may point to some
upward selection bias in the data
sampling that I will discuss further later
in the paper.
While the Dissimilarity Index can
give an idea of the geographic densities
of different populations, it cannot
measure the level of contact that blacks
and non-blacks have with one another.
To get at the root of racial tension, it
seems that such an Isolation Index would
be
more
helpful since Table 2: Summary Statistics (388 observations)
the
Mean
Dissimilarity Variable
Index
is Dissimilarity Index
0.740
more likely
to be heavily Isolation Index
0.428
influenced
0.291
by people’s Former Confederate
homophilic
Former Union
0.483
preference
Std. Dev.
0.103
0.215
0.455
0.500
Former Border
0.047
0.211
Former Western Territories
0.174
0.380
Log Tract Size
2.413
0.856
Log Population
12.752
1.005
Note: Log Tract Size = ln (Size of City / Number of
Tracts)
Log Population is in 1,000,000s.
141
Figure 4: Map of Black Density
where, again, blacki represents the
number of blacks in area i, blacktotal the
total number of blacks in a city,
nonblacki the number of non-blacks in
area i, nonblacktotal the total number of
non-blacks in a city, personsi the total
number of people in a tract, and
personstotal, the total number of people in
a city. The intuition behind the model is
that a small number of blacks are going
to be less isolated, on average, than a
larger black population simply because
there are fewer blacks to be isolated with.
By starting with the percentage black of
the area occupied by the average black
and then subtracting the percentage
black of the city as a whole, the index can
eliminate the effect coming from the
overall size of the black population in the
city. The indexing denominator makes it
so that the adjusted index ranges from
zero to one. Cutler, Glaeser, and Vigdor
(1999) say that a city has a ghetto if the
Dissimilarity Index is greater than 0.6
and the Isolation Index is greater than
0.3. Looking at the means in Table 2, this
means that the average city in my sample
has a ghetto. Other summary statistics
show the disparity between the number
of Former Union cities (48.3%) and
Southern cities (29.1%).
With the Cutler, Glaeser, and
Vigdor data at my disposal, I went
through all of the cities and determined
whether each was formerly in the Union,
in the Confederacy, or a Border state
based on the geographic boundaries on
the map in Figure 3. If none of the above
(i.e. the city was in a state colored white
in Figure 3) I dropped the data for the
purposes of this study due to my focus
on the difference between cities in
Former Union and Confederate states.
142
Figure 3: Civil War Boundaries
As Figure 4 clearly shows the
population distribution in the South is
very different than it is in the North1.
While the South has more blacks, they
are spread across the agrarian Southern
landscape, while in the North blacks are
focused in the cities. This could lead to
sample selection bias since only MSAs
with over 50000 people and 1000 blacks
are included in the analysis. So even
though the 2000 Census showed 54% of
restrictions gets around this selection
problem, but brings back the problems
related to the consistency of the
segregation indices that the imposition
of the selection criteria did away with.
Table 5 shows the movement over
time in the dissimilarity index for all 3
geographic regions. Former Border states
show virtually no change. Former Union
states show a 0.038 increase in the
dissimilarity index from 1950 to 1970.
Former Confederate states on the other
hand show a decrease of 0.035 in moving
from 1950 to 1970.
blacks lived in the South, a large deal
lived in areas that were not picked up in
1950, 1960, or 1970 given the restrictions
on the data selection. Lifting the
1
Even though the figure is from 2000, the
distribution has remained similar since the Civil War.
143
Table 5: Dissimilarity Index Trends
Table 6: Isolation Index Trends
Former Former
Former
Union Confederate Border
Former
Union
Former
Former
Confederate Border
Dism
1950
0.729
(0.128)
0.746
(0.090)
0.757
(0.095)
Isol
1950
0.406
(0.215)
0.600
(0.123)
0.583
(0.100)
Dism
1960
0.753
(0.092)
0.717
(0.105)
0.759
(0.074)
Isol
1960
0.365
(0.205)
0.526
(0.156)
0.483
(0.240)
Dism
1970
0.767
0.711
(0.096) (0.109)
0.756
(0.092)
Isol
1970
0.411
(0.206)
0.526
(0.154)
0.439
(0.288)
Note: Standard
parentheses
deviation
listed
in
Note: Standard
parentheses
deviation
listed
segregation, as the isolation index
increases by 0.005. The Isolation Index
for Former Confederate states falls by
0.074. It is interesting to note that in
1950 Former Confederate states had an
Isolation Index nearly 0.20 greater than
Former Union states. So while the two
converged between 1950 and 1970, the
.4
.6
dism
.8
1
Table 6 shows the movement in the
isolation index for all 3 geographic areas.
Former Border States here show a
dramatic decrease of 0.144 in terms of
isolation index from 1950 to 1970. Former
Union states again show an increase in
0
.2
in
.4
.6
.8
isol
Figure 7: Isolation and Dissimilarity Indexes – Former Confederate
states
144
1
Econometric Approach and Results:
high levels of isolation in the South could
not be overcome. It is important,
however, to keep in mind that the
isolation index more or less measures the
percentage of a population that is black.
With this in mind, the high initial level
of isolation in the South could simply
reflect more blacks living in the south at
that time and the trend over time could
merely reflect migration patterns.
Since implicit racial tension is hard
to quantify, I use a model where two
different measures of segregation serve
as my dependent variables to try to
capture
racial
division
through
segregation outcomes, which have been
empirically shown to have negative
consequences for blacks (Cutler and
Glaeser 1997).
Figures 7 and 8 are included to
map the difference between Former
Confederate states and Former Union
states. Former Confederate states tend to
have a higher Isolation Index relative to
their Dissimilarity Index while Former
Union states tend to have a higher
Dissimilarity Index relative to their
Isolation Index.
The basic model I employ is a
difference-in-difference model where:
.4
.5
dism
.6
.7
.8
.9
Dissimilarityc,t = β0 +
β1*FormerConfederatec +
β2*PostCivilRightst +
β3*(FormerConfederate*PostCivilRights)c
,t + εit
0
.2
.4
isol
.6
Figure 8: Isolation and Dissimilarity Indexes – Former Union states
145
.8
Or
dummy and the Post Civil Rights dummy
on the Dissimilarity Index. As you can
see, none of the coefficients are
significant, except for the constant,
which is statistically significant at the 1%
level. Of note is the fact that there is no
sign of the South responding less to the
changes from 1950 to 1970 in comparison
to the North, as the coefficient
corresponds to 3.2% less black people
having to move in the South to create an
equal distribution relative to the North.
Since this coefficient is not statistically
significant, it goes against my initial
supposition of cities in the former
Confederacy responding less favorably to
the Civil Rights Movement than Former
Union cities. It implies that there is no
statistically
significant
difference
between segregation responses across the
regions and, if anything, the South
improved slightly more than the North
did in terms of segregation across this
time
period.
Isolationc,t = β0 + β1*FormerConfederatec
+
+
β2*PostCivilRightst
β3*(FormerConfederate*PostCivilRights)c
,t + εit
With either segregation index as the
dependent variables, the parameter
estimate β1 can be interpreted as the
ceteris paribus effect that being in the
South has on segregation. β2 can be seen
as the ceteris paribus effect of moving
from 1950 to 1970 on segregation. β3 is
the difference-in-difference estimator,
the coefficient of interest. It says whether
or not the period from 1950-1970 (which
was dominated by the Civil Rights
Movement as Table 1 shows) affected
segregation differently for the South than
for other areas.
Table 9 shows the regression of
the Former Confederate dummy, the
Post Civil Rights dummy, and the
interaction of the Former Confederate
Table 9: Difference-in-difference Estimation for Dissimilarity
Coefficient
Standard
Error
95% Conf.
Interval
FormerConfederate
-0.022
0.016
[-0.053,
0.009]
PostCivilRights
0.020
0.015
[-0.009,
0.049]
FormerConfederate*PostCivilRights
-0.032
0.024
[-0.080,
0.016]
Constant
0.745***
0.009
[0.727,
0.763]
146
Table 10 shows the regression of
the Former Confederate dummy, the
Post Civil Rights dummy, and the
interaction of the Former Confederate
dummy and the Post Civil Rights dummy
on the Isolation Index. This time, the
Former
Confederate
dummy
was
statistically significant at the 1%
significant level. The coefficient means
that a city in a former Confederate state
has on average, a 14.8%-point higher
level of isolation, ceteris paribus. While
this does not say anything about a
difference in segregation outcomes
across the Civil Rights Movement (as
again, the interaction term coefficient is
insignificant), it does point to a possible
reason for the stereotypical perception of
the racist Southerner: there does seem to
be historically more racial isolation in
the South (in part because of the larger
black
population).
Table 10: Difference-in-difference Estimation for Isolation Index
Coefficient
Standard
Error
95% Conf. Interval
FormerConfederate
0.148***
0.030
[0.090, 0.206]
PostCivilRights
0.019
0.028
[-0.035, 0.074]
FormerConfederate*PostCivilRights -0.037
0.046
[-0.127, 0.054]
Constant
0.017
[0.361, 0.430]
0.395***
With no significant difference-indifference estimates emerging from my
initial specification, I added 2 controls to
see if my zero result would still hold: Log
of Population and Log of Tract Size, with
population in millions and tract size
equal to the area of the city in square
miles divided by number of tracts in year
t.
size, and Log Population on the
Dissimilarity
Index.
The
Former
Confederate dummy was statistically
significant at the 5% level with a
coefficient that corresponds with the
South having a 3.1%-point lower level of
dissimilarity on average relative to the
North, ceteris paribus. Log Tract Size was
positive and statistically significant, not
surprising considering that smaller tracts
seem intuitively more likely to have
higher levels of dissimilarity since small
neighborhood clusters, which generally
are filled with similar residents, carry
more weight. Log Population and the
Table 11 shows the regression of
the Former Confederate dummy, the
Post Civil Rights dummy, the interaction
of the Former Confederate dummy and
the Post Civil Rights dummy, Log Tract
147
constant term were also significant,
while the difference-in-difference
estimate become even less significant
with the addition of the controls.
Table 11: Difference-in-difference Estimation for Dissimilarity with Controls
Coefficient
Standard Error
95% Conf. Interval
FormerConfederate
-0.031**
0.015
[-0.060, -0.002]
PostCivilRights
0.004
0.013
[-0.021, 0.031]
FormerConfederate*PostCivilRight
s
-0.017
0.022
[-0.059, 0.026]
Log Tract Size
0.031***
0.008
[0.016, 0.046]
Log Population
0.055***
0.006
[0.043, 0.068]
Constant
-0.031***
0.095
[-0.220, 0.156]
Table 12 shows the regression
of the Former Confederate dummy,
the Post Civil Rights dummy, the
interaction of the Former Confederate
dummy and the Post Civil Rights
dummy, Log Tract Size, and Log
Population on the Isolation Index. As
in Table 12, the coefficient on the
Former Confederate dummy is
significant at the 1% level, suggesting
that, on average, a city in a former
Confederate state has a 17.4%-point
higher level of isolation than a state in
the North, ceteris paribus. The
coefficients on Log Tract Size and Log
Population are significant and
positive
once
again.
Most
interestingly,
the
difference-indifference estimator is once again
statistically insignificant, as it has
been for all of the regressions. This
suggests that cities from Former
Confederate states did not, on
average, have different segregation
outcomes between 1950 and 1970
relative to cities in Former Union
states.
148
Table 12: Difference-in-difference Estimation for Isolation with Controls
Coefficient
Standard
Error
95% Conf. Interval
FormerConfederate
0.174***
0.028
[0.119, 0.229]
PostCivilRights
-0.009
0.025
[-0.057, 0.040]
FormerConfederate*PostCivilRights -0.035
0.041
[-0.116, 0.045]
Log Tract Size
0.035
0.015
[0.006, 0.064]
Log Population
0.126***
0.012
[0.102, 0.150]
Constant
-1.299***
0.180
[-1.654, -0.945]
Discussion
The primary problem in making a
causal statement about the impact of the
Civil Rights Movement across geographic
lines is the fact that “racism” is not
something that can be measured.
Segregation seems to be a reasonable
enough proxy for racial tension, and has
been shown to be bad for black outcomes
(Cutler and Glaeser 1999), but it is also
influenced by countless other factors
(Bayer 2002) so it is far from perfect. One
such other factor is the fact that
segregation change is predicated on
people changing homes, so while people
may be eager to sort in a Tiebout model
after hearing the message of Civil Rights,
real people do not just up and relocate.
Moving is usually driven by some
external factor, and once someone has
decided to move, then and only then, can
that person even begin to think about
neighborhood choice. I did not extend
my time period, however, because
opening up the time series to other
decades would open it up to countless
other drivers of segregation outcomes. It
is fortunate that the Civil Rights
The goal of this paper was to see if
former Confederate cities responded
differently to the Civil Rights movement
than Northern cities in terms of
segregation outcomes. While I will
discuss several reasons why my
difference-in-difference
estimation
method is not flawless in trying to make
a causal statement about segregation
across the Civil Rights Movement, it is
important to consider the possibility
that, in the context of my result
continually being insignificant from zero,
the South responded as well or even
more favorably to the Civil Rights
message than the North did. It certainly
had some ground to make up with
regards to racial isolation, so the
consistently insignificant and slightly
negative coefficients may be close to the
truth. What they mean, however, must
be put into context.
149
Movement fit nicely in between two
Census years, but even so, it would be
foolish to think that the Civil Rights
Movement would be the sole driver of
outcomes even while it was in progress,
as things do not happen in a vacuum.
was no statistically significant difference
in segregation outcomes across the Civil
Rights Movement between cities in
Former Confederate states and Former
Union states.
Conclusion
Another issue to consider in that
same vain is the limited nature of
difference-in-difference
estimation.
Bertrand, Duflo, and Mullainathan
(2004) say that most difference-indifference papers should not be trusted
because they rely on many years of data
and focus on serially correlated
outcomes. They claim that almost all
these papers ignore the bias in the
estimated standard errors that serial
correlation
introduces,
which
is
especially
troubling
because
the
independent variable of interest in
difference-in-difference
estimation
(usually some historical event) is itself
very serially correlated, which will only
exacerbate the bias in standard errors. To
illustrate the severity of the issue, they
randomly generated placebo laws (to
serve as historical pivot points) and used
OLS to compute the difference-indifference estimate of its “effect” as well
as the standard error for this estimate.
They found the standard errors to be
severely biased: with about 20 years of
data, difference-in-difference estimation
finds an “effect” significant at the 5%
level for up to 45% of the laws which, by
the nature of the experiment, in reality
had no effect. If anything, this bolsters
my insignificant finding as the model
seems designed to want to find
significance
in
the
difference-indifference estimator. In that context,
finding a consistently insignificant result
seems to conclusively show that there
While
finding
statistical
significance is often the goal going into
an academic paper, the lack thereof in
this paper is perhaps just as interesting.
Popular culture and Bertrand, Duflo, and
Mullainathan (2004) would seem to
suggest a significant difference-indifference estimator emerging, but the
data does not agree. There is no clear
difference between the change in
segregation outcomes between cities in
the former Confederacy versus cities in
the former Union during a period of
great social change and I think that this
result is important enough to be
considered when we as a culture
collectively develop stereotypes based on
where someone is from.
One takeaway point from this
study that I think could be useful for
future work is the high level of
significance on the Former Confederate
variable in the Isolation index regression.
It
seems
that
while
historical
dissimilarity trends are relatively close
between the North and South, the South
systematically has a higher Isolation
Index so it could have potential as an
instrumental variable in a future 2SLS
model where percentage of a city that is
black is not, on its own, suitable as an
instrument. Otherwise, I hope this paper
can provide some clarity on the murky
issue of how this country is still divided
along Civil War lines, if it still even is.
150
Works Cited
Cutler, D.M., E.L. Glaeser, and J.L.
Vigdor. 1999. "The Rise and Decline of
the
American Ghetto." Journal of
Political Economy 107:455-506.
Bayer, Patrick, Robert McMillan and Kim
Reuben. 2002. "What Drives Racial
Segregation? New Evidence using Census
Microdata." Working paper.
Duncan, Otis Dudley and Duncan,
Beverly. "A Methodological Analysis of
Segregation Indexes."
American
Sociological Rev. 20 (April 1955): 210-17.
Bell, Wendell. "A Probability Model for
the
Measurement
of
Ecological
Segregation." Soc. Forces 32 (May 1954):
357-64.
Ellen, I. Gould. 2000. Sharing America's
Neighborhoods: The Prospects for Stable
Integration. Cambridge, MA: Harvard
University Press.
Bertrand, M.; Duflo, E.; Mullainathan, S.
"How
Much
Should
We
Trust
Differences-in-Differences
Estimates?"
The Quarterly Journal of Economics, v.
119, n. 1, p. 249275, February 2004.
Farley, R., E.L. Fielding, and M. Krysan.
1997. "The Residential Preferences of
Blacks and Whites: A Four-Metropolis
Analysis." Housing Policy Debate 8:763800.
Blau,
Peter
M.
Inequality
and
Heterogeneity: A Primitive Theory of
Social Structure.
New York: Free
Press, 1977.
Logan, J., B. Stults, and R. Farley. 2004.
"Segregation of Minorities in the
Metropolis: Two Decades of Change."
Demography 41:1-22.
Brasington, David M. & Donald R.
Haurin. "Parents, peers, or school inputs:
Which
components of school
outcomes are capitalized into house
value?"
University
of
Cincinnati,
Economics Working Papers Series, 2008.
Massey, D.S. and N.A. Denton.
1993. American Apartheid: Segregation
and the Making
of the Underclass.
Cambridge, MA: Harvard University
Press.
Card, David; Mas, Alexandre and
Rothstein, Jesse (2007), “Tipping and the
Dynamics of Segregation
in
Neighborhoods and Schools,” Working
Paper, UC Berkeley Center for Labor
Economics.
Massey, D.S. and A.B. Gross. 1991.
"Explaining Trends in Racial Segregation,
1970- 1980." Urban
Affairs
Quarterly 27:13-35.
McPherson, M., L. Smith-Lovin and J.M.
Cook, "Birds of a Feather: Homophily in
Social Networks," Annual Review of
Sociology, 27 (2001), 415-444.
Collins, William J. and Robert A. Margo.
“The Economic Aftermath of the 1960s
Riots: Evidence from Property Values.”
NBER Working Paper No. 10493, May,
2004.
Rhode, P.W. and K. S. Strumpf (2003).
“Assessing the Importance of Tiebout
Sorting: Local Heterogeneity from 1850
Cutler, D., Glaeser, E., 1997. “Are ghettos
good or bad?” Quarterly Journal of
Economics 112, 827–872.
151
to 1990,” American Economic Review 93,
1648– 1677.
Schelling, T. 1971. “Dynamic Models of
Segregation.” Journal of Mathematical
Sociology 1: 143-186.
Schuman, H., C. Steeh, L. Bobo, and M.
Krysan. 1998. Racial Attitudes in America:
Trends and Interpretations, rev. ed.
Cambridge, MA: Harvard University
Press.
Taeuber, Karl E., and Taeuber, Alma F.
Negroes in Cities: Residential Segregation
and Neighborhood Change. Chicago:
Aldine, 1965.
Tolnay, S. E., and E. M. Beck (2002)
‘‘Race, regional origin, and residence in
northern cities at the beginning of the
Great Migration.’’ American Sociological
Review
67:
456–75.
152
Segregation and Incarceration: How Life in the Ghetto
Leads to Life in Prisons for Young Black Men
Jess Mawhirt, Class of 2010
This paper seeks to further understand the causes of high crime rates among young black
men. I have extended the work of Cutler & Glaeser (1997) who determined that blacks in
highly segregated metropolitan statistical areas (MSAs) have worse outcomes than blacks
in less segregated MSAs. One outcome not included in the study is the likelihood of being
incarcerated. By obtaining state-level segregation measures and individual level
incarceration data from the U.S. Census (1980-2000), I have determined that there is a
correlation between these two measures; that black men age 18-65 who live in more
segregated states have a higher probability of being in jail than those living in less
segregated states. By using an Instrumental Variable, I can determine that this is in fact a
causal relationship. Additionally, I examine the mechanism through which segregation
influences life incarceration risks of blacks and how improving these areas can have a
significant impact on life-trajectories for young black men.
Introduction:
disproportionate incarceration of blacks
(Hartney, 2006.)
Incarceration levels for blacks
have increased dramatically in the past
few decades to astonishing levels. For a
black man, the estimated risk of going to
prison at some point in his life is 28.5%
while that statistic is only 4.4% for white
men (Pettit & Western, 2004.) The
percentage of the black male population
that is in prison at any given time has
been rising faster than any other group
(see Figure 1 from the U.S. Department of
Justice.) The statistics for black men born
between 1965 and 1969 (now ages 41-45)
are shocking; 30% of black men without
college education and 60% of black men
without a high-school education had
been to prison by 1999. Only 3% of white
men born in that same period had served
time by 1999. (Pettit & Western, 2004.)
The U.S. in general has staggering
incarceration rates; the highest in the
world and four times the world average,
which is in large part due to
The causes and consequences of
this trend are extremely important to
study and understand. Having a prison
record
significantly
impacts
life
trajectories, including the ability to find
and keep a job. It also further stigmatizes
poor and minority individuals and their
neighborhoods as being dangerous. In
addition, recidivism rates are extremely
high; one study finds rates close to 70%
within 3 years of release.1 Many studies
have suggested that young black men
have turned to crime as a result of a lack
of
education,
fewer
economic
opportunities, lower wages, high returns
to
criminal
activity
and
social
interactions with criminals, (Petit &
Western, 2004; Freeman, 1996; Grogger,
1998; Case & Katz, 1991; Glaeser,
1
153
Langan, 2002
Sacerdote & Scheinkman, 1996.)
However, it seems plausible that many of
these causes could in fact be symptoms
of segregation.
incarceration risks for black men. I have
incorporated an instrumental variable,
using a shift-share methodology to
predict segregation rates that would
occur if people did not move across state
borders. This instrumental variable will
be necessary to establish a causal
relationship. In addition, I have created a
mechanism
through
which
this
relationship could potentially operate.
Through previous studies, I show that
segregation leads to low spending on
education which leads to low wages and
economic opportunities that lead to high
crime rates. I show that this mechanism
is a plausible explanation for why
younger black men from more
segregated states are more likely to go to
jail than black men from less segregated
states.
If incarceration rates are higher
for black men living in segregated areas
than for black men living in less
segregated areas, it could be true that
segregation is responsible for the above
mentioned causes leading to worse
outcomes. For examples, lower wages
could certainly lead to high crime for any
group and in any area, but if lower wages
combine with other problems that are
specific to segregated areas to create a
higher risk of incarceration, then
segregation itself could be to blame.
In order to explore this thesis, I
have formulated a regression to quantify
the relationship between segregation and
This paper is organized as follows:
In the next section I will present the
background
literature
for
both
154
incarceration studies and segregation
studies. Section three describes the data
acquisition
and
explains
the
measurements used. Section four
describes summary statistics. Section five
includes the primary model and the
initial results of that model. In section six
I discuss the controls and results. In
section seven, I discuss the mechanism
through which the causal relationship
operates and in section eight I conclude
the paper.
earning potential far exceed the
punishments. He argues that because
real earnings for the least educated fell
significantly from the 1970s to the 1990s,
uneducated individuals turned to the
drug trade in response to the increased
demand for drugs. Incomes from selling
drugs or other criminal activity increased
relative to legitimate earnings such that
youths and adults could easily make $1030 per hour. He argues that the
opportunity cost of crime is lower for the
uneducated, especially for those growing
up in low-income neighborhoods where
there is less stigma attached to criminals.
He estimates the cost of imprisonment is
significant for the rest of society to bear,
with an average annual cost per prisoner
of $22,000, which could arguably go to
better uses if in fact incarceration does
not deter crime.
Background Literature and Trends
There is extensive literature on
segregation measurements, trends and
consequences as well as incarceration
causes, trends and consequences. Much
of the crime literature that focuses on
race differentials suggest causes that
could be particular problems of
segregated neighborhoods.
Grogger (1998), like Freeman,
finds that incentives toward crime are
higher when wages are low and that the
rising crime rates of youth in the 1970’s
and 80’s was probably a response to
falling wages during that time period.
However, he finds that because blacks
earn less than whites, they participate in
crime to a greater extent. He argues that
the black-white wage gap explains a
substantial amount of the differential in
crime between blacks and whites.
Because blacks earn a lower wage than
whites, they have a lower opportunity
cost to committing crime in the same
way that young people (who earn lower
wages than older people) have a lower
opportunity cost.
Petitt
&
Western
estimate
comparative incarceration risks by race,
education level, and age (determined by
birth year groups.) They calculate the
cumulative risk of imprisonment and
find that men have a much higher
chance of going to jail if they are
uneducated, young, and black and briefly
discuss the possibility of these findings
being related to living in poor urban
neighborhoods. They find that a black
male high-school drop-out had a 61.8%
chance of being either in jail or dead by
the age of 34, compared to only 14.0% for
white male drop-outs.
Freeman (1996) assesses why this
might be the case for uneducated men.
He finds that the rewards of crime for
those who have low skills and low
Both Freeman and Grogger are
supported by Licher (1988) who finds
155
that underemployment increases in the
1970’s and 80’s for young blacks and
whites with low-education. However,
this effect is worsened for blacks due to
urban racial polarization. In addition,
Glaeser, Sacerdote and Scheinkman
(1996) find that social interactions play a
role in encouraging younger urban
individuals to commit crimes, and that
the presence of intact family units
lessened this effect, suggesting that areas
with a prevalence of single-parent or
incomplete households might be more
susceptible to peer-pressure related
criminal behavior among teens and
young people. Case & Katz (1991) find
similar results; the behavior and
examples of older family members have
significant impacts on the actions of the
youth in terms of criminal behavior, drug
abuse, schooling etc. They find that for
young people living in neighborhoods
where other youths are engaged in
criminal activities, the probability of
being involved in crime increases. In
addition, Massey, Condran and Denton
(1987)
find
that
in
segregated
environments,
even
higher-income
blacks have lower average SAT scores
than low-income whites because of the
neighborhood effects of being educated
in poor schools with less motivated classmates. These studies have shown that
many of the attributes we ascribe to
segregated areas do in fact cause crime.
My study will determine if segregation
itself causes crime and incarceration.
Massey and Denton (1989)
introduce
five
major
ways
to
conceptualize measure and analyze
segregation levels. The five dimensions
are evenness, exposure, clustering,
centralization and concentration. I will
rely on evenness and exposure as they
are arguable better measures of
segregation. If a region is uneven, some
areas have a higher percentage of
minority residents than other areas. If it
lacks exposure, the chances for people of
different races to interact are slim.
Massey and Denton find that segregation
does exist, that 29% of all urban blacks in
the U.S. live in very highly segregated
MSAs, including the six most segregated
cities, Baltimore, Chicago, Cleveland,
Detroit, Milwaukee, and Philadelphia.
For my purposes, the states that contain
these hyper-segregated cities, Maryland,
Illinois, Ohio, Michigan, Wisconsin and
Pennsylvania should possess higher than
average state segregation levels.
Cutler and Glaeser (1997) advance
the literature by finding that segregation
seriously worsens outcomes for blacks.
While they consider the positive
externalities of ghettos, they conclude
that a one-standard deviation reduction
in segregation (or a 13% reduction) will
eliminate all differences in black and
white outcomes, including high school
graduation rates, idleness, earnings, and
the likelihood of being a single mother.
We already know that being a college
drop-out
can
have
serious
life
consequences, including imprisonment,
so it is reasonable to conclude that
segregation can likely be an indirect
cause of incarceration via education
levels. I will explore this possibility later.
Data Acquisition & Variables
I acquired U.S. decennial Census
data from 1980, 1990, and 2000 from the
Integrated Public Use Microdata Samples
156
(IPUMS-USA2). The analysis is restricted
to men age 18 to 65. I did not include
women because they make up a small
portion of people in jail or prison, for
example, less than 8% in 1997.3 I used the
IPUMS variable “race” (general version)”
to determine if an individual was black
(race=2) or white (race=1). I specifically
left out Asians and Hispanics from my
data because most (not all) of the
previous literature specifies segregation
as largely a black problem and because I
am predominantly interested in the lifecourse of blacks that can lead to gang
violence and imprisonment. Education
was measured by the IPUMS variable
“yearsed” where the min was 0 years and
the max was 17 years.
individuals with mental health issues
living in prison (283,800 people in 1998)
amounted to more than double those
living in actual mental health institutions
(less than 100,000.)4 On the whole
1,734,900 Americans were in jail or in
prison in 1997. If the proportion of those
in mental institutions to those in
correctional institutions is constant over
time, we can assume that those in mental
institutions only account for 5% of our
group quarters data. Even so, we must
keep this in mind when drawing
conclusions from the data.
State diversity can be measured
through
ELF
(Ethno-Linguistic
Fractionalization) (Easterly & Levine,
1997.) ELF ranges from 0 and 1 and in this
case measures the probability that two
randomly selected individuals in a state
belong to different racial groups. In a
highly diverse state, ELF is expected to
be high and vice-versa. The highest ELF
is 0.66 for California in 2000 and the
lowest ELF is 0.028 for Vermont in 1980.
Because there is no variable in the
U.S. Census to indicate whether or not a
particular individual is in jail at the time
of the census, I used the method used by
Borjas, Grogger & Hanson, 2006 to
isolate prisoners. The census has a
variable for people living in “Group
Quarters” (IPUMS variable GQ). These
people could be living in college dorms,
retirement homes, mental institutions,
jail, prison, or other group living
arrangements. Borjas, Grogger, and
Hanson determine that a person is
institutionalized (in a correctional or
mental or other type of institution) if gq
=2 or 4. (In my data, institutionalized
individuals have a gq value of 3 or 4.) It is
impossible to determine which of these
individuals are in prison or in mental
institutions. However, according to the
Department of Justice, in the 1990’s,
Segregation can be measured
through dissimilarity and isolation
measures (Massey & Denton, 1988a.)
Dissimilarity is a measure of “evenness,”
or the degree to which the percentage of
minorities in an area equals the
percentage of minorities in the larger
area. In this study, the variable disb is
measured from 0 to 1 and it represents
how well the percentage of blacks in the
zip codes of a state reflects the
percentage of blacks in that state. The
following is the equation for dissimilarity
used by Massey and Denton.
2
Ruggels, 2010
USDOJ, Correctional Populations of the
U.S., 1997
3
4
157
Karaim, 2002
inaccurately
represent
non-prison
populations. (For example, 30.5% of the
rural Brown County, IL population is
incarcerated. The county appears to be
diverse but 99.6% of blacks there were
actually transplanted inmates.)5 For
these
reasons,
segregation
and
incarceration cannot be measured on the
zip-code level in my analysis, but must
be aggregated to the state level. This
causes potential problems. Certain areas
of some states might not be segregated at
all, while cities in those states can be
hyper-segregated.
Where t and p are total population and
minority proportion of the zip codes and
T and P are the same measures for the
whole state.
Isolation measures “exposure,” or
the degree of potential contact between
minorities and majorities. In my dataset,
the variable isob is measured from 0 to 1
and measures the state average
probability that blacks will come into
contact with each other. As expected,
states with low black populations like
Vermont have low isob values. The
following is the equation, also from
Massey and Denton, for isolation.
In addition, it is important to
address the possibility of inter-state
prisoner transfers. If prisoners are not
incarcerated in their home-states, then it
would be difficult to draw any
conclusions from my models. Due to
over-crowding and security-risks, most
states can and do choose to transfer a
small number of prisoners to other
states. There are both regional
agreements between states as well as
cost-saving
contracts
between
correctional facilities. There are multiple
protections and stipulations that must be
in place before states can transfer
prisoners, for example, they must be
considered to be ‘rarely-visited’ or ‘longterm.’ (State restrictions vary, for
example, in the state of Alaska, no
Alaskan natives can be transferred, and
in Ohio, sex offenders and those awaiting
execution cannot be transferred.)6
Recently, however, Pennsylvania sent
2,000 prisoners to Michigan and Virginia
to ease the severe and extremely
expensive over-crowding in its prisons.
Where x and t are number of blacks and
the total population of zip-code i and
where X is the total population of the
state.
The segregation data was also
obtained from the U.S. Census using the
method seen in Massey, Denton, 1989.
The original race data was obtained by
zip-code. However, there is evidence that
prisoners are often moved to zip codes
other than their zip code of residence. In
fact there is increased frustration over
the gerrymandering of political districts
based on large prison populations.
Opponents argue that prisoners (who
cannot vote) from largely urban areas are
moved to rural area prisons and
5
6
158
Paley, 2010
Biasca, 2006
Figure 2:
However, this transfer agreement
of only 2,000 people was seen as
controversial.7 This is a good
indication that large transfers are
still rare.
Due to restrictions and
expenses, the number of inter-state
transferred prisoners in 2005 was
less than 5,000 men and women
nation-wide.8 The total prison
population of the U.S. in 2005 was
2,320,359.9 This means that 99.8% of
prisoners in 2005 were incarcerated in
their home-state, with only 0.2%
transferred. With so few inter-state
transfers, it is reasonable to assume that
the overwhelming majority of prisoners
in my sample are in fact incarcerated in
their state of residence.
over the last 30 years, the average in my
sample is 28%. This differential is in fact
due to the increased proportion of those
in jail that are black. In 1980 only 20.1%
of those in jail were black; that number
rose to 28.1% in 1990 and 33.7% in 2000,
such that the average conceals this
important upward trend. (See Figure 2
below)
Summary Statistics
The average number of years of
education is 12.61. The average years of
education for whites in the sample is
12.73 years while the average for blacks is
11.72 years, which is a difference of a
whole year. Figure 3 shows that
educational
attainment
for
all
observations (including those not who
are not incarcerated) is lower for blacks
than for whites. In addition, the pie
charts below (Figure 5 & 6) show that of
those who are in “jail” in my sample,
blacks are less likely to have any
education past high school than whites.
In fact, 41% of white prisoners had some
college education while only 23% of
black prisoners did. In addition, 35% of
black prisoners were high school drop
outs compared to only 20% of whites.
My data contains 2,236,017
observations, 2,064,195 of which were
either black or white. I dropped 171,822
observations that were of another race.
11.78% of the resulting observations were
black and 88.22% were white. There are
very revealing differences between
whites and blacks in my sample.
According to my calculation of
the variable “jail,” 4.01% of the
observations were in jail; 9.6% of the
blacks were in jail compared to 3.3% of
whites. In the U.S. in 1997, 34.5% of
inmates were black males10, however,
7
Barnes, 2009
Biasca, 2006
9
Harrison & Beck, 2006
8
10
Bureau of Justice Statistics, USDOJ
159
Figure 3:
Figure 4:
The average age of the individuals
in the sample is approximately 39 years
old. However, the average age of those in
jail was only about 29 years, ten years
younger. 59% of my sample is married;
however, while 61% of the white
population is married, only 39% of the
black population is married. In addition,
the average state population density is
higher for blacks (248.819 people/square
mile) than it is for whites (235.101) in my
sample.
Arizona in 2000 has the lowest measure
(0.006) and Illinois in 1980 has the
highest (0.65).
As I mentioned before, previous
work by Massey & Denton has shown
that the most segregated cities are
located in six states (Maryland, Illinois,
Ohio,
Michigan,
Wisconsin
and
Pennsylvania). In all years, these six
states are in the top thirteen most
segregated states when measured by
isolation.
When
measured
by
dissimilarity, five out of these six states
fall into the top nine most segregated
states for all three census years; the one
exception in all cases being Maryland.
Michigan and Illinois, home to Detroit
and Chicago, are consistently the top two
segregated states when measured by
isolation and in the top four when
measured by dissimilarity. These results
increase my confidence in using these
statistics as accurate measures for
segregation.
The
average
racial
fractionalization (or average state
diversity) is 0.38 with the maximum and
minimum by state and year being 0.66
for California in 2000 and 0.028 for
Vermont in 1980. The average state
black-specific dissimilarity (or evenness)
is 0.59. The minimum was South
Carolina in 1980 at 0.36, while the
maximum was Wisconsin in 1980 with a
very high 0.82. The average state blackspecific isolation (or exposure) is 0.38.
160
Below, Table 1 shows the dissimilarity
data sorted by values for year 2000.
Only the 20 most segregated
states are shown. The aforementioned
six states are highlighted to
demonstrate their high levels of
segregation.
Empirical Model / Initial Findings
My first ordinary least squares
regression is to see the preliminary
relationship between my variables
and the probability of going to jail for
blacks. I have defined the regression
as follows:
Jailisy = f(agei, marriedi, yearsedi, blacki,
isobsy,
disbsy,
dropouti,
elfsy,
popdensitys, y1990, y2000) + random
errorijt
Table 1:
North Dakota
Kansas
Iowa
Maryland
Connecticut
Minnesota
Massachusetts
New Jersey
Colorado
Arkansas
Tennessee
New York
Ohio
Nebraska
Pennsylvania
Indiana
Illinois
Missouri
Michigan
Wisconsin
Jail originates from group
quarters and is defined as I have
previously discussed and represents the
probability for an individual i living in
state s in year y will be incarcerated. The
variables age, married, and black are selfexplanatory. Yearsed is
1980
0.711329
0.613902
0.629343
0.553557
0.610418
0.680671
0.668681
0.650763
0.650976
0.59158
0.611998
0.673047
0.692788
0.767836
0.726254
0.722335
0.760745
0.742566
0.763684
0.821375
1990
0.637969
0.582375
0.607512
0.578105
0.583934
0.637085
0.611289
0.621204
0.633571
0.602176
0.625357
0.663504
0.674489
0.713212
0.712848
0.702566
0.72048
0.710704
0.775979
0.812679
2000
0.540646
0.549115
0.550264
0.556451
0.560976
0.578137
0.578235
0.586884
0.599502
0.608947
0.624735
0.643495
0.656717
0.659391
0.674941
0.676884
0.681031
0.706228
0.746995
0.762923
the number of years of education, elf is
diversity, isob is isolation of blacks, disb
is dissimilarity of blacks, dropout is a
dummy variable to control for dropouts,
popdensity is a measure of population
density per state, and y1990 and y2000
are year dummy variables. The first
regression is for both blacks and whites.
The second regression is for whites only
and the third is for blacks only.
Figure 5:
In table 2 you see that being black
increases your probability of going to jail
by 3.9 percentage points. In addition,
being older, being married, and having
more education make blacks and whites
less likely to be in jail. Being a high
school dropout makes blacks 3.2
percentage points more likely to be in
jail, but being a high school dropout
161
actually makes whites less likely to go to
jail. In addition, we see that whites
become less likely to be in jail in 1990
and 2000 while blacks and more and
more likely. In fact, in 1990, they are 2.2
percentage points more likely and in
Table 2
age
married
yearsed
black
elf
isob
disb
dropout
popdensity
y1990
y2000
Observations:
R-squared:
Adjusted R-squared:
*** Significant at 1% level
** Significant at 5% level
*Significant at 10% level
2000 4.1 percentage points more likely.
This finding is expected because as we
have seen from Figure 2, blacks become a
higher portion of inmates are the
decades progress.
White & Whites only
Blacks
-.0010*** -.0009***
(0.0000) (0.0000)
-.0809*** -.0738***
(0.0003) (0.0003)
-.0047*** -.0052***
(0.0001)
(0.0001)
0.0393*** dropped
(0.0001)
-.0059*** -.0118***
(0.0011)
(0.0010)
-.0059*** .0097***
(0.0012)
(0.0011)
-.0228*** -.0423***
(0.0019)
(0.0018)
-.0133***
(0.0005)
-0.000002***
.0123**
(0.006)
-.1840***
(0.0063)
.0776***
(0.0072)
-.0225***
(0.0004)
-0.000002***
(0.0000)
-.0006***
(0.0003)
.0014***
(0.0004)
2,063,910
0.0710
0.0710
(0.0004)
-.0034***
(0.0003)
-.0038***
(0.0003)
1,843,507
0.0618
0.0618
State diversity makes whites less
likely to be in jail while it increases the
Blacks only
-.0022***
(0.0001)
-.1310***
(0.0013)
-.0039***
(0.0002)
dropped
.0315***
(0.0019)
.00001***
(0.0000)
.0215***
(0.0016)
.0411***
(.0017)
220,403
0.0831
0.0831
likelihood for blacks. In addition, blacks
living in more dissimilar (more uneven /
162
Segregation by state: Evenness of Blacks
Note: more than 0.6 is dark grey, 0.5-0.6 is light grey, and less than
0.5 is white.
However, when segregation is
measured by isolation (exposure), we see
the opposite effect. The problem with
isolation is that even if all the blacks in
one state live in the same city, they
might still be exposed to whites on a
regular basis because there are few
more segregated) states are 7.8
percentage points more likely to be in jail
while whites living in more segregated
states are actually 4.2 percentage points
less likely to be in jail. This means that
dissimilarity of blacks helps whites and
hurts blacks.
Segregation by State: Isolation of Blacks
Note: more than 0.4 is dark grey, 0.25-0.4 is light grey, and less than
0.25 is white.
163
blacks in that state and hence blacks are
unable to avoid whites. This explains
why some of the less populated and
states with lower black populations show
up as being somewhat segregated when
measured by dissimilarity but not by
isolation. In figures 7 and 8 below, you
can see that isolation is working more
like a proxy for the population of the
state that is black.
For example, northern cities are
notoriously more segregated than
southern cities due to historical events
such as “The Great Migration” and “white
flight,” but the southern states are
segregated when measured by isolation
and not when measured by dissimilarity.
Isolation is still a good measure, in that it
in some way controls for the population
of blacks in the state, but it will not be as
representative
of
segregation
as
dissimilarity will be, especially on the
state level.
Instrumental Variable
Any endogenetiy in my results
could be caused by reverse causality. The
dependent variable (incarceration) could
be causally related to one of my
covariates (segregation) which would
create a case of reverse-causality. For
example, it is possibly that high crime
rates cause increased incarceration but
also cause a fear of living in high crime
areas. This could in turn reduce the
portion of white and educated people
living in a particular area leading to
higher segregation indices. However, I
find the opposite story more convincing;
that segregation leads to limited
opportunities that lead to crime and
incarceration. As I mentioned above,
Table 3
Blacks only/
without IV
age
-.0022***
(0.0001)
married
-.1310***
(0.0013)
yearsed
-.0039***
(0.0002)
elf
.0123**
(0.006)
isob
-.1841***
(0.0063)
disb
.0776***
(0.0075)
dropout
.0315***
(0.0019)
popdensity
-.00001***
(0.0000)
y1990
.0215***
(0.0016)
y2000
.0411***
(.0017)
Observations: 220,403
Blacks only/
with IV
-.0022***
(0.0001)
-.1310***
(0.0013)
-.0042***
(0.0003)
0.0471**
(0.0122)
-.2251***
(0.0140)
.1822***
(0.0329)
.0322***
(0.0019)
-.00002***
(0.0000)
0.0218***
(0.0016)
.0409***
(0.0017)
220,403
R-squared:
0.0831
Adjusted R- 0.0831
squared:
0.0823
0.0823
there is evidence that segregation
originally occurred due to racist zoning
and selling of property (Cutler, Glaeser &
Vigdor, 1999) thus we can reasonably
assume that even though other factors
perpetuated segregation levels, the initial
cause was probably not incarceration
levels or crime rates. However, I will
attempt to control for this possibility
regardless.
To control for endogeneity I have
to implement an instrumental variable
which is correlated to the explanatory
variable (segregation) but not to the
164
dependent variable (incarceration.) The
IV that I use is a shift-share IV similar to
those used by Card (2001), Ottovani &
Peri (2006) and Sparber (2008).
interesting to consider why isolation has
a negative relationship to incarceration
rates. Cutler and Glaeser (1997) suggest
that there are some positive effects for
blacks from segregation, such as gains
from homogeneity, community support,
reduced competition with whites and the
increased presence of skilled blacks in
poor urban areas. They argue that if
segregation by race doesn’t also create
segregation by skill, blacks will benefit
from segregation more than they lose.
However, these gains would only come
about through isolation from whites. A
simple lack of evenness would not
produce these effects. Although it is
unlikely, it is a possible explanation for
why fewer blacks go to jail in states with
high isolation.
First, we assume that 1970 U.S.
demography is exogenous and that each
racial group in each zip code grows at its
own exogenous rate. Then we can make
predictions, based on the 1970 data,
about 1980, 1990, and 2000 zip code
demographies.
Theoretically,
those
predicted values, which are not actual
values, should be correlated with
segregation, but not correlated with
incarceration rates.
To estimate the new model, I have
included
the
dissimilarity-specific
instrumental variable (disIV).
One omitted variable that could
be a problem is income. Income and
poverty are highly correlated to the
incidence of crime and incarceration as
well as segregation levels. However,
because incomes for those who are
currently in jail are “0” in the census
records, it is impossible to know if in fact
they were poor when the crime was
committed. However, I believe it is
reasonable to assume that education
levels are correlated to income because
children who are poor have a higher
opportunity cost of attending school
rather than working and brining home
income. Because of this, I think it is
reasonable to assume that the education
variable can be used as a proxy for
income to some extent.
(1)
Jailisy = f(agei, marriedi, yearsedi,
blacki, elfsy, disbsy, dropouti, popdensitys,
y1990, y2000) + disbIV + random errorijt
With the instrumental variable,
none of the results have changed
significantly, making me confident in the
coefficient on my primary measure of
segregation (dissimilarity.) In fact, the
coefficient on dissimilarity has increased
such that now segregation increases the
probability that a black man will be in
jail by 18.22 percentage points. This
means that dissimilarity indeed has a
causal relationship to incarceration
because the reverse causality bias has
been controlled for.
However, the new results increase
the coefficient on isob and show that
living in a state where blacks are more
isolated makes them 22.5 percentage
points less likely to go to jail. It is
Another issue with my model is
that there is no way of knowing when the
prisoners began their sentences. This is a
165
problem because their presence in prison
does not necessarily reflect the
segregation levels of their state in the
year they committed the crime. This is
especially the case for those with longer
sentences. For example, if someone
committed murder in 1960 and is found
in jail in the 2000 census, his crime
should be paired with the segregation
level of 1960, not 2000. However, there
are no clear solutions for this problem.
The theory that I will describe and
examine is as follows. Segregation is
somehow created and perpetuated in the
U.S, which in turn lowers expenditure on
public goods such as education, which in
turn worsens outcomes for blacks living
in segregated neighborhoods, which
reduces the expected wage and
opportunity cost of committing crime,
which in turn increases crime and
incarceration rates, which eventually has
significant impacts on life trajectories for
blacks.
in midcentury, greater levels of
segregation resulted from collective
actions on the part of whites to exclude
blacks:
legitimized
forms
of
discriminations such as restrictive
housing covenants and explicit or
implicit threats of violence. These factors
made blacks pay relatively more for
housing in more segregated cities…” By
the 1970s, they explain, the average black
American lived in a census tract that was
68% percent black. They argue that
segregation declined from the 1970s
onward but note that while educated
blacks moved into more white
neighborhoods, “The share of tracts that
were at least 90% black doubles in both
cities and suburbs.” (Cutler, Glaeser,
Vigdor, 1999.) This means that during my
time period of study, 1980 to 2000, the
existing ghettos became even more
concentrated with poorer, less educated
minorities as other neighborhoods
became more diverse.
Once segregation exists, spending
on public goods decreases. Alesina, Baqir
and Easterly (1999) show that because
heterogeneous and polarized societies
value public goods less, spending on
public goods will be relatively less in
segregated cities than in non-segregated
cities. They find examples of poor public
goods leaving disadvantaged groups even
farther behind, like public transportation
and schooling.
There is evidence that segregation
began due to post-civil war migration
northward and collective racism leading
up to restricted housing options. Cutler,
Glaeser and Vigdor (1999) attempt to
explain segregation’s origins by testing
their theories and they determine that “
The lack of a good quantity and
quality of schooling results in poor
outcomes. As previously mentioned,
Massey, Condran and Denton (1987) find
that in segregated environments, high
income blacks have lower average SAT
scores than low income whites.
Discussion
Through
the
use
of
an
instrumental variable, I am able to show
that
the
relationship
between
segregation and incarceration is not just
correlated,
but
causally
related.
However, it is useful to consider how this
causal link operates.
166
Education is clearly correlated with life
income; I find with my own data that one
more year of education is associated with
a highly significant 17.89 percentage
point increase in the census variable
“poverty” (which is basically a measure of
income up to a certain level.) When
wages and incomes are low, people will
turn to crime (Grogger, 1998.)
young black
dramatically.
men
have
increased
Because poor beginnings lead to
poor life outcomes which again lead to
poor beginnings for the next generation,
this cycle is important to fix. In addition,
high levels of imprisonment remove men
from the outside work force, resulting in
reduced productivity. There have been
some suggestions that are worth noting.
Freeman (1996) proposes that the U.S.
cannot simply increase the consequences
of crimes, or the likelihood of
imprisonment if caught, which have
clearly risen in the past few decades but
been largely disregarded by criminals. He
argues for a “carrot as well as a stick”
approach,
such
that
economic
opportunities must be increased for
those with limited educations.
This therefore explains the high
propensity of crime in segregated areas
and the likelihood of being incarcerated.
Pettit & Western (2004) argue that being
incarcerated has become a life-stage for
many young black males in the last 30
years and interrupts the life course.
“Persistent offending is more likely for
those who fail to secure the markers of
an adult life,” like marriage and steady
jobs. This significantly affects life
trajectories because men who lack those
“markers” have difficulty attaining them
at a later age. Pettit and Western state
that ex-prisoners earn lower wages, are
less likely to be married or live with the
mothers of their children, and are less
likely to get and hold a job due to
stigmas against criminals.
Case & Katz (1991) show that
although bad behaviors and examples are
contagious and can increase youth crime,
good behaviors and examples can also be
contagious. They write, “…shocks or
policy interventions that positively affect
individuals will have positive multiplier
effects within neighborhoods through
peer influences and across generations
through family influences.” This leaves
some room for optimism that perhaps
social programs, positive role-models
and educational funding can do a lot to
reverse this trend.
Conclusion
In this paper I have established a
causal link between segregation and
incarceration. I have laid out the
mechanism through which this occurs. I
have determined that the primary
problem resulting from segregation is the
lack of a quality education. I have shown
the importance of understanding this
problem because due to the recent “war
on drugs,” the incarceration risks for
Works Cited
Alberto Alesina, Reza Baqir, and William
Easterly (1999) “Public Goods and Ethnic
Divisions” The Quarterly Journal of
Economics, Vol. 114(4): 1243-1284
167
Andrew Paley, “Census engaged in
'prison-based gerrymandering,' report
says” Medill Reports, Northwestern
University, (02/04/2010) Retrieved from:
http://news.medill.northwestern.edu/chi
cago/news.aspx?id=155730
Debra Biasca, (2006) “Interstate Transfer
of Prison Inmates in the United States”
National Institute of Corrections, USDOJ
Douglas S. Massey and Nancy A. Denton
(1989)
“Hypersegregation
in
U.S.
Metropolitan Areas: Black and Hispanic
Segregation along Five Dimensions”
Demography, Vol. 26(3): 373-391
Anne C. Case and Lawrence F. Katz,
(1991) “The Company You Keep: The
Effect of Family and Neighborhood on
Disadvantaged Youth,” NBER Working
Paper No. 3705
Douglas S. Massey, Gretchen A. Condran,
Nancy A. Denton (1987) “The Effect of
Residential Segregation on Black Social
and Economic Well-Being” Social Forces,
Vol. 66(1): 29-56
Becky Pettit and Bruce Western, (2004)
“Mass Imprisonment and the Life
Course: Race and Class Inequity in U.S.
Incarceration.” American Sociological
Review, Vol. 69(2): 151-169
Edward L. Glaeser, Bruce Sacerdote, and
Jose A. Scheinkman (1996) “Crime and
Social Interactions” The Quarterly
Journal of Economics, Vol. 111(2): 507-548
Chad Sparber, (2008) “Racial Diversity
and Aggregate Productivity in U.S.
Industries:
1980-2000”
Southern
Economic Journal
George J. Borjas, Jeffrey Grogger, and
Gordon H. Hanson (2006) “Immigration
and
African-American
Employment
Opportunities: The Response of Wages,
Employment, and Incarceration to Labor
Supply Shocks” NBER Working Paper
No. 12518
Christopher Hartney, “Fact Sheet: US
Rates of Incarceration: A Global
Perspective.” National Council on Crime
and Delinquency (NCCD), November
2006
Jeff Grogger (1998) “Market Wages and
Youth Crime” Journey of Labor
Economics, Vol. 16(4): 756-791
Daniel T. Lichter (1988) “Racial
Differences in Underemployment in
American Cities” The American Journal of
Sociology, Vol. 93(4): 771-792
Patrick A. Langan and David J. Levin
(2002) “Recidivism of Prisoners Released
in 2004” Bureau of Justice Statistics
Special Report, USDOJ, NCJ 193427
David M. Cutler and Edward L. Glaeser
(1997) “Are Ghettos good or bad” The
Quarterly Journal of Economics, Vol.
112(3): 827-872.
Paige M. Harrison and Allen J. Beck
(2006) “Prisoners in 2005” Bureau of
Justice Statistics Bulletin, USDOJ, NCJ
215092
David M. Cutler, Edward L. Glaeser, and
Jacob L. Vigdor (1999) “The Rise and
Decline of the American Ghetto” The
Journal of Political Economy, Vol. 107 (3):
455-506
Reed Karaim, “Special Report: Housing
First, People with Mental Illness” NPR,
168
2002
Retrieved
from:
http://www.npr.org/news/specials/housi
ngfirst/whoneeds/mentallyill.html
Richard B. Freeman (1996) “Why Do So
Many Young American Men Commit
Crime and What Might We Do About
It?” The Journal of Economic Perspectives,
Vol. 10(1): 25-42
Steven Ruggles, J. Trent Alexander, Katie
Genadek, Ronald Goeken, Matthew B.
Schroeder,
and
Matthew
Sobek.
Integrated Public Use Microdata Series:
Version 5.0 [Machine-readable database].
Minneapolis: University of Minnesota,
2010
Tom Barnes, “Pa. will transfer 2,000
inmates to Va., Mich.” Pittsburgh PostGazette, (12/22/2009) Retrieved from:
http://www.postgazette.com/pg/09356/1022675-454.stm
169
The Effects of Taurine and Caffine Alone and in
Combination on Locomotor Activity in the Rat
Marissa Dombovy-Johnson, Class of 2010
Background Information
History of Energy Drinks
Energy drinks were first marketed
in Europe and Asia in the 1960’s, but they
did not reach the United States market
until the 1980’s. Red Bull, the leader of
energy drink sales since its introduction,
was created by Dietrich Mateschitz and
released in Austria in 1987 and in the
United States ten years later (Ressig,
Strain, and Griffiths, 2009). By 2006
there were over 500 brands worldwide
(Johnson, 2006) and approximately 31%
of American teens consumed energy
drinks (Lord, 2007). In 2008 the leaders
in the energy drink market were Red
Bull, Monster, Rockstar, and AMP. They
had 40, 23, 12.3, and 8 percent of the
sales, respectively (Energy Fiend, 2007).
energy gets you focused and ready for
everything life throws at you each day,”
according to each of their websites (AMP
Energy, 2010; Red Bull, 2010).
The ingredients in energy drinks
are designed to create activating effects
on the human body. The most common
ingredients found in energy drinks are
ginseng, taurine, caffeine, glucose, Bvitamins, and glucuronolactone. Each
energy drink is unique in its combination
of ingredients; however, caffeine and
taurine are present in all of the above
mentioned
energy
drinks.
Glucuronolactone and glucose are
present in Monster and Red Bull, ginseng
in AMP, while B-vitamins are in Red Bull
and AMP (Amp Energy, 2010; Monster
Energy Drink, 2010; Red Bull, 2010;
Rockstar, 2010).
Energy drinks, unlike their
caffeine counterparts, coffee and soda,
were associated with the emerging
lifestyle of action/extreme sports (Ho,
2006). They are marketed towards those
wanting increased alertness and energy,
particularly in athletic endeavors. Thus
today’s youth, especially young males,
have become the target for energy drink
companies. Miller (2008) surveyed
undergraduate students in 2008 and
found that masculinity and risk taking
behaviors were positively correlated with
energy drink consumption. Red Bull
“vitalizes body and mind,” while “AMP
The
above
ingredients
are
included because
of proven or
hypothesized effects of increased
alertness, stimulation, and energy
(Clauson, Shields, McQueen, and Persad,
2008).
Glucose,
which
is
a
monosaccharide, is the main source of
energy in the human body. In addition to
providing energy, glucose has been found
to enhance memory, and concentration
(Clauson et al., 2008). Panax ginseng
170
grows naturally and has been used to
prevent fatigue, liver disease, amnesia,
and colds. It has been suggested that has
immuno-modulating activity and can
enhance mental capacity (Kiefer and
Pantuso, 2003). B-vitamins include a
wide variety of vitamins, such as folate,
B12, and B6. Deficiencies in B-vitamins
can lead to impaired neurological and
psychological
functioning
(Selhub,
Bagley, Miller, and Rosenberg, 2009). A
naturally occurring substance in the
body, glucuronolactone is believed by
energy drink manufacturers to detoxify
the body. However, no studies have
studied this (Kim, 2003). Taurine, an
essential amino acid that is not used to
build proteins (Birdsall, 1998), has many
different effects, possibly because of its
interaction at the GABA, glycine, and
“unknown”
taurine
receptors
(Ramanathan, Chung, Giacomini, and
Brett, 1997; El Idrissi et al., 2008). This
amino acid’s best known role involves
the synthesis of taurine bile acid
conjugates (Clauson et al., 2008).
However, its function in improving
activity level and mental performance
has been controversial. The last main
ingredient is caffeine, which stimulates
the central nervous system, heart, and
skeletal muscle, and has a diuretic effect
(Clauson et al., 2008).
maybe due to its suggested stimulant
effects or, when combined with caffeine
it further increases the stimulant
properties of caffeine. However, Whirley
and Einat (2008) concluded that
taurine’s stimulant properties were
hypothetical because they failed to find
increased activity in mice. However, they
studied taurine alone and not in
combination with caffeine, thus further
research is needed to verify this
possibility. Furthermore, Zeratsky (2008)
states that athletic and maybe mental
performance is increased by the
interaction of taurine and caffeine. If
caffeine and taurine acted in a synergistic
way then it would support their dual
inclusion in energy drinks. In this study I
examined whether or not the locomotor
activity of a rat, as measured by the
number of rotations in a rodent wheel
and activity in the open field test,
increased in the presence of combined
taurine and caffeine, when compared to
taurine or caffeine alone or control
saline. As well as measuring activity, the
open field test is also useful in evaluating
anxiety, a potentially detrimental side
effect of caffeine. Before stating the
details of this experiment, research
involving taurine and caffeine alone and
in combination will be further examined.
Caffeine
As shown above, the ingredients
found in energy drinks have separate
potentially different effects on the brain
and body. As these effects are likely to
interact, the effects of ingesting an
energy drink are likely due to the
combined interaction of these substances
on the brain and body systems. The
inclusion of taurine in energy drinks
Caffeine, the most commonly
used psychoactive substance, is believed
by most researchers to cause the increase
in energy, memory and alertness
associated with energy drinks. Caffeine is
a competitive A1 and A2 adenosine
receptor antagonist (Hughes, McHugh,
and Holtzman, 1998; Smit and Rogers,
171
2000; Angleucci, Cesário, Hiror, Rosalen,
and Da Cunha, 2002; Karcz-Kubicha et
al., 2003). Adenosine slows down neural
activity and, by preventing this inhibitory
effect, caffeine stimulates brain activity
(Rao, Hu, and Nobre et al., 2005). The A1
adenosine receptors are expressed in the
cerebral cortex, while the A2 receptors
are present in the hippocampus and the
striatum (Angleucci et al., 2002).
Angelucci et al. (2002) reported that
acetylcholine (ACh) release is inhibited
by activating the A1 receptor, thus
caffeine positively impacts memory
storage. Motor activity is believed to be
regulated by adenosine’s interaction with
both the A1 and A2 receptors (KarczKubicha et al., 2003). Synder, Katims,
Annau, Bruns, and Daly (1981) showed
that the A1 receptors mediated the
motor-activity effects of
caffeine.
However, later studies have determined
the A2 receptors as the cause of the
activity increase (Ferré, Fuxem von Euler,
Johansson, and Fredholm, 1992).
high doses of caffeine. The timing of the
dosing was another important finding
because the rats who received caffeine
after training experienced more of an
impact on memory than those who
received it before training, implying a
possible effect on memory consolidation.
Enhancement in memory and other
higher level cognitive functions possibly
results from caffeine’s antagonism by
enhancing cholinergic release (Smit and
Rogers, 2000).
Numerous past studies have been
conducted on caffeine’s effects both
alone or in association with other energy
drink ingredients. Scholey and Kennedy
(2004) found improved memory and
speed in attention tests, but no other
mood or cognitive effects, when humans
drank a mixture of caffeine, glucose, and
herbal flavoring fractions. Similarly,
Kennedy and Scholey (2004) found that
after cognitive demand, a caffeine and
glucose drink improves cognitive
performance and decreases fatigue.
Increased behavioral performance in a
selective-attention
task,
specifically
improved accuracy and reaction times,
was
discovered
after
individuals
consumed a caffeine, glucose drink (Rao
et al., 2005). These researchers postulate
that caffeine and glucose, the major
source of energy in the body, may
interact in a synergistic way. Various
studies (Johnson, Spinweber, and
Gomez, 1990; Rusted, 1999; Walburton,
1995) found increased alertness after
caffeine was administered. Frewer and
Lader (1991) and Fine et al. (1994)
replicated the findings that caffeine has
beneficial effects on psychomotor speed
and vigilance.
The ingestion of caffeine has
many effects on the human body. This
substance helps prevent fatigue (Smit
and Rogers, 2000), decreases reaction
time (Clubley, Bye, Henson, Peck, and
Riddington, 1979; Lieberman, Wurtman,
Emde, Roberts, and Coviella, 1987; Kerr,
Sherwood, and Hindmarch, 1991),
facilitates
memory
attention
and
stimulates the heart, and voluntary
muscle (Angleucci et al., 2002).
Angleucci et al. (2002) found that the
effects of caffeine are dose related, with
peak stimulating effects occurring at
moderate doses. Behavioral activity,
learning, and memory were actually
suppressed when rats were administered
172
(Whirley and Einat, 2008). High levels of
this amino acid can be found in the
cerebral
cortex,
basal
ganglia,
hippocampus,
hypothalamus,
cerebellum, and the caudate-putamen
(Olive, 2002; El Idrissi et al., 2008;
Whirley and Einat, 2008). In order to
reach the brain, taurine must cross the
blood-brain-barrier by way of a sodium
and chloride dependent carrier (Benrabh,
Bourre, and Lefauconnier, 1995).
As mentioned above, caffeine has
been found to be a competitive
adenosine antagonist and have dose
related effects. Karcz-Kubicha et al.
(2003) found that administering caffeine
acutely
produced
motor-activating
effects by blocking both the A1 and A2a
receptors. Despite being an antagonist of
both adenosine subtypes, Ferré et al.
(1992) determined that the A2a receptors
are responsible for the motor-activating
effects of caffeine in rodents. Therefore,
they may also be the cause for decreased
reaction times experienced by humans
after caffeine ingestion. Past research has
shown positive effects on arousal when
lower doses of caffeine were used (12.5100mg) (Smit and Rogers, 2000, 2002;
Smit, Cotton, Hughes, and Rogers, 2004),
while a dose of 500mg produces negative
effects by impairing performance on tests
of attention and vigilance (Frewer and
Lader, 1991). The best interpretation of
the above literature review is that
caffeine exerts stimulant effects in a
dose-dependent manner.
Taurine
Taurine has been found to impact
many different biochemical actions
(Huxtable, 1992), thus it has been
difficult to locate the exact taurine
receptors. It can act as a trophic factor
and inhibitory modulator (Chepkova,
Sergeeva,
and
Hass,
2005),
a
neuroprotectant (Kim, 2003), as well as
playing a role in detoxification,
membrane
stabilization,
and
osmoregulation (Birdsall, 1998; Imagawa
et al., 2009). Taurine levels were altered
in some patients with neuropsychiatric
illness, suggesting a possible use as
medication. In addition it already has
been used to treat people with seizure
disorders (Whirley and Einat, 2008).
Taurine is the most abundant
amino acid in many tissues and, most
importantly, in the brain’s extracellular
fluid (Birdsall, 1998; Olive, 2002).
Cysteine sulfinic acid decarboxlyase
(CSAD)
synthesizes
taurine
from
cysteine (Olive, 2002). Since it is not
incorporated into proteins (Birdsall,
1998), taurine is found free in tissues, for
example the brain and muscles (Whirley
and Einat, 2008). The consumption in
food and synthesis by astrocyctes are the
main ways taurine enters the body
Since taurine is structurally
similar to GABA, it may mimic GABA’s
effects in enhancing chloride ion flux or
possibly displace the agonists associated
with the GABAa receptor complex
(Olive,
2002).
Furthermore,
its
interaction with GABA receptors, as well
as glycine, cholinergic, and adrenergic
receptors, may provide support for its
role in stress, mood, and behavior (El
Idrissi et al., 2008). The interaction at
GABA receptors is important for
taurine’s anxiolytic effect because
173
anxiolytic drugs interact at the same
receptor (Whirley and Einat, 2008).
Similar to taurine’s use in treating people
with seizure disorders, its interaction at
GABA receptors demonstrate is sedative
effect when used alone. However, when
combined with caffeine it may interact in
a synergistic way to further increase
caffeine’s stimulant properties.
Other behavioral studies have
produced conflicting results regarding
taurine’s role in locomotor activity as
measured through the open-field test.
Baskin et al. (1974) found a dosedependent decrease in activity following
i.p. injections of taurine (0.3-3.0 mg/kg).
Similarly, Sanberg and Ossenkopp (1977)
found that increasing doses of taurine
(25-200 mg/kg) decreased ambulation in
rats, with doses greater than 50 mg/kg
causing a significant decrease. Mice
exhibited the same decrease in activity
following i.p. injections (Hruska, Thut,
Huxtable,
and
Bressler,
1975).
Contradictory to the Sanberg and
Ossenkopp’s (1977) finding, Kong et al.
(2006) found that a dose of 126 mg/kg
increased the number of times a rat
entered the center during the open field
test. Comparable to Kong et al.’s finding,
Imagawa et al. (2009) found an increase
in endurance performance following two
weeks of taurine administration. Possible
explanations for these differences could
be environmental conditions, duration of
administration, and age of the animals.
Higgins, Jones, and Oakley (1992) and
Serrano et al. (2002) have stated that
environmental conditions and the age
impact the way taurine interacts in the
rodent body.
El Idrissi et al. (2008) found that
the way taurine was administered, either
injected (43 mg/kg) or supplemented in
drinking water, resulted in suppressed or
increased
locomotor
activity,
respectively. The injection produced an
anti-anxiety
effect,
while
supplementation induces the opposite
effect, anxiety. In another study in mice,
Whirley and Einat (2008) gave two
different doses of taurine that were
significantly higher than the amounts in
energy drinks. However, they stated that
mice, when compared to humans, are
given higher doses of drugs due to
metabolism differences. Since the high
doses had no significant stimulant
effects, they speculated that lower doses
maybe more effective. Furthermore, they
only tested taurine, and other research
implies a possible synergistic effect with
caffeine. Even though they found that
sub-chronic or chronic administration
did not have a stimulant effect, they saw
an initial reduction in locomotor activity,
which is typical of other stimulants,
followed by an increase in activity. They
propose that this may be due to taurine
affecting different pathways. After
analysis of their findings and past
research, they consider the claim of
taurine as a stimulant to be speculative
and further research is needed.
Caffeine and Taurine
As energy drinks often contain
caffeine and taurine it is vital to examine
past research involving the combination.
Many researchers have looked at the
impact of Red Bull, which contains
caffeine and taurine, in humans. Alford,
Cox, and Wescott (2001) found that a
174
Red Bull energy drink caused increased
diastolic
blood
pressure,
aerobic
endurance and anaerobic performance.
The observed increase in blood pressure
may suggest an interaction between
taurine and caffeine because taurine is an
antihypertensive, thus opposing the
caffeine mediated increase in both the
diastolic and systolic blood pressure. No
effect on short term recall memory in
college students was found after they
ingested a Red Bull drink (Birchler,
Swenson, and Harris, 2006). The heart
rate in subjects was decreased, and
cardiac stroke volume was increased
(Baum and Weis, 2001). Baum and Weis
(2001) suggested that the increase in
ventricular function in young athletes
may be due to the combined effects of
caffeine and taurine. However, these
effects may also be due to caffeine, since
they did not look at that substance alone.
performance in mice. In addition, their
results also show that caffeine and
taurine use for two weeks had an
additive effect, when compared to both
substances alone and to administration
for only a week. Therefore, the duration
of administration could be essential to
caffeine’s and taurine’s effects in energy
drinks.
The research on the components
of energy drinks is very limited. Since the
literature suggests a possible synergistic
effect between taurine and caffeine on
causing the increased alertness and faster
processing associated with energy drinks,
it is important to study this in further
detail. Based on past research I
hypothesize that rats will display
increases in locomotor activity after an
injection caffeine alone and an additive
effect when caffeine and taurine are
combined.
Walburton,
Bersellini,
and
Sweeney (2001) looked at a glucose,
caffeine, and taurine drink and
discovered that information processing
increased after consumption of caffeine
and taurine. Glucose was not included
because there was no difference between
the control drinks, one sugar (glucose)
free and the other with just glucose.
However, similar to the Baum and Weis
(2001) study, caffeine was only looked at
in association with taurine, therefore it
could be caffeine’s effect alone.
Methods
Animals
Adult male Sprague-Dawley rats
raised at Colgate University weighing
472-578g (avg. 514g) at the beginning of
the experiment were used. The rats were
born in either February or March 2009
making them an average of a year old
during the experiment. The animals were
housed on a 12hr light/dark cycle (lights
on at 08:00) at 22ºC ± 1ºC with food and
water available ad libitum, except from
the time of the injection until the
beginning of the exercise wheel. The rats
were housed in cages (47 X 26 X 20 cm)
with the same cagemates during the
duration of the experiments. The cages
were switched every week to allow the
Imagawa et al. (2009), as far as I
know, were the only study to examine
the effect of taurine and caffeine, alone
and in combination, on performance.
They found that taurine use for two
weeks significantly increased endurance
175
Exercise Wheel
rats to receive equal distribution of light.
After the habituation period, which will
be
explained
in
the
procedure
subsection, the rats were broken into
four groups: control (C), taurine (T),
caffeine (F), and taurine and caffeine, the
combined group (B). The experiments
and treatments were approved by
Colgate University’s Animal Care
Committee.
The exercise wheel consists of a
rodent wheel containing an automated
counting device secured to a plastic
storage bin covered with bedding and
supplied with water and food. Thus,
while in the wheel, each full rotation was
automatically recorded.
Procedure
Drugs
Twenty four rats went through
three 16-hr habituation periods (which
included their active period) in the
exercise wheels before the experiment
began in order to choose twelve that ran
an average of 5 complete spins. Rats that
never ran or ran over 15 times in any
habituation period were eliminated.
After the habituation period, the twelve
rats chosen experienced each treatment
group. In order to prevent order effects,
the rats were divided into the four
groups described above with three rats in
each. Thus, each rat went through four
trials, one for each treatment group.
There was at least 36 hours in between
injections on the same rat. The open field
test began thirty minutes after the
injection, which occurred in the same
room as the open field. The interval of
30-min
was
chosen
because
methylxanthine, a metabolite of caffeine
formed in the liver, has a half-life of 0.71.2 hr in rats, compared to 2.5-4.5 hr for
humans (Morgan, Stults, and Zabik,
1982). Taurine has a longer full body halflife at 18.6 days (Huxtable, 2000). If a
longer duration was chosen, then
caffeine would be mostly metabolized,
thus the results would not display an
accurate portrayal of caffeine’s effects.
Taurine (cell culture tested), 2aminoethanesulfonic acid, and caffeine,
1,3,7-trimethylxanthine, were purchased
from Aldrich Supply Company. The
average weight of 514g was used as the
weight multiplied by the drug dose,
which was 50 mg/kg taurine and 30
mg/kg caffeine. After the drugs
completely dissolved, 1mL aliquots were
transferred to twelve 1mL micro-vials and
frozen until needed. This ensured that
the doses were as equal as possible. All of
the drugs were injected intraperitoneally
(i.p.) in a 1mL amount 30 min before the
open field test. Control rats received an
i.p. injection of 1 mL of 0.9% NaCl and
the combined group received the same
amount of both drugs.
Open-field Apparatus
A rectangular (105 X 90 cm) open
field divided into 42 equally sized
squares (15 X 15 cm) with a plexiglass
floor and 28 cm high walls was used. A
video camera was positioned above the
arena, high enough to cover just the
field.
176
Statistical Analysis
Furthermore, past research using the
open-field test waited thirty minutes
after the injection before beginning the
experiment (Kong et al., 2006; Whirley
and Einat, 2008).The rats were placed in
the center to start and were videotaped
for ten minutes, after which they were
placed in the exercise wheel for four
hours. The open-field was cleaned after
each rat with 70% ethanol to remove any
odor traces. The open field test was
analyzed for total number of squares
entered with all four paws, number of
rears (forelegs raised from the floor or
against the wall), total duration of
grooming (total time in seconds spent
grooming; considered as any time the
paws are touching fur or whiskers), and
percentage of perimeter squares entered
compared to the middle squares (rats
usually move around by hugging the
walls with one side of their body).
Perimeter squares were considered the
squares that directly bordered a wall.
Videotapes were later scored by 3
observers blind to the treatments,
allowing an average of the 3 to be used
for each trial (every rat went through 4
trials, one for each drug). See figure 1 for
the sheet used by the observers. After the
open field, the rats were immediately
placed into an exercise wheel in an
adjacent room for 4-hrs. At the end of
the 4-hrs, the number of rotations were
recorded and the rat was returned to its
original cage in the housing room. All
the behavioral experiments occurred
during the hours of 10:00 and 16:00h,
besides the habituation period, which
occurred during a 16-hour window
encompassing their active period.
Results are reported as means ±
SEM. Data were analyzed by means of
repeated measures analysis of variance
(ANOVA). Whenever the ANOVA was
significant, further multiple comparisons
were made using the paired t-test. All
analyses were performed using the
software SPSS V16.0 for windows. The
level of statistical significance adopted
for the ANOVA was P < 0.05, while for
the paired t-test it was P < 0.0083, which
corresponds to the Bonferroni correction
for 6 pair-wise comparisons.
Results
Open Field Test
Table 1 shows the complete data
for the open field test and the exercise
wheel. The results only include the data
from ten of the twelve rats. One rat died
after receiving only two drugs, while the
other suffered a hematoma after the
injection, thus did not participate in one
trial. Therefore, in order to make sure
the results were from rats that
experienced all the drugs, the data for
these two rats were eliminated. The
results for the open field test are also
shown in figure 2. The ANOVA for total
squares showed a significant druginduced effect [F (3,27) = 12.9, P < 0.05].
The paired t-test showed a significant
difference when the combined group of
taurine and caffeine was compared to the
control [t (10) = 6.9, P < 0.0083] and
taurine groups [t (10) = 5.8, P < 0.0083].
Caffeine showed a tendency to increase
the number of squares entered when
compared to the control [t (10) = 3.2, P <
177
0.05] and taurine groups [t (10) = 2.6, P <
0.05]. There was no significant difference
between the number of squares entered
for the combined group when compared
to caffeine [t (10) = 0.9, P > 0.0083]. Both
taurine and saline have no significant
effect on the locomotor activity, as
shown through total squares entered.
taurine’s stimulant properties has
produced conflicting results possibly due
to large differences in the dosages used
(Sanberg and Ossenkopp, 1977; Kong et
al., 2006). Examination of their
combined effect has been sparse with the
majority finding an increase in activity
(Alford et al., 2001; Walburton et al.,
2001). Only one study to my knowledge
has considered the drugs in combination,
as well as alone. In this study, Imagawa
et al. (2009) found that taurine and
caffeine had an additive effect on
endurance.
Figure 2 shows the results for the
number of rears. The repeated measures
ANOVA showed a significant druginduced effect [F (3,27) = 5.54, P < 0.05].
The use of a paired t-test showed that
both the caffeine and combined groups
had a tendency to increase the number of
rears when compared to the control and
taurine groups [P < 0.05]. No pairings
showed a significant difference.
Based on past research and
conflicting data, the present study was
designed to evaluate the potential
stimulant interaction caffeine and
taurine may have. The open-field test
and the exercise wheel were chosen as
the apparatuses because of their ability
to show multiple behaviors at once.
For grooming [F (3,27) = 2.8, P >
0.05] and the percentage of outer squares
[F (3,27) = 2.9, P > 0.05] the ANOVAs
failed to show a significant difference
among treatment groups.
One dose each of caffeine and
taurine were tested, 30 and 50 mg/kg,
respectively. These dosages were chosen
based on previous research (Sanberg and
Ossenkopp, 1977; Smit and Rogers, 2000,
2002; Smit et al., 2004; Kong et al., 2006).
Both of these doses are higher than the
ones contained in a typical energy drink.
An 8.0 oz can of Rockstar contains 80mg
caffeine and 1000mg taurine, thus for a
70kg person the doses would be 1.14 and
14.3 mg/kg, respectively. Even though
these doses are significantly greater, it is
typical for rats to be given greater
amounts because they metabolize the
drugs at a faster rate (Einat, Yuan, and
Manji, 2001).
Exercise Wheel
There
was
no
significant
difference among treatment groups for
the exercise wheel data [F (3,27) = 0.78, P
> 0.05].
Discussion
The interaction of caffeine and
taurine has not been a well studied in
past research, despite their presence
together in energy drinks. Past research
has concluded that caffeine alone
increases alertness (Smit and Rogers,
2000) and decreases reaction time
(Clubley et al., 1979). Past research on
178
The open-field test, ever since its
introduction by Calvin Hall in the 1930’s,
has been one of the most common
apparatuses used to study rat behavior
because of the various measures one is
able to analyze after one trial in the field
(Walsh and Cummins, 1976). In this
study, the open-field was used to assess
how four different behaviors were
affected by saline, taurine, caffeine, and
both drugs. The behaviors considered
were total squares entered or total
locomotion, rearing, grooming, and
percentage of perimeter squares entered.
reached significance [F (3,27) = 5.54, P <
0.05], but none of the pairwise
comparisons were statistically different
(P < 0.0083). However, there was a
tendency for the combined and caffeine
groups to increase the number of rears
when both were compared to control and
taurine groups. As with total locomotion,
there was no significant difference
between the combined and caffeine
groups.
Grooming has long been seen as
an anxiogenic, or anxious response
(Choleris, Thomas, Kavaliers, and Prato,
2001; Nosek et al., 2008). In other words,
when a rat grooms itself, it is anxious.
Stimulants, such as caffeine, have been
found to promote an anxious state, thus
the total time of grooming would be
expected to be the highest from the
combined and caffeine groups. Despite
there not being a significant difference
between the different groups [F (3,27) =
2.8], caffeine’s average was the highest
amongst the 4 groups (65.3 ± 22.8).
Total locomotion is an important
factor for determining whether or not a
substance causes a stimulant effect
because stimulants increase activity. The
opposite of locomotion is called freezing
and it coincides with high amounts of
stress (Walsh and Cummins, 1976). If
caffeine and taurine interact in a
synergistic way to strengthen caffeine’s
stimulant properties, then the total
locomotion (total squares entered)
experienced by this group would be
greater than for any of the other groups.
The finding that the combined
administration of caffeine and taurine
produced a significant increase in
locomotion when compared to the
control and taurine groups, suggests that
caffeine and taurine somehow interact.
However, there was no difference
between the results of the combined
group versus caffeine alone. Rearing, like
total locomotion, is another measure of
activity specifically, excitability (Walsh
and Cummins, 1976). When the rat is less
anxious, they are more likely to explore
the open-field. Rearing is a way for them
to explore vertically. The ANOVA
The last measure of the open-field
test was the percentage of perimeter
squares entered. Rats are typical wallhuggers because it allows them to not
enter open and dangerous spaces
(Wilson et al., 1976). Thus, when they are
placed into the open-field rats remain in
close proximity to the walls. When rats
travel away from the perimeter of the
open-field it is normally because they are
not as anxious and feel more comfortable
exploring the remaining field (Lister,
1990). The ANOVA for the percentage of
perimeter squares entered failed to reach
significance [F (3,27) = 2.9].
179
The exercise wheel is a measure of
activity because the rat has free access to
the wheel. Before the experiment began,
the rats went through three habituation
periods in order to become acquainted
with surroundings and figure out how to
work the wheel. The ANOVA failed to
show
any
statistically
significant
difference among groups for the exercise
wheel data [F (3,27) = 0.78]. This could
be due to the fact that all of the rats ran
very little, thus making comparisons
difficult. Reasons for the lack of running
in the exercise wheel could include the
nocturnal nature of rats (i.e. rats are
more active at night). The rats chosen
ran an average of 1.5-8.5 spins during the
habituation period, which encompassed
their active period. However, when the
rats were placed into the exercise wheel
after the open-field test, they were in
there for a 4-hr period that was during
their normal sleep time. Furthermore,
the rats only experienced the exercise
wheels
three
times
before
the
experiment, so it is possible that further
habituation time could have led to more
spins during the experiment.
synergistic effect does indeed happen
and at what doses.
Second, the rats used in this
experiment were around 12 months old,
which brings in the issue of age. Most
experiments use rats that are younger
because as they get older their activity
levels are not as high. Future research
should use a group of rats that are
younger, in order to produce more
accurate results. In addition to activity
levels, many drugs exert different effects
depending on the age of the test animals
(Serrano et al., 2002). In addition, the
data was only from ten rats, which is a
very small sample size. Larger samples
tend to increase the likelihood of finding
a significant difference.
The caffeine treatment group had
high variability, specifically for the total
number of squares entered. The standard
deviation for this measure was 96.79.
Therefore, if the rats were grouped based
on their response to caffeine (little
response, average, high response) and
then compared to the control, there
might have been a significant difference.
A significant difference between the high
response group and control would most
likely occur. Increasing the sample size,
as mentioned above, will help to
diminish
the
variability
amongst
treatment groups.
There are limitations of the
present study. First, there was only one
dose of each drug used, which raises a
possibility that different amounts of
caffeine and taurine could have an
additive effect. The best way to
determine this would be to replicate the
study, but add a lower and higher dose of
each drug, in addition to the one used.
Even though this proposal will
dramatically increase the number of
pairing between the drugs, it will allow a
detailed analysis to better conclude if a
Each rat experienced all four
conditions, thus experienced the openfield on four separate occasions, which
allows the rat to habituate to the field.
The more the rat experiences the openfield, more the locomotion tends to
decrease (Walsh and Cummins, 1976).
180
The procedure used in this experiment
corrected for this habituation by each
group
experiencing
the
different
treatments in a varied order, thus
eliminating order effects.
the drugs, and longer habituation
periods in the exercise wheel). This study
provides results that justify the inclusion
of caffeine in energy drinks, but does not
provide evidence for a synergistic
relationship between caffeine and
taurine.
There was a minimum of 36 hours
in between injections; however the
taurine’s half-life exceeds this time
frame. Future studies should increase the
time in between injections in order for
taurine to be almost completely
metabolized before the next injection
occurs. There may have been taurine
present from the previous injection when
the next injection occurred, which may
have affected the results. However,
administering the treatments in different
orders for each group should have taken
this into consideration. Furthermore, by
increasing time between injections, the
rat will also have a longer period in
between the open-field test, which may
decrease habituation effects.
Works Cited
Alford C, Cox H, Wescott R (2001) “The
effects of red bull energy drink on human
performance and mood.” Amino Acids
21:139-50.
Amp
Energy
website,
2010.
http://www.ampenergy.com.
Accessed
on April 14, 2010.
Angelucci MEM, Cesário C, Hiror RH,
Rosalen PL, Da Cunha CD (2002) “Effects
of caffeine on learning and memory in
rats tested in the Morris water maze.”
Braz J Med Biol Res 35:1202-8.
Imagawa et al. (2009) found an
additive
effect
on
endurance
performance when caffeine and taurine
were supplemented in drinking water for
two weeks. Therefore, if the rats received
daily injections of taurine and caffeine
for two weeks there maybe an additive
effect on activity levels compared to
caffeine alone.
Baskin SI, Hinkamp DL, Marquis WJ,
Tilson HA (1974) “Effects of taurine on
psychomotor activity in the rat.”
Neuropharmacology 13:591-4.
Baum M, Weiss M (2001) “The influence
of a taurine containing drink on cardiac
parameters before and after exercise
measured by echocardiography.” Amino
Acids 20:75-82.
In summary, the present study
provides further research into the
interaction between caffeine and taurine.
Even though there wasn’t a significant
difference between the combined group
and caffeine, there might be under
different experimental parameters (e.g.,
use of younger rats, varying dosages of
Benrabh H, Bourre JM, Lefauconnier JM
(1995) “Taurine transport at the bloodbrain barrier: an in vivo brain perfusion
study.” Brain Res 692:57-65.
Bichler A, Swenson A, Harris MA (2006)
“A combination of caffeine and taurine
181
has no effect on short term memory but
induces changes in heart rate and mean
arterial blood pressure.” Amino Acids
31:471-6.
El Idrissi E, Boukarrou L, Heany W,
Malliaros G, Sangdee C, Neuwirth L
(2008) “Effects of taurine on anxiety-like
and locomotor behavior of mice.” In:
Taurine 7 (Advances in Experimental
Medicine and Biology) (Azuma J et al.,
eds), pp207-15. New York: Springer.
Birdsall
TC
(1998)
“Therapeutic
applications of taurine.” Alternative Med
Rev 3:128-136.
Energy Fiend (2007) The top 15 energy
drinks.
http://www.energyfiend.com/2007/04/th
e-15topenergy-drink-brands. Accessed on
November 10, 2009.
Chepkova AN, Sergeeva OA, Haas HL
(2005) “Long-lasting enhancement of
corticostriatal transmission by taurine:
role of dopamine and acetylcholine.” Cell
Mol Neurobiol 25:767-76.
Ferré S, Fuxe K, von Euler G, Johansson
B, Fredholm BB (1992) “Adenosinedopamine interactions in the brain.”
Neuroscience 51:501-12.
Choleris E, Thomas AW, Kavaliers M,
Prato FS (2001) “A detailed ethological
analysis of the mouse open-field test:
effects of diazepam, chlordiazepoxide
and an extremely low frequency pulsed
magnetic field.” Neurosci Biobehav Rev
25:235-60.
Fine BJ, Kobrick JL, Lieberman HR,
Marlowe B, Riley RH, Tharion WJ (1994)
“Effects of caffeine or diphenhydramine
on visual vigilance.” Psychopharmacology
114:233-8.
Clauson KA, Shields KM, McQueen CE,
Persad N (2008) “Safety issues associated
with commercially available energy
drinks.” J Am Pharm Assoc 48:e55-63.
Frewer LJ, Lader M (1991) “The effects of
caffeine on two computerized tests of
attention
and
vigilance.”
Hum
Psychopharmacol Clin and Exp 6:119-28.
Clubley M, Bye CE, Henson TA, Peck
AW, Riddington CJ (1979) “Effects of
caffeine and cyclizine alone and in
combination on human performance,
subjective effects and EEG activity.” Br J
Clin Pharmacol 7:157-63.
Higgins GA, Jones BJ, Oakley NR (1992)
“Effect of 5-HT1A receptor agonists in
two models of anxiety after dorsal raphe
injection.” Psychopharmacology 106:2617.
Einat H, Yuan P, Manji HK (2001)
“Increased anxiety-like behaviors and
mitochondrial dysfunction in mice with
targeted mutation of the Bcl-2 gene:
further support for the involvement of
mitochondrial function in anxiety
disorders.” Behav Brain Res 165:172-80.
Ho, M. “For Red Bull, it’s here, there and
everywhere.” Washington Post, August
23,
2006.
http://www.washingtonpost.com/wpdyn/content/article/2006/08/22/AR2006
082201516.html. Accessed on November
10, 2009.
182
Hruska RE, Thut PD, Huxtable RJ,
Bressler R (1975) “Suppression of
conditional drinking by taurine and
related compounds.” Pharmacol Biochem
Behav 3:593-9.
A2A receptors in the motor effects of
caffeine after is acute and chronic
administration.”
Neuropsychopharmacology 28:1281-91.
Kerr JS, Sherwood N, Hindmarch I (1991)
“Separate and combined effects of the
social
drugs
on
psychomotor
performance.”
Psychopharmacology
104:113-9.
Hughes JR, McHugh P, Holtzman S
(1998) “Caffeine and schizophrenia.”
Psychiatr Ser 48:1415-7.
Huxtable RJ (1992) “Physiological actions
of taurine.” Physiol Rev 72:101-63.
Kennedy DO, Scholey AB (2004) “A
glucose-caffeine
‘energy
drink’
ameliorates subjective and performance
deficits during prolonged cognitive
demand.” Appetite 42:331-3.
Huxtable, RJ (2000) “Expanding the
circle 1975-1999: sulfur biochemistry and
insights on the biological functions of
taurine.” In: Taurine 4: Taurine and
Excitable Tissues (Corte LD et al., eds),
pp1-26. New York: Springer.
Kiefer D, Pantuso T (2003) “Panax
Ginseng.” Am Fam Physican 68:1539-1542.
Imagawa TF, Hirano I, Utsuki K, Horie
M, Naka A, Matsumoto K, Imagawa S
(2009) “Caffeine amd taurine enhance
endurance performance.” Int J Sports
Med. 30:485-8.
Kim W (2003) “Debunking the effects of
taurine in Red Bull energy drink.”
Nutrition Bytes 9:1-7.
Kong WX, Chen SW, Li YL, Zhang YJ,
Wang R, Min L, Mi X (2006). “Effects of
Taurine on rat behavior in three anxiety
models.” Pharmacol Biochem Behav
83:271-6.
Johnson, CK. “Caffeine-Stoked energy
drinks worry docs.” The Washington
Post,
October
29,
2006.
http://www.washingtonpost.com/wpdyn/content/article/2006/10/29/AR20061
02900290.html. Accessed on November
14, 2009.
Lieberman HR, Wurtman RJ, Emde GG,
Roberts C, Coviella IL (1987) “Effects of
low doses of caffeine on human
performance
and
mood.”
Psychopharmacology 92:308-12.
Johnson LC, Spinweber GL, Gomez SZ
(1990) “Benzodiazepines and caffeine:
effect
on
daytime
sleepiness,
performance,
and
mood.”
Psychopharmacology 101:160-7.
Lister RG (1990) “Ethologically-based
animal models of anxiety disorders.”
Pharmacol Ther 46:321-40.
Karcz-Kubicha M, Antoniou K, Terasmaa
A, Quarta D, Solinas M, Justinova Z, et al.
(2003) “Involvement of adenosine A1 and
Lord, J. “Energy drinks’ sales explode
despite concerns.” News and Tribune,
January
5,
2007.
http://news183
tribune.net/homepage/x519361428/Energ
y-drinks-sales-explode-despiteconcerns?keyword=leadpicturestory.
Accessed on November 10, 2009.
Reissig CJ, Strain EC, Griffiths RR (2009)
“Caffeinated energy drinks- a growing
problem.” Drug Alcohol Depend 99:1-10.
Rockstar Energy Drink website, 2010.
http://www.rockstar69.com/
Accessed
on April 14, 2010.
Miller, KE (2008) “Wired: energy drinks,
jock identity, masculine norms and risk
taking.” J Am Coll Health 56:481-9.
Rusted J (1999) “Caffeine and cognitive
performance: effects on mood or mental
processing?” In: Caffeine and Behavior:
Current Views and Research Trends
(Gupta BS, Gupta U, eds), pp221-30. Boca
Raton, Florida: CRC Press.
Morgan KJ, Stults VJ, Zabik ME (1982)
“Amount and dietary sources of caffeine
and saccharin intake by individuals ages
5 to 18 years.” Regul Toxicol Pharmacol
2:296-307.
Monster
Energy
website,
http://www.monsterenergy.com.
Accessed on April 14, 2010.
2010.
Sanberg PR, Ossenkopp KP (1977) “Doseresponse effects of taurine on some open
field
behaviors
in
the
rat.”
Psychopharmacology 53:207-9.
Nosek K, Dennis K, Andrus BM,
Ahmadiyeh N, Baum AE, Woods LC,
Redei EE (2008) “Context and straindependent behavioral response to stress.”
Behav Brain Funct 4:23.
Scholey AB, Kennedy DO (2004)
“Cognitive and physiological effects of an
“energy drink”: an evaluation of the
whole drink and of glucose, caffeine and
herbal
flavouring
fractions.”
Psychopharmacology 176:320-30.
Olive MF (2002) “Interactions between
taurine and ethanol in the central
nervous system.” Amino Acids 23:345-57.
Selhub J, Bagley LC, Miller J, Rosenberg
IH (2001) “B vitamins, homocysteine, and
neurocognitive function in the elderly.”
Am J Clin Nutr 71:614S-620S.
Ramanathan VK, Chung SJ, Giacomini
KM, Brett CM (1997) “Taurine transport
in cultured choroid plexus.” Pharm Res
14:406-9.
Serrano MI, Goicoechea C, Serrano JS,
Serrano-Martino MC, Sánchez E, Martín
MI (2002) “Age-related changes in the
antinociception induced by taurine in
mice.” Pharmacol Biochem Bev 73:863-7.
Rao A, Hu H, Nobre AC (2005) “The
effects of combined caffeine and glucose
drinks on attention in the human brain.”
Nut Neurosci 8:141-53.
Smit HJ, Cotton JR, Hughes SC, Rogers PJ
(2004) “Mood and cognitive performance
effects of “energy” drink constituents:
caffeine, glucose, and carbonation.” Nutr
Neurosci 7:127-39.
Red
Bull
website,
2010.
http://www.redbull.com. Accessed on
April 14, 2010.
184
Smit HJ, Rogers PJ (2000) “Effects of low
doses
of
caffeine
on
cognitive
performance, mood and thirst in low and
higher
caffeine
consumers.”
Psychopharmacology 152:167-73.
Warburton DM, Bersellini E, Sweeney E
(2001) “An evaluation of a caffeinated
taurine drink on mood, memory and
information processing in healthy
volunteers without caffeine abstinence.”
Psychopharmacology 158:322-8.
Smit HJ, Rogers PJ (2002) “Effects of
‘energy’ drinks on mood and mental
performance: critical methodology.” Food
Qual Pref 13:317-26.
Whirley BK, Einat H (2008) “Taurine
trials in animal models offer no support
for
anxiolytic,
antidepressant
or
stimulant effects.” Isr J Psychiatry Relat
Sci 45:11-18.
Snyder SH, Katims JJ, Annau Z, Bruns RF,
Daly JW (1981) “Adenosine receptors and
behavioral actions of methylxanthines.”
Proc Natl Acad Sci USA 78:3260-3264.
Wilson RC, Vacek T, Lanier DL,
Dewsbury
DA
(1976)
“Open-field
behavior in muroid rodents.” Behav Biol
17:495-506.
Walsh RN, Cummins RA (1976) “The
open-field test: a critical review.” Psychol
Bull 83:482-504.
Zeratsky, K (2008) “Taurine in energy
drinks: what is it?” Mayo Medical Clinic.
http://www.mayoclinic.com/health/tauri
ne/AN01856. Accessed on September 20,
2009.
Warburton DM (1995) “Effects of caffeine
on cognition and mood without caffeine
abstinence.” Psychopharmacology 119:6670.
185
Appendix
RAT # _________
TREATMENT GROUP_________
TOT
AL # of Squares Entered _________
TOTAL # of Rears _________
TOTAL Cleaning Time (secs) _________
186
Figure 1: Coding Sheet Used By the Blind Observers.
Test
Open Field
Exercise
Wheel
Measure
Total Sqaures (4
paws)
Rears (front paws
off floor)
Grooming Time
(secs)
% of Perimeter
Squares
Total Rotations
Control
(mean±SEM)
Taurine
(mean±SEM)
Caffeine
(mean±SEM)
Combined
(mean±SEM)
63.7 ± 12.4
76.2 ± 17.9
168.1 ± 30.6
191.8 ± 15.8*, #
7.4 ± 1.9
7.9 ± 1.3
13.5 ± 2.0
14.6 ± 2.2
16.7 ± 4.1
38.2 ± 8.8
65.3 ± 22.8
25.6 ± 4.5
96.7 ± 1.6
92.7 ± 1.7
94.8 ± 1.3
91.5 ± 1.7
0.1 ± 0.1
0.5 ± 0.3
0.7 ± 0.4
0.3 ± 0.2
Table 1: Mean Values For Total Squares, Total Rears, Cleaning Time, % of Perimeter Squares,
and Total Rotations. Taurine (50 mg/kg), caffeine (30mg/kg), combined (same amounts), or
control (0.9% saline) was injected 30 min prior to the start of the open field test. * P < 0.0083
(Bonferroni correction) combined group vs. control. # P < 0.0083 combined group vs. taurine.
Repeated measures ANOVA followed by paired t-test. n = 10.
187
*, #
A. Total Squares Entered
B. Total # of Rears
C. Total Duration of Grooming
D. % of Perimeter Squares
Figure 2: Effect of taurine (50 mg/kg), caffeine (30 mg/kg), control, and the combined treatment
groups on rat behavior in the open-field test: (A) the total number of squares entered; (B) the
total number of rears; (C) the total duration of grooming; (D) the percentage of perimeter squares
entered. All drugs were injected 30 min prior to the start of the open-field test. Results are
expressed as mean ± SEM. Significant differences: compared to control * P < 0.0083; compared
to taurine # P < 0.0083. n = 10.
188
Measuring the Persistence of Output Shocks: A Study of
Output Behavior using ARMA and Monte Carlo Methods
By Michael LoFaso, Class of 2010
This paper examines whether output fluctuations are better characterized as shocks with
measurable and persistent effects or transitory deviations from macroeconomic growth
trends. Building on the work of Nelson and Plosser [1982) and Campbell and Mankiw
[1987), the paper employs ARMA and Monte Carlo techniques to provide robust evidence
for the persistence of shocks. After accounting for the Great Moderation period of the late
20th century, the models suggest that the magnitude and volatility of impulse responses
are comparable, yet slightly smaller than Campbell and Mankiw’s findings. When applied
to the economic crisis beginning in 2007, the paper’s findings of persistence suggest that
the shock will affect economic growth into the foreseeable future, with output levels
remaining below potential capacity well into the next decade.
Introduction
In recent months, the health and
stability of the economy has vaulted to
the forefront of national issues, garnering
vast media, public, and governmental
attention. Needless to say, sorting
through and understanding the many
potential factors and events affecting the
economy are of great interest to
economists, government officials, and
citizens alike. While the short-term
effects of the economic slowdown have
unquestionably caused a number of
difficulties for Americans, in order to
fully understand the current economic
environment, it is important to
understand how the recent shocks to the
United States economy fit within the
larger scope of business cycles and
domestic economic growth.
and persistent effects on the long-term
health of the economy.
The paper’s primary focus will not
be to solve a longstanding debate
regarding the existence of a unit root in
the output process, but rather to
construct
quantitative
measures
persistence. In order to explore the
question of persistence, the paper will
analyze United States GDP data with
several autoregressive and moving
average processes to determine whether
or
not
macroeconomic
shocks
demonstrate persistent and measurable
effects.
Building on the work of Nelson
and Plosser (1982) and Campbell and
Mankiw (1987), the main contributions
of this paper will be to: (1) quantitatively
measure the impact and persistence of
economic shocks using a stationary
ARMA model of the differences in log
GDP, (2) use a Monte Carlo procedure to
The goal and objective of this
paper will be to examine whether shocks
to the economy represent temporary
deviations from macroeconomic growth
trends, or, if the shocks have measurable
189
forecast future prospects of economic
growth in light of the current and
significant GDP shock, and (3) discuss
alternative theoretical interpretations of
persistence.
consistent with the Congressional
Budget Office’s projections for economic
recovery.
Lastly, we conclude with a
discussion about interpretations of
persistence in light of the
paper’s findings.
The paper will begin by discussing
relevant time series topics, including
issues of stationarity, autoregressive
integrated moving average models, and
unit roots. We will then focus specifically
on trend vs. difference stationary
techniques for measuring the persistence
of fluctuations. In this section of the
paper we show that an estimation of the
change in log GDP as a stationary ARMA
process is consistent with a variety of
statistical processes in the level of GDP,
while it also leaves open the possibility
that output fluctuates around a
deterministic trend.
Time Series Theory and Approaches
When we analyze time series data
our first consideration is whether or not
our
variables
demonstrate
stationarity.
Regressions using time series data are
different than regressions using solely
cross-section data mostly because they
might include dynamics: explanatory
variables and/or random shocks might
have effects that persist and there might
be systematic or stochastic trends.
We then turn to a literature
review that discusses business cycle
theory and previous studies that have
tested for unit roots and the persistence
of shocks. The next section presents
results for our statistical models,
providing evidence that shocks remain
persistent into the long-run horizon.
However, the models also suggest that
the magnitude of and volatility of
impulse
response
functions
has
diminished after accounting for the
Great Moderation period.
A time series is stationary when it
represents a random process and its
statistical properties do not vary over
time. In order to achieve unbiased
estimates, it is important to work with a
stationary process (as opposed to a nonstationary process in which statistical
properties are different at different time
periods).
The unit root concept is also
important for the considerations of this
paper. A linear stochastic process has a
unit root if 1 is a root of the time series
process’ characteristic equation. The
presence of a unit root has important
implications for the estimation of ARMA
models. We will first discuss the concept
of roots in the context of our time series
To provide context about output
behavior after the current economic
shock, we then construct a Monte Carlo
experiment and create probability
distributions for forecasts of future
growth paths. We also show that our
statistical approach yields results
190
models, and then explain the statistical
implications of a unit root.
Let us consider a discrete
stochastic process written as an AR (p)
model:
We can then show that this AR(1)
model with a unit root is non-stationary
because its variance is dependent on
time t:
yt = ρ1yt −1 + ρ2 yt−2 + . . . + ρpyt−p + εt.
.
If we express this process in lag
polynomial notation, we have:
For our purposes, the main
implication of a unit root is that a shock
to a unit root (or near-unit root) process
may have permanent effects that would
otherwise decay if the process were
stationary. However, because output can
be written as a difference stationary
model, we are able to analyze the
persistence of shocks by converting the
unit root process into a difference
stationary series that can be analyzed
through our time
series modeling.
(1 − ρ1L − ρ2L2 − . . . − ρp Lp )*yt = εt.
We can then obtain the
characteristic equation by replacing the
lag operator L by a variable (call it ‘m’),
and setting the resulting polynomial
equal to zero:
1 − ρ1m − ρ2m2 − . . . − ρp m = 0.
The characteristic roots are the
values of ‘m” that solve this equation.
A brief overview of autoregressive
integrated moving average (ARIMA)
models will also provide helpful context
for this paper’s empirical analysis.
ARIMA models are characterized by
autoregressive (AR), integrated (I), and
moving average (MA) components.
Specifications
of
autoregressive
integrated moving average models are
generally referred to as ARIMA (p,d,q)
models, where p, d, and q represent the
order of the autoregressive, integrated,
and moving average parts of the model
respectively.
Let us consider, for example, a
first order autoregressive model:
Yt = ρ1Yt-1 + εt
where the process demonstrates a unit
root when ρ1 = 1.1 We can show that this
process is non-stationary by illustrating
that the moments of the stochastic
process depend on the time periods, t. By
repeated substitution, our first order
autoregressive model with ρ1 = 1 can be
written as:
The autoregressive (AR) process
of an ARIMA model is used to forecast
future values of a time series based on
the past values of the series. It does so by
regressing the dependent variable on its
1
In this case, the characteristic equation would
be m-ρ1 = m-1 and the root of the equation is ρ1 =
1
191
lagged values. In its general form, the
autoregressive model can be written as:
When a series demonstrates nonstationarity, integration can be useful if
the process is first-order stationary (i.e.
its
first-order
probability
density
function does not change across different
time periods).
Yt = c +ρ1Yt-1 + ρ2Yt-2 + … + ρpYt-p + εt
where c represents a constant and εt
represents white noise.
Now that we have introduced the
concepts of ARIMA models, we will turn
to a more in depth discussion regarding
the appropriateness of this paper’s
estimation
technique—a
stationary
ARMA (p,0,q) model in the difference in
log GDP—in light of trend vs. difference
stationarity and persistence issues.
The moving average (MA) process
is used to analyze how a weighted
average of the lingering effects of
random shocks move through time. A
general moving average model can be
written as:
Yt = εt + α1εt-1 + α2εt-2 + … + αqεt-q
Statistical Considerations for
Measuring the Persistence of Shocks:
where εt represents the shock in the
current period, εt-1 represents the shock
in the previous period, an so on.
In order to formulate a response
to the main question of this paper—how
does a current shock to output affect
future output over a long-term
horizon?—we will begin by discussing
the appropriateness of our statistical
approach.
We can then combine these two
components into one model, the
autoregressive moving average model:
Economic output in the United
States has historically drifted upward. In
order to account for this upward drift,
previous studies have used two different
approaches: detrending and differencing.
A non-stationary process that becomes
stationary after accounting for a trend
(i.e. detrending) is called a trendstationary process, while a nonstationary
process that can be made stationary after
differencing is referred to as a differencestationary process.
The integrated component of the
model is used as an initial differencing
step in cases where data show evidence
of
non-stationarity.
For example,
consider an ARIMA (1,1,1) model, in
which we incorporate a first difference
into a model with one AR parameter and
one MA parameter:
Yt - Yt-1 = c + ρ1(Yt-1 - Yt-2) + α1εt-1 + εt.
More generally, we can write a
first-order integrated ARMA process as:
Detrending
macroeconomic
processes has been a technique used by
many, such as Blanchard (1981) to
ΔYt = c + ρi(ΔYt-i) + αiεt-i + εt.
192
prepare data for time series analysis.
However, as Campbell and Mankiw
(1987) explain, this technique is not
appropriate for answering the question
of
persistence
we
have
raised.
Detrending the data series leads the
process to be trendreverting, essentially
eliminating the possibility that a current
output shock will have a long-run effect
on output. That is, detrending the time
series would presuppose the answer to
our primary question at an infinite
horizon.
series before estimation would severely
bias the process towards zero.
Similarly, Frisch and Waugh
(1933) note that including time as an
explanatory variable is numerically
identical to detrending any of the
variables before estimation, and thus
inappropriate for our purposes.
The differencing method provides
an alternative approach to address the
upward drift present in output data.
According to Campbell and Mankiw, the
differenced series of log GNP "appears
stationary, allowing one to invoke
asymptotic
distribution
theory"
2
(Campbell and Mankiw 1987). Figures 1
and 2 illustrate the log GDP series in
Campbell and Mankiw (1987)
illustrate this by exploring the possibility
that output follows a random walk with
drift:
Yt = α + Yt-1 + εt
where Yt (the log of GDP) is
characterized by an AR(1) process in
which α is the drift term representing
long-run growth. As Nelson and Kang
(1981) point out, detrending the output
2
This assertion remains accurate when
we account for our updated time series
and use of log GDP as our
measure of growth.
193
comparison to the differenced series of
log GDP.
φ(L)ΔYt =θ(L)εt ,
where
Campbell and Mankiw then
consider the potential issues involved
with using differenced data. They first
consider an IMA (1,1) process:
and
Yt - Yt-1 = α + εt - θεt-1
θ(L) =1+θ1L +θ2L2 + ...+θqLq .
Here, a unit impulse in Yt changes
one's forecast of Yt+n by (1-θ). In this
case, θ represents measure of persistence
and, depending on its value, allows for a
current shock to have a range of effects—
anywhere from virtually no effect to a
significant long-term effect—on future
outcome forecasts.
By deriving the moving average
representation of this equation Campbell
and Mankiw show that the differenced
model leaves open the question of
whether the level of log GDP is stationary
(Campbell and Mankiw 1987). That is,
the estimation employed in this paper
remains consistent with processes in the
level of log GDP
φ(L) =1−φ1L −φ2L2 − ...−φ pLp ,
This
paper’s
estimation
technique—modeling the change in log
GDP as a stationary ARMA (p,0,q)
process—is written as follows:
Campbell and Mankiw (1987) also
discuss the possibility that outcome is
not characterized by a random walk with
drift. However, they find that if the
194
followed
by
economic
activities3,
similarly
general
recessions,
contractions, and revivals which merge
into the expansions phase of the next
cycle; this sequence of changes is
recurrent but not periodic; in duration
business cycles vary from more than one
year to ten or twelve years; they are not
divisible into shorter cycles of similar
character with amplitudes approximating
their own.
output process is best captured by
traditional business cycle theory—where
growth is stationary around a trend—
estimating the difference in log GDP
does not bias estimates toward
misleading or excessive persistence. In
other words, the model leaves open the
possibility that output is stationary
around a deterministic linear trend.
Our estimation equation is
bounded by the stability condition that
the inverse characteristic roots of the
equation (i.e. the roots of 1 − ρ1L − ρ2L2 −
. . . − ρp Lp), both real and complex, lie
outside the unit circle. In this case, our
process will be stable and stationary.
However, if the inverse characteristic
roots demonstrate unity or lie within the
unit circle, the process may not be
stationary.
Robert Lucas (1977) characterizes
this notion of a business cycle by asking,
“Why is it that, in capitalist economies,
aggregate variables undergo repeated
fluctuations about trend, all of essentially
the same character?” This view of
economic growth suggests that the
effects of any given technological shock
to output will eventually approach zero,
and that output is characterized by a
trend stationary process.
Literature Review
This section will discuss literature
regarding the trend stationary and
random walk with drift theories behind
time series economic growth models. We
will summarize how time series growth
concepts have developed through
literature, and then turn to the findings
and interpretations of previous studies.
Conventional
business
cycle
literature has since evolved to explain
that fluctuations in output represent
temporary
deviations
from
a
deterministic trend. The implications of
this conventional view are described by
Romer: If output movements are
temporary
deviations
around
a
deterministic trend, then “output growth
will tend to be less than normal when
output is above its trend and more than
In their analysis of the economy
during expansions and contractions,
Burns and Mitchell (1946) explain the
general characteristics of business cycles:
A cycle consists of expansions occurring
at about the same time in many
3
While literature has examined the
empirical relationship between the
aggregate business cycle and various
aspects of the macroeconomy—such as
interest rates, employment, investment,
etc.—this paper will focus
solely on the structural behavior of GDP.
195
normal when it is below its trend”
(Romer 1996).
long horizon, the economy should return
to its natural rate; the time series for
output should be trend-reverting
(Campbell and Mankiw 1987).
However, competing literature
has also argued that changes in
technology may have a permanent effect
on long-run growth. According to
Romer, “An innovation today may have
little impact on the likelihood of
additional innovations in the future, and
thus on the expected behavior of the
growth of technology in the future. In
this case, the innovation raises the
expected path of the level of technology
permanently” (Romer 1996).
However,
based
on
their
examination of quarterly postwar United
States Data (spanning from 1947:1 to
1985:4) Campbell and Mankiw are
skeptical about this implication.
Before considering their results,
let us turn to Cochrane (1988) who
provides guidance on the interpretation
of the effects following a one-unit shock
to output. Cochrane explains that
different results lead to different
conclusions
about
the
structural
components of output growth (i.e.
whether or not output should be
characterized by a trend-stationary
process or if it is better captured by a
random walk). According to Cochrane, if
a one-unit shock to GNP affects forecasts
in the far future "by one unit, it finds a
random walk; if by zero, it finds a trend
stationary process" (Cocrane 1988).
The question of persistent output
fluctuations was addressed by McCulloch
(1975) and Nelson and Plosser (1982) who
consider whether output shocks have
permanent
effects.
Estimating
regressions that test for trend-reversion
versus permanent shocks, Nelson and
Plosser conclude that they cannot reject
the null hypothesis that fluctuations
have a permanent component. That is,
they cannot reject the existence of a unit
root in the autoregressive representation
of time series.
However, the resulting effects of a
one-unit shock do not always yield such
a simplistic interpretation. For example,
a one-unit shock may also affect forecasts
in the future by "numbers between zero
and one, characterizing a series that
returns toward a 'trend' in the far future,
but does not get all the way there. Or it
can find a number greater than one,
characterizing a series that will continue
to diverge from its previously forecast
values following a shock" (Cochrane
1988).
While the work of Nelson and
Plosser began to challenge conventional
business cycle theory, Campbell and
Mankiw (1987) advanced persistence
literature by quantitatively measuring
the permanent component of output
fluctuations. Campbell and Mankiw
contend: If fluctuations in output are
dominated by temporary deviations from
the natural rate (of growth), then an
innovation in output should not
substantially change one’s forecast of
output, in, say, five or ten years. Over a
196
Campbell and Mankiw's results
provide evidence that shocks to output
are generally followed by further output
movements in the same direction. They
estimate several models in which a 1
percent innovation to real GNP should
change one's forecast of GNP over a long
horizon by over 1 percent. While they
find "some evidence of short-run
dynamics that makes GNP different from
random walk with drift, the long-run
implications of (their) estimates suggest
that shocks to GNP are largely
permanent” (Cambell and Mankiw 1987).
15.6 percent—on output being trend
stationary, suggesting instead that the
current shock to economy may have
persistent negative effects.
Paul Krugman (2009), who
challenges the unit root hypothesis set
forth by Mankiw and Campbell, provides
a distinctively different commentary:
I always thought the unit root thing
involved
a
bit
of
deliberate
obtuseness…For one thing is very clear:
variables that measure the use of
resources,
like
unemployment
or
capacity utilization, do NOT have unit
roots: when unemployment is high, it
tends to fall. And together with Okun’s
law, this says that yes, it is right to expect
high growth in future if the economy is
depressed now.
Greg Mankiw and Paul Krugman's
comments on recent forecasts made by
the Council of Economic Advisors (CEA)
highlight the interpretational differences
between the trend stationary—or
business
cycle—theory
and
the
difference stationary—or unit root—
theory. According to the March 2009
CEA statement: A key fact is that
recessions are followed by rebounds.
Indeed, if periods of lower-than- normal
growth were not followed by periods of
higher-than-normal
growth,
the
unemployment rate would never return
to normal.
Whereas Mankiw sees the forecast
estimates as optimistic, Krugman rejects
the notion that the effects of a negative
shock in the economy will be persistent.
Based on previous literature studies, it
seems as if there is evidence to argue
for both sides of the debate. Unit root
literature has suggested that indeed,
GDP may have a unit root; James Stock
and Mark Watson (1998) find GDP to
have a near unit root measured within an
interval of 0.96-1.10 for a 90 percent
confidence level.
The CEA continues, "because we
are now experiencing below-average
growth, we should raise our growth
forecast in the future to put the economy
back on trend in the long run" (CEA
2009)
Dissenters, however, argue that
while the time-series econometrics may
demonstrate a unit root, the actual
implications of the findings are not
realized in real economic growth
patterns.
Mankiw disagrees with the CEA
premising its economic forecast—which,
as of March 2009, suggested that GDP
will shrink only by 1.2 percent in 2009
and will grow between now and 2013 by
Cocharane
(1988)
provides
important perspective on unit root
197
findings within a time series process.
According to Cochrane:
fluctuations. According to Romer (1996),
there are two major considerations—one
statistical and one theoretical—that
should not be overlooked. He begins by
explaining the former:
Since a series with a unit root is
equivalent to a series that is composed of
a random walk and a stationary
component, tests for a unit root are
attempts to distinguish between series
that have no random walk component
(or for which the variance of shocks to
then random walk component is zero)
and series that have a random walk
component (or for which the variance of
shocks to the random walk component is
between zero and infinity) (1988).
The statistical problem is that it is
difficult to learn about long-term
characteristics of output movements
from data from limited time spans. The
existence of a permanent component to
fluctuations and the asymptotic response
of output to an innovation concern
characteristics of the data at infinite
horizons. (Romer ).
Given this description, it becomes
clear why tests for a unit root are not
particularly powerful: it is difficult to tell
a stationary series from a stationary
series with a very small random walk
component.
Romer continues by discussing
the risks of using the relationship
between current output growth and its
most recent lagged values to make
inferences about output's long-run
behavior:
Similarly, Campbell and Perron
(1991) argue that it can be very difficult to
distinguish between a unit root and near
unit root process, suggesting that the
usefulness of unit root tests is “not to
uncover some true relation,” but rather
to “impose reasonable restrictions on
data and suggest what asymptotic
distribution theory gives the best
approximation to the finite-sample
distribution of coefficient estimates and
test statistics” (Campbell and Perron
1991).
Suppose, for example, that output
growth is actually AR-20 instead of AR-3,
and the coefficients on the 17 additional
lagged values of Δ ln y are all small, but
all negative. In a sample of plausible size,
it is difficult to distinguish this case from
the AR-3 case. But the long-run effects of
an output shock may be much smaller
(Romer 1996).
That being said, this difficulty
arises from the brevity of the sample, and
not from the differencing procedures
used to estimate autoregressive and
moving average parameters.
Thus, the focus of this paper will
not be to solve a longstanding debate
regarding the existence of a unit root in
the output process, but rather to
construct measures of persistence.
Romer (1996) provides insight into the
econometric issues to consider when
investigating
the
persistence
of
There are two factors that help to
attenuate this concern: (1) with the
benefit of a significantly longer time
series dataset, we will have the flexibility
to measure longer ARMA models with
greater variation and (2) when using
information provided by the current
198
Parameterization
GDP shock to simulate the output
growth path with a Monte Carlo
procedure, we will build a probability
distribution to capture the likelihood of
different growth trajectories.
To estimate our ARMA process, φ(L)ΔYt
=θ(L)εt , for the change in log GDP we
must determine the appropriate number
of autoregressive and moving average
parameters to analyze. We will use
Akaike Information Criteria (AIC) and
Schwartz Bayesian Information Criteria
(SBC)
as
guidance
for
our
parameterization. Both the AIC and SBC
selection criteria penalizes a model for
over-parameterization.
As Romer points out, there is also
a theoretical issue worth considering
when investigating the persistence of
fluctuations. Namely, “there is only a
weak case that the persistence of output
movements, even if it could be measured
precisely, provides much information
about the driving forces of economic
fluctuations” (Romer 1996).
The Akaike Information Criteria is
calculated as:
-2*ln(likelihood) + 2k,
While this may be true, the paper
will attempt to explain its results in light
of a theoretical discussion regarding the
potential effects of fiscal and monetary
policies, Keynesian theory, and real
business cycles on economic growth.
and the Schwartz Bayesian Information
Criteria is calculated as:
-2*ln(likelihood) + ln(N)*K,
where K equals the number of
parameters in the model and N equals
the number of observations.
This discussion will also be useful
when considering the hypothesis of
Blanchard and Watson (1984), who
examine business cycles within the
United States. They suggest that another
key area of consideration is “whether
fluctuations in economic activity are
caused by an accumulation of small
shocks, where each shock is viewed in
isolation, or rather whether fluctuations
are due to infrequent large shocks.”
While the paper's measures of
persistence are not suited to sort out the
precise causes of shocks, we will again
attempt to address this issue through a
theoretical discussion.
To provide robust evidence about
the persistence of shocks we will
consider several different combinations
of autoregressive and moving average
parameters.
Models
with
more
parameters allow for greater flexibility
with impulse response functions, but
also
run
the
risk
of
overparameterization. The selection criteria
will be used to provide context for each
model in relation to the other models
estimated.
Results
To analyze the persistence of
output fluctuations, we estimate ARMA
199
processes for the difference of log GDP
with up to three AR and three MA
parameters. We consider 15 total models,
the most simple being the MA(1) and
AR(1) processes, and the most complex
being the ARMA (3,3) model. The
quarterly time series data is reported as
billions of chained 2005 dollars and
spans 1947:1 to 2009:3.
of output persistence, rather than use a
single model as the basis for evidence
and interpretation. From the table, it also
becomes evident that models more
complex than our ARMA (3,3) will be
penalized for having too many
parameters. Thus our selection for our
most complex model—one with three AR
parameters and three MA parameters—
seems appropriate.
Table I reports the AIC and BIC
selection criteria for each of the 15
models. While our ARMA (3,2) model
minimizes the value of the selection
criteria—providing evidence that it is the
most parsimonious model—our focus
with Table I is shifted towards the fact
that all of the models have comparable
AIC and SBC values. This is important
because we will use each of the models to
provide a statistically robust account
Table II reports the inverse
characteristic AR and MA roots for each
of the models. These results seem
interesting in light Campbell and
Mankiw’s (1987) findings.
Using a data set that spans from
1947:1 to 1985:4, Campbell and Mankiw
find that high order MA processes have
unit roots that are cancelled by near-unit
AR roots, causing
shocks to die out
over
a
long
horizon.
While
Campbell
and
Mankiw report
moving average
roots
with
perfect unity and
corresponding
autoregressive
roots near unity
for the ARMA
(1,3),
ARMA
(2,3), and ARMA
(3,3) models, we
do not find this
phenomenon.
The
major
implication
of
this finding is
200
(3,3) model aligns
with
our
interpretation
of
the Akaike and
Schwartz Bayesian
selection
criteria:
adding more than
three AR and three
MA
coefficients
leads
to
overparameterizatio
n and imprecise
estimates.
that all of our ARMA models provide
evidence for the persistence of shocks
over a long horizon because we do
not find the cancellation of near-unit
AR and MA
roots.
The parameter estimates for
each of the ARMA4 models are
reported in Table III.
The results are consistent with our
statistical theory that more complex
models allow for greater flexibility, as
well as greater persistence, in our
parameter estimates.
When we move from the
simpler models to more complex
models
by
adding
additional
parameters, the magnitude of our
coefficients generally become larger,
yet
often
remain
statistically
significant. That being said, the
insignificance reported for the ARMA
We now turn to our impulse response
functions, reported in Table IV, to
provide
an interpretation of these coefficients.
For each model, we calculate impulse
4
The large magnitude of this inverted root
suggests that the AR parameter is for the ARMA
(1,3) model is small, and most of the effects
captured by the model are attributed to the
moving average components.
201
responses for 80 quarterly periods into
the future. The impulse responses are
calculated by
imposing a one percent shock to the
error term of each respective model in
the current time period (period zero). In
the ensuing time periods, it is assumed
that the error term assumes a value of
zero, so that the persistence of the shock
in the current period is only carried by
parameter estimates reported in Table
III.
There are a number of important
differences to consider when comparing
these
results with Campbell and Mankiw’s
(1987) findings. First, as noted while
discussing Table II, we do not find
evidence that high order MA processes
have unit roots are cancelled by nearunit AR roots, and thus, we do not find
that shocks die out over a long horizon
for any of the models. We also find that
the ARMA (2,3) and ARMA (3,2) process
do not completely stabilize by the end of
the 80-quarter period.
The functions reported in Table
IV represent the level of the impulse
response. The level is calculated by
summing the change in the impulse
response function from the current
period until the reported period. All of
the impulse responses suggest that a 1
percent shock to output in the current
period leads to over a 1 percent increase
in the forecast of output over a long-run
horizon. That is, the impulse responses
increase above 1 percent and settle
between 1.28 and 1.69 percent over the
course of 10 to 20 years into the future.
Figures 1A and 1B illustrate the
impulse response paths of varying ARMA
models from one quarter to the next. The
ARMA (3,0) model—which has an
impulse response similar to the other
pure
autoregressive
models—
demonstrates a path that stabilizes over a
short period of time (roughly 1 to 2 years
for each of the pure AR models).
With more parameters and fewer
limitations on the impulse response
path, the ARMA (2,2) model depicts a
function that oscillates for approximately
7-10 years before reaching an equilibrium
202
or stable path. The ARMA (2,3) process,
however, does not reach a steady state
even after 20 years. Although its path is
slowly collapsing on an equilibrium, the
effects of a 1 percent shock in the current
period fluctuate over a longrun horizon.
absorbed by an economy characterized
by greater stability.5
One possible explanation for this
finding is that our current dataset
includes The Great Moderation- a period
from the mid-1980’s until the early 2000’s
in which the volatility of economic
variables such as GDP, industrial
production, and unemployment declined
(Bernanke 2004). According to Bernanke,
the decrease of volatility associated with
economic shocks can be explained by
structural
changes,
improved
macroeconomic policies, and good luck.
Bernanke classifies structural changes as
changes in economic institutions,
technology, business practices, or other
structural features of the economy that
improve the ability of the economy to
absorb shocks. He notes, for example,
that improved management of business
inventories, made possible by advances
in computation and communication, has
reduced the amplitude of fluctuations in
inventory stocks, which in earlier
decades played an important role in
cyclical fluctuations” (Bernanke 2004).
When we compare this paper’s
findings with those of Campbell and
Mankiw, we uncover two important
developments: (1) unconstrained by the
autoregressive and moving average near
unit root cancellation phenomenon, we
can provide additional robust evidence in
favor of persistent output shocks, and (2)
while shocks demonstrate persistent
effects, the magnitude and volatility of
these effects has diminished.
As we have discussed the
statistical properties that help to explain
our first finding—namely, we no longer
see a near unit root cancellation
phenomenon—we will turn our attention
to our second finding. For models with
no unit moving root average—all models
except the ARMA (1,3), ARMA (2,3), and
ARMA (3,3)—Campbell and Mankiw find
that the “impulse responses increase
above and then settle between 1.3 and 1.9
at about the eighth quarter, remaining
there even at ten or twenty years”
(Campbell and Mankiw 1987). However,
when we compare these results to our
impulse responses, we find that, on
average, persistence measures are both
smaller and less volatile: impulse
responses settle between about 1.2 and 1.7
percent and demonstrate smaller
fluctuations from one quarter to the
next. That is, a shock to output is
Describing the implications of
improved
macroeconomic
policies,
Bernanke suggests that corresponding
changes in the volatilities of output
growth and inflation gives some
5
While it is possible that this paper’s findings
and those of Campbell and Mankiw may
impacted by a
change in data compilation methods (a chaintype price index is now used in favor of
Paasche/Laspeyres indices) or switching from
GNP to GDP, the impact is expected to be small
given that the differences in results using
different methods were found to be very small, if
not negligible.
203
credence to the idea that better
monetary policy may have been a major
contributor to increased economic
stability (Bernanke 2004). According to
Bernanke, since “few disagree that
monetary policy has played a large part
in stabilizing inflation,” the fact that
“output volatility has declined in parallel
with inflation volatility, both in the
United States and abroad, suggests that
monetary policy may have helped
moderate the variability of output as
well” (Bernanke 2004).
The Great Moderation more as a period
of increased economic stability rather
than a long-term insurance policy
against volatility and the persistence of
output shocks. With this in mind, we will
turn to a Monte Carlo experiment that
forecasts future output in light of the
current output shock.
Monte Carlo Experiment
This section of the paper uses a
Monte Carlo experiment to replicate a
datagenerating process. We simulate a
data set with the essential characteristics
of or time series process in order to
forecast output projections after the
current economic shock.
The third possible explanation for
a period of moderation, good luck,
contends that The Great Moderation did
not result from structural changes or
improved macroeconomic policies, but
rather the frequency and magnitude of
shocks declined within the time period,
perhaps by chance as opposed to a more
stable
economy.
Quantitatively
measuring the factors of The Great
Moderation, Stock and Watson estimate
that the dampening of GDP volatility is
“attributable to a combination of
improved policy (20-30%), identifiable
good luck in the form of productivity and
commodity price shocks (20-30%), and
other unknown forms of good luck that
manifest themselves as smaller reducedform forecast errors (40- 60%)” (Stock
and Watson 2002).
After calculating the model error
for each of our ARMA processes, we use a
random number generator—assuming
that each model’s error has a mean of 0
and is normally distributed6—to build a
probability distribution of errors for each
model (the variance is derived from the
different random errors generated for the
various iterations of each model). We
generate error terms for 80 periods, and
using our parameter estimates from
Table III, construct a sequence for each
specified model. We are able to capture
some of the effects of the current shock
by tying our forecasts in the early periods
to the estimated residuals of recent
quarters.
While The Great Moderation
helps to reconcile our findings with those
of Campbell and Mankiw’s, we cannot
overlook the significant influence of
output volatility that has accompanied
shocks caused by the economic and
financial crisis beginning in 2007. The
recent economic crisis leads us to view
6
The validity of this assumption is checked by
generating the density function for the residuals
estimated
from our time series. The density function shows
that our error terms take on close to a standard
normal distribution.
204
We then repeat this process to
generate approximately 1,000 iterations
for each model, with the goal of ensuring
that the statistical properties of the
constructed sequence are in accord with
the true distribution (Enders 2004).
205
206
After
generating
numerous
iterations for a sequence of 80 quarters
into the future, we then calculate a mean
and band it by standard deviations in
order to construct a confidence interval
around our point estimates. Figures 2A2C provide illustrations of Monte Carlo
experiments for three different ARMA
models.
shock. That is, we remove the negative
shock and forecast output as if it were to
continue a stable growth path.
The simple MA(1) model shows
that forecasted GDP continues to grow at
levels lower than potential GDP. That is,
the slack between potential and
forecasted GDP remains throughout the
foreseeable future. In this case, a
negative unexpected change in output
will cause future growth levels of output
be lower than in the period before the
shock.
The figures graphically represent
the level forecast for the change in log
GDP. Figure 2A shows the point
estimates for an ARMA (2,1) model
banded by a 1 and 2 standard deviation
distribution. On the graph, time period 1
represents the first forecasted period,
and all time periods leading up to that
show the path of actual log GDP.7 As we
can see, the standard deviation bands
around the point estimates widen over
the period of our forecast—this is not
surprising given that we expect estimates
to become less precise as we move
further into the long-run horizon.
However, the ARMA (3,3) model
tells a slightly different story.8 After an
unexpected drop in observed output,
output slowly recovers and eventually
reaches its growth rates from before the
shock. Compared to the ARMA (0,1)
model, the additional parameters in this
process likely pick up a greater portion of
the historical upward trend in economic
growth, thus allowing for an output path
that eventually resumes its natural
rate of growth in the time period
following the shock.
While Figures 2B and 2C provide
further evidence for the persistence of
shocks reported with our impulse
response functions, the ARMA (0,1) and
ARMA (3,3) models tell slightly different
stories about the trajectory of output
over the long-run horizon. In both
figures, we add a dashed line to simulate
the point estimates for output growth in
the absence of the current economic
7
The graphs show 100 previous observations only
to shift the focus onto the Monte Carlo forecast.
It is
important to not that 251 observations dating
back to 1947 were used to construct the statistical
properties
of the Monte Carlo experiment.
8
The 1 standard deviation lines are moved only to
more clearly show the point estimate growth
path.
207
While
this
Monte
Carlo
experiment is helpful for our persistence
analysis, it is important to mention a few
limitations of the experiment. First, given
the volatility of the current economic
environment, it is difficult to claim with
certainty that the negative shock has
ended. It is possible that the economy
may still experience unexpected losses
(or gains) in the near future that would
be important to consider when
attempting to forecast output over a long
horizon.
To provide further perspective on
our findings, we compare the forecasts
generated by our econometric models
with the Congressional Budget Office’s
(CBO) judgment-based projections of
potential and projected GDP.10
Figure 3 maps the ARMA (2,1)
models forecasted output path on top of
the CBO’s potential GDP figures.
For comparison purposes, the CBO
projection—which is also mapped on top
of the potential GDP—is shown in the
lower right panel of the figure. As we can
see, our model predicts a similar recovery
path to the projections made by the CBO
Additionally, a limitation of any
Monte Carlo experiment is that “the
results are specific to the assumptions
used to generate the simulated data”
(Enders 2004). Thus, when we consider
the ARMA (3,2) model—which has two
complex roots within the unit circle—we
find that our results are difficult to
interpret and not overly useful for our
analysis.9
growth rates prior to the shock. This result seems
unlikely given the structure of the experiment,
and may
be attributable to the model’s complex roots.
10
The CBO’s economic projections are based on
trends in the labor force, productivity, and
saving. In
preparing its forecasts, the CBO examines and
analyzes recent economic developments and
relies on the
advice of a distinguished panel of advisers [CBO
2010].
9
For this model, we find that output growth
increases substantially after a negative shock, far
surpassing
208
growth,11 we will discuss these premises
and our results in light of several
theoretical interpretations.
(although the CBO projections, which
show projected GDP reaching potential
GDP by 2015, seem slightly more
optimistic) As we can see, our estimation
techniques to measure the persistence of
the shock provide results highly
consistent with the CBO’s projected
recovery path. The findings, in
accordance with the CBO projections,
suggest that the effects of the current
shock may keep output levels below
potential well into the upcoming decade.
Nelson and Plosser argue against
the first premise in favor of real business
cycles, where fluctuations in output are
attributable to changes and innovations
in aggregate supply, rather than
temporary
demand
shocks.
They
abandon the notion that fluctuations
are driven by aggregate demand—
particularly monetary disturbances—
suggesting
that
technological
innovations are the source changes in
economic growth patterns.
Conclusions
Our findings of persistence
suggest that a current shock to the
economy has longrun implications for
future economic growth. According to
Campbell and Mankiw, many traditional
business cycle theorists, monetarists, and
neo-Keynesians
maintain
two
fundamental premises: (1) fluctuations in
output are assumed to be driven
primarily by shocks to aggregate
demand, such as monetary or fiscal
policy, and (2) shocks to aggregate
demand are assumed to have only a
temporary effect on output, and the
economy returns to a natural growth rate
in the long run (Campbell and Mankiw
1987).
However, completely abandoning
our first premise need not be necessary.
For example, we can consider theories
that attribute fluctuations to both supply
and demand shocks. A Keynesian
position may suggest that demand
shocks affect supply, which, in turn,
permanently effects on the economy.
Similarly, it could be argued that demand
shocks, in it of themselves, have
permanent or near permanent effects.
Our second premise—shocks to
aggregate demand are assumed to have
only a temporary effect on output and
the economy returns to a natural growth
rate—can also be met be alternative
theoretical interpretations. For example,
models of multiple equilibria might
explain a long-lasting effect of aggregate
demand if shocks to aggregate demand
can move the economy between
However, if output shocks remain
persistent into the foreseeable horizon,
both of these premises cannot be strictly
maintained. Given the challenges
associated with distinguishing between a
trend stationary process and a process
that allows for permanent changes in
11
For example, even if the output process is
actually trend-stationary it may take quite a long
time to get
back to a trend.
209
equilibria (Diamond 1984). As well, the
hysteresis theory suggests that supply or
demand shocks may have permanent
effects because they affect the natural
rate of output, unemployment, inflation,
etc. Thus, perhaps our findings of
persistence suggest that there may be a
“new” natural rate after an economic
shock.
longstanding debate regarding the
existence of a unit root in the output
process, they do provide context to our
findings of persistence. While it remains
challenging, if not impossible, to prove
or disprove theories business cycles,
fluctuations,
and
output
growth
behavior, our ARMA and Monte Carlo
methods shift our efforts away from
championing a single exclusive theory
and towards realizing the impact and
persistence
of
economic
shocks.
Although
these
theoretical
interpretations fall short of settling the
Works Cited
Bachman, Daniel, Peter Jaquette, Kurt
Karl, and Pasquale Rocco. “The WEFA
U.S. macro model with chain-weighted
GDP.” Journal of Economic and Social
Measurement, 1998.
Macroeconomics.
Cambridge,
Massachusetts: MIT Press, 1984.
Enders, Walter. Applied Econometric
Times Series, Second Addition. Wiley
Series in Probability and Statistics. 2004.
Blanchard, Oliver, and Mark Watson.
“Are Business Cycles All Alike?” NBER
Working Paper Series. Working Paper
No. 1392, June 1984.
Evans, Martin. Where Are We Now? RealTime Estimates of the Macroeconomy.
International Journal of Central Banking,
September 2005.
Campbell, John and Gregory Mankiw.
“Are Output Fluctuations Transitory?”
The Quarterly Journal of Economics:
November 1987.
Jones, Charles. Sources of U.S. Economic
Growth in an World of Ideas. University
of California-Berkeley.
Christiano, Lawrence, and Martin
Eichenbaum. “Unit Roots in Real GNP:
Do We Know and Do We Care?” National
Bureau of Economic Research. October
1989.
Lucas, Robert. Understanding Business
Cycles. Carnegie-Rochester Conference
on Public Policy. 1977.
Nelson, Charles, and Charles Plosser.
“Trends and Random Walks in
Macroeconomic Time Series: Some
Evidence and Implications.” Journal of
Monetary Economics. September: 139-162.
Cochrane, John. “How Big is the Random
Walk in GNP?” The Journal of Political
Economy, Vol. 96, No. 5. October 1988.
Diamond, Peter. A Search-Equilibrium,
Approach to the Micro Foundations of
210
Nelson, Charles, and Heejoon Kang.
“Spurious Periodicity in Inappropriately
Detrended Time Series: Some Evidence
and Implications.” Journal of Monetary
Economics. 1982.
Papell, David, and Ruxandra Prodan. The
Uncertain Unit Root in U.S. Real GDP:
Evidence with Restricted and Unrestricted
Structural
Change.
University
of
Houston.
Romer, David. Advanced Macroeconomic.
McGraw Hill Companies Inc., New York.
35
211
A Generalized Leslie Matrix Model for Projections of Area
V Humpback Whale Population Demographics
Vildana Hajric, Class of 2011
Australian Humpback Whales have been harvested since the late 1700’s, helping to spur the
country’s first primary industry. Such continual hunting of the species diminished its stock
to near-extinction, forcing the International Whaling Commission to create a ban on
commercial whaling of humpbacks in the 1960’s. However, countries such as Japan and
Norway that depend on whaling as an integral part of their meat industries and cultural
traditions have commenced whaling under the guise of scientific research. This paper aims
to calculate humpback whale population growth rates of the eastern Australian area V
stock using survivorship and fecundity rates of different classes within the group.
Furthermore, it aims to show that scientific whaling has a negative effect on population
growth rates, helping to diminish the stock’s population numbers. Lastly, the paper gives
recommendations for future recovery plans that could potentially ensure that the area V
humpback population increases with time.
near extinction in the Pacific Ocean
along eastern Australia. Its numbers
decreased drastically between the 1780’s
and 1960’s causing concern among the
international community about its
preservation.
Introduction
In the early 1600’s, European
explorers and traders came in contact
with
the
previously
unexplored
Australian continent looking for new
land to claim under the name of their
home countries. The area was officially
claimed by Captain James Cook in 1770, a
British explorer who created early
settlements in the Sydney area
(McKernan, n.d.). By 1788, however, not
many areas had been developed for
settlement and colonizers were faced
with hunger due to lack of jobs and food
(“Australia’s,” 1995). Looking to build
industries and jobs, they started hunting
whales, using their meat and products as
trade items. Since then, whales have
been hunted for a number of different
reasons around the coasts of Australia.
The humpback whale, a baleen whale
often found off the coasts of New South
Wales, was targeted and hunted into
History of Whaling
Whaling off the coasts of Australia
commenced as a smaller industry used
primarily for food resources in the late
1700’s. By the 1830’s, whaling officially
grew into a commercial business when
Australia became a valuable region for
the export of raw materials, such as
whale products, to Britain and the
United States (“Australia’s,” 1995).
Whaling initially used smaller boats, but
methods were later aided with harpoons,
guns, and other weapons.
Whalers
sought after all types of whales, including
killer, minke, and humpback whales.
The animals were being harvested for a
number of different reasons. As Australia
offered little food sources elsewhere,
212
hunters mainly focused on whales
because they provided a reliable source
of food (“Killers,” 2009). The settlers
were in luck because whales were found
in large stocks off the coast of Australia.
Within a couple of years, a number of
different whaling businesses sprung up
along the coast near modern Sydney, and
traders around the world, especially
Britain and America, started visiting
Australia to profit from this industry.
The European colonizers, thus, turned
whaling into Australia’s first and oldest
primary
industry
(“Australia’s,” 1995).
Regulation of Whaling (ICRW) in 1946.
The ICRW then established the
International
Whaling
Commission
(IWC) whose aim was to protect and
conserve
rapidly-depleting
whale
populations. The IWC worked towards
protecting certain species of whales,
specifying sanctuaries for stocks, and
setting limits on how many whales could
be harvested per year. Through its
practices, the IWC has been able to help
recover whale species from near
extinction and plans on working towards
eliminating other threats
to whale populations in
the
future
(“IWC:
History,” 2008).
Yet
whaling
industries did not only
spring up in Australia—
Whaling as Part of
countries all around the
Scientific Research
world also harvested
Even though the
whales for their products.
IWC aims at protecting
The profession drew
and conserving whales, it
thousands of men to
does allow its member
work in the whaling
countries to apply for
business
as
whale
permits to hunt whales
products were coveted
for scientific purposes.
and
brought
many
These provisions are
Figure 1: Japanese boat whaling
financial
rewards
for scientific research (Modified
given
out
because
(McNamara,
2009).
from: “Samir’s Weblog,” 2008).
countries, such as Japan,
However, as the industry
that once heavily relied
grew over the years, the
on
whale
products
for uses in many of
number of whales declined and many
their industries found scientific whaling
countries became concerned with the
as a loophole to illegally engage in
rapid depletion of whale populations.
whaling
as they had done before
The League of Nations became
restrictions were imposed. Therefore,
particularly alarmed and drafted the
countries with permits are allowed to
Geneva Convention on the Regulation of
“kill, take and treat whales for scientific
Whaling treaty in 1931. The treaty was
purposes and requires that the animals
signed
by
twenty-two
countries,
be processed so far as practicable once
including Australia, and was revised a
the scientific data has been collected”
number of times, eventually leading to
(“IWC: Scientific,” 2008).
the establishment of the current
International
Convention
for
the
213
However, since the IWC started giving
out permits for this purpose, more than
11,000 whales have been taken by only
four countries: Japan, Iceland, Norway,
and Korea. This poses a problem for
whale conservation because a large
number of whales are killed in the name
of science, yet scientific whaling has
failed to conduct valuable research about
whales that could not have been done
through non-lethal research.
These
practices have been subject to heavy
diplomatic and international criticism,
yet progress to end scientific whaling has
been nothing but slow (Rothwell, 2007).
coasts of New Zealand and throughout
the South Pacific Islands” (Anderson et
al., 2004). One of the most shocking
discoveries, however, is that Japan sells
whale meat on its markets, all of which
comes from whaling done under the
guise of such scientific research and
purpose, as has been Japanese tradition
for hundreds of years (Head, 2005).
Nevertheless, “Japan’s increase in the
number of whales it kills has become an
abuse of a legal right to engage in
scientific whaling” and many members of
the IWC have proposed that an end
needs to come to JARPA programs
(Rothwell, 2007).
The country of Japan has been
particularly spotlighted as it has tried to
overturn the IWC’s efforts to stop
whaling worldwide. Under its research
programs known as Japan’s Research in
the Antarctic (JARPA) and JARPA II,
Japan has continued whaling humpbacks
that travel along the coasts of Australia.
Japan claims that not enough scientific
research has been done to fully
understand how to successfully manage
whale stocks (“ICRW: Background,”
n.d.). A majority of the whales Japanese
researchers harvest are ones that take
seasonal migrations along Australia’s
eastern coast, known as area V
humpbacks. All humpback whales
traveling in the Southern Hemisphere
are divided into six different stocks based
on migration patterns and aggregations
in the Antarctic during their feeding
season, and the whale stock most
commonly observed along the eastern
coast of Australia is the area V
humpback. It is contained within 130ºE 170ºW and is the same stock that is
“observed migrating along the east coast
of Australia, past Norfolk Island, both
Question
With current trends of whaling,
the area V humpback whale population
is being affected. This paper aims to
discuss whether or not the stock
population is decreasing, what the effects
of whaling under the name of scientific
research are, and how the population
would be affected if scientific research
whaling did not exist.
Humpback Whales
Humpback Whaling
When
Australia’s
whaling
industry was in its prime, humpbacks
were targeted, specifically because
hunters could use their oil for soaps,
lubricants, perfumes, candles, etc. and
their baleen for the creation of umbrellas
and corsets (Egan, 1995). However, their
harvest was unregulated until recently,
bringing them to a point of near
extinction. Nearly 95% of humpback
whales were eliminated as Australia and
New Zealand harvested more than
214
40,000
humpbacks
has special sieves in its
migrating from the
mouth known as baleen
Antarctic
along
plates.
In Australia,
Australia’s eastern coast
they are one of the
in the 1800’s and 1900’s.
most commonly sighted
The area V stock
whales on the eastern
contained
around
and western coasts of
Figure 2: Mother and calf on southern
10,000 whales in its unthe country (“Whale
migration (Modified from: “Conforth
harvested state, but
Watching,” n.d.). They
Images,” 2009)
only 500 individuals
are stocky, stout, and
survived
by
the
year
1962
extremely large with outsized pectoral
(Chittleborough, 1963, p. 34). Whaling of
fins, which it uses for breaching and
humpbacks officially ceased in 1963 in
slapping water in acrobatic ways (Perrin,
Australia, and ended worldwide in 1965
2008, p. 584). It receives its name
after the international community
“humpback” because it arches its back
recognized a dramatic decline in
out of water in preparation for a dive for
population numbers.
This decline
air (Harrison, 2005). The top of its back
incidentally led to the demise of
is dark blue or black and its underside
Australia’s
whaling
industry
and
can wary anywhere from black to white
commercial whaling in Australia ceased
to marbled patterns, characteristics
in 1978 with the closure of Australia’s last
scientists use in demographic studies of
whaling station in Western Australia
the whales. Each individual has his or
(“Department,” 2007).
Currently,
her own particular pattern and shading,
Australia has listed its humpback whale
with variations being noticed among
populations as vulnerable under the
whales originating from different parts of
Environmental
Protection
and
the world (Paterson, 2000).
More
Biodiversity Conservation (EPBC) Act
interestingly,
however,
is
that
(“Australian Government,” 2005, p. 5).
humpbacks grow to about fifty-two feet
long and can weigh thirty to fifty tons,
Life History
making them a very slow moving animal
(“Australian
The
Antarctic,” 2009).
humpback whale
(Megaptera
novaeangliae)
is
found all over the
world including in
the North Pacific,
North Atlantic, and
Southern
Hemisphere
(Perrin, 2008, p.
582). It is a baleen
whale, meaning it
Figure 3: Distribution of humpback whales
around Australia (Modified from: “Department
of Environment,” 2009)
215
All
humpback whales
take
long
and
arduous seasonal
migrations—the
longest of any
mammal.
They
mate in warmer
waters
during
winter seasons and
travel to colder
waters during the spring, summer and
fall seasons to feed (Perrin, 2008, p. 583).
Area V humpback whales breed and
calve on the northern side of the east
coast in warmer areas such as the Great
Barrier Reef and the Northwest Shelf
during the winter (“Australian Antarctic,”
2009).
They then embark on their
migrations, traveling in large, loose packs
with their calves from northern Australia
towards colder, polar waters during the
summers in search of food (Perrin, 2008,
p. 583). Adult whales do not feed during
the winter, managing to survive off their
layers of fat during that time, and feed
enough during summer, fall, and spring
seasons to produce milk for their future,
young calves (“Animal Info,” 2005).
Once in the polar waters, they eat twice a
day and tend to feed on planktonic
crustaceans, including krill, small fish,
and plankton (Perrin, 2008, p. 583).
Female humpbacks give birth
every one to three years. Twins are very
rare and over 99% of the time only one
calf is born to a female. The gestation
period of humpbacks is around eleven to
twelve months. Calves are born about
fourteen feet long and weigh around two
tons. Babies are nurtured with their
mothers’ milk and are weaned sometime
between a year and 18 months and reach
puberty between ages three and six
(Paterson, 2000). Once they are born,
the mother and calf pair travels to the
colder polar waters together in
conjunction with other, looser groups of
whales. The only strong bonds among
groups are found between mother and
baby pairs. Because of their large size,
they are only able to travel three to nine
miles an hour and have even slower
speeds during the babies’ feeding times
(“Australian Antarctic,” 2009). Due to
their slow speeds and extremely large,
noticeable bodies, humpback whales
were prime targets for early whaling
industries (“Australian Antarctic,” 2009).
Demographic Model
A demographic model is used to
calculate and describe population change
over time as it is affected by certain
factors. For this study, a Leslie Matrix
model is implemented to determine
population growth rates for Area V
humpback whales using survivorship
abilities and fecundity rates for specific
age classes.
Leslie Matrix Model and Methods
Figure 4: Area V Humpback whale
migration calendar (Modified from:
Paterson, 2000).
A Leslie Matrix model is a basic
tool used for population studies where
specie extinction or survival rates are
216
model of population growth rates by
calculating how fecundity rates and
survivorship abilities affect the number
of individuals in each class, thus
projecting growth or decline within the
overall population.
calculated using parameters of age or
class-specific survivorship abilities and
age or class-specific fecundity rates
(Liebler, 2003, p. 163). It can be used to
determine population growth rates as
well as age distribution within
populations over time. Individuals of
studied populations are grouped by
maturity and each group records the
number of individuals for every age
bracket (Liebler, 2003, p. 163). For this
study, Area V humpback whales have
been condensed into four groups,
according to their maturity levels:
infants, sub-adults, adults, and postadults. Infants (class 1) were defined as
age groups 0-2 because once whales are
born the yearlings stay with their
mothers for 12 to 18 months
(Chittleborough, 1957, p. 2). Sub-adults
(class 2) were defined as age groups 3-6
because these are weaned but not yet
sexually mature whales (Chittleborough,
1963, p. 54). Adults (class 3) were defined
as age groups 7-20, as this is the age
period when humpback whales are
sexually mature. The post-mature group
(class 4) was defined as age groups 21-45
and whales in this class do not reproduce
(Chittleborough, 1963, p. 55).
Basic Assumptions
Because only mature adults are
capable of reproduction, only class 3 (F3)
can be used to calculate fecundity rates
for the different classes. That means that
F1 = 0, F2 = 0, and F4 = 0. In order to
calculate F3, however, the average
number of calves female humpback
whales have during their adult lifetime is
used, which is 6-9 calves—6 being the
lowest end of reproduction and 9 being
the highest. This was calculated based
on the fact that humpbacks are sexually
mature between ages 7-20 and have, on
average, one calf every two years with a
one-year resting period every two cycles
(Chittleborough,
1963,
p.
55).
Furthermore, 37% (0.37) of females give
birth on a two-year cycle (one calf in two
years) (Chittleborough, 1963, p. 58, 1957,
p. 1). If it is assumed that the most
conservative number for reproduction is
6, then the average number of female
offspring produced by a female whale
during this timeframe for class 3 (F3) is (6
* 0.5 * 0.37) = 1.11. If it is assumed that
the high end for reproduction is 9 calves,
then F3 = (9 * 0.5 * 0.37) = 1.67.
Therefore, F3 equals 1.11 when the average
number of calves used is 6 to calculate
fecundity rates, and equals 1.67 when the
average number used is 9 calves.
Each
class
or
group
is
characterized by fecundity (F) or the
average number of female offspring
produced by a female. Class 1 is denoted
by F1, class 2 by F2, class 3 by F3, and class
4 by F4. Groups are also characterized by
probability of survivorship through each
age class (P) which is the survival rate of
each age group from infant to sub-adult
to adult to post-adult. Class 1 is denoted
by P1, class 2 by P2, class 3 by P3, and class
4 by P4. From these assumptions it is
possible to construct a deterministic
To
estimate
survivorship
probabilities, approximations from a
previous study were used which
217
calculated that P2 = 0.7103, and P3 =
0.9100 (Jackson, 2009). Because the
study did not have approximations for P1,
the survivorship rate of infants,
information was used from the
Chittleborough 1965 paper to determine
P1. The study states that the average
number of births per 100 females was 18.6
in three years of sampling. The average
number of calves surviving to age 3 (subadult) was 11.9, 13.4, and 15.1.
By
averaging those numbers together and
dividing them by 18.6, it was calculated
that P1 = 0.7240. Therefore P1 = 0.73, P2 =
0.71, and P3 = 0.91 (Chittleborough, 1963,
p. 115).
Basic Results
Using the calculated numbers for
survivorship (6 calves per year, F3 = 1.11),
fecundity (P1 = 0.73, P2 = 0.71 and P3 =
0.91), and class sizes (infants=717, subadults = 520, adults = 2710) it is simple to
determine the population growth rates
given these parameters and the Leslie
Matrix model using the program
Populus. The results show that the
population growth, characterized by λ, is
0.83, or λ = 0.83. Because the resulting
number is less than one, the population
is shrinking under these circumstances.
When projecting population change for
the next couple of decades using this
result, it is clear that humpback whales
would become extinct by 2050 under
these assumptions.
However, if the
population growth rates using 9 calves
per year, or F3 = 1.67, were to be
calculated, then λ = 0.95. The resulting
number is still less than one. Therefore,
when using these calculations to project
population change in the next couple of
decades, the area V humpback whale
population would still be shrinking
under such assumptions and whales are
likely to be extinct by 2150.
Lastly, in order to calculate
population growth rates, fecundity and
survivorship rates need to be evaluated
in conjunction with the population sizes
of the four different classes. To estimate
current population sizes in area V
humpback whales, the Chaloupka et al.
1999 study was used that states that the
stock has 156 sub-adults and 813 adults
(Chaloupka, 1999). The study also states
that these numbers represent between
30-50% of the entire population. To
calculate the number of calves for the
same time period, it was assumed that
approximately 72.4% of calves survive to
sub-adults.
Therefore, the estimated
number of calves is 215. If it is assumed
that these class size numbers represent
30% of the entire population, then there
are 717 calves, 520 sub-adults, and 2710
adults at the first time point of the
modeling (Chaloupka, 1999). Finally, the
number of post-adults is presumed to
start at 0 at the first time point in the
modeling because it does not influence
the population growth rate.
218
0 1.11 0
 0
0.73 0
0
0

 0
0.71 0
0


0
0.91 
 0
them for a specific year.
Japanese
researchers claim to be taking fifty
whales per year from the region or
around 0.1% of the total population of
humpback whales (“Embassy of Japan,”
2009). They claim that these numbers
have no effect on the population
numbers. However, fifty whales amount
to 1% of the estimated area V population,
ten times the number the Japanese claim
to harvest. These practices need to be
calculated into population growth rates
in order to get a more accurate estimate
of population change over the years.
0 1.67 0
 0
0.73 0
0
0

 0
0.71 0
0


0
0.91 
 0
Leslie Matrix Equations depicting
Calculations using F3 = 1.11 and F3 = 1.67
As stated before, the P values
indicate the probability rate that a
female survives to the next class. It is
known that P1 = 0.73 and that whales
spend two years in that class before
moving to the next one. Therefore, their
yearly survivorship equals (0.73)1/2 or
0.85. P2 = 0.71 and whales spend four
years in that class before moving to the
next one. That means that their yearly
survivorship equals (0.71)1/4 = 0.92. So, if
it is assumed that whaling takes 1% of the
population each year (50 whales / approx
5000 total), then putting those numbers
back into the survivorship calculations
makes P1 = 0.86 * 0.86 instead of
0.85*0.85, and P2 = 0.934 instead of 0.924.
The new survivorship rates therefore are
P1 = 0.74, P2 = 0.72 and P3 = 1.0. Using
these numbers it is easy to calculate the
effect no whaling would have on the
humpback population as was done
previously. The calculations show that
with 6 calves per year λ = 0.84, and with
9 calves per year λ = 0.96. Because both
these numbers are less than 1, the
population would be projected to be
shrinking in either case.
An explanation as to why
humpback whale numbers are decreasing
can be found by calculating whether or
not scientific whaling impacts stock
numbers, as area V humpback whales are
still being harvested under the guise of
scientific research. In order to calculate
true population growth rates, current
levels of scientific whaling must be
included to investigate the effect of
greater or lesser whaling which can be
done by manipulating the survivorship
probability values of the stock.
Effects of Whaling on Calculated Results
Whaling of area V humpback
whales happens in the Antarctic once
they migrate to the region during their
mating seasons. It is mainly done by
Japanese hunters who claim to harvest
whales in order to conduct scientific
experiments that could later be used to
create better protection programs for
whales. However, their whaling greatly
impacts
the
population
of
the
humpbacks as they have been known to
take more whales than the IWC allots
219
However, it is a possibility that
the whaling numbers provided by the
Japanese are not accurate. Therefore, a
more representative number of whales
7000
6000
5000
4000
3000
2000
1000
2141
2130
2119
2108
2097
2086
2075
2064
2053
2042
2031
2020
2009
0
1.0. Calculating population change with
6 calves per year produces λ = 1.0034.
This number is very close to 1.0, which
means that the population stays stable,
not growing nor shrinking.
When calculating it with 9
calves per year λ = 1.15. This
number gives a 15% population
Whaling (6)
increase per year, meaning that
Whaling (9)
without
the scientific whaling
Assumed (6)
done by Japanese researchers,
Assumed (9)
Reality (6)
humpback populations are
more likely to be stable or
growing.
Figure 4: Graph depicts humpback population predictions
according to demographic calculations affected by Japanese
whaling numbers using first 6 then 9 calves, calculations
with assumed Japanese whaling numbers using first 6 then 9
calves, and finally demographics without whaling using 6
calves (calculations using 9 calves have been excluded due
to extraordinary quick growth rates)
Conclusion and
Recommendations
Under
the
current
conditions,
it
has
been
calculated
that
area
V
humpback whales are facing
extinction, perhaps as soon as
2050 or 2150. In reality, because
of the way the model is calculated with
age groups, extinction rates are probably
closer to 2070 or 2170 and have a 20-year
lag within age groups. If the current
amount of whaling was stopped at levels
cited
by
Japanese
researchers,
populations would continue to shrink.
This is very strange because without
external mortality through whaling one
would think the population would be at
carrying capacity, meaning that the
population growth rate or λ would be
close to 1.0. Under perhaps a more
realistic assumption of whaling (10%
instead of 1%), calculations show that if
whaling were to stop, λ would indeed be
1.0034 (under an average of 6 calves per
female), which would indicate a
population near carrying capacity.
Under the assumption of 9 calves per
harvested or killed each year using more
realistic factors can be calculated. It is
know that the Japanese take a total of
930 whales in the Antarctic including
minke and blue whales (“Act Now,”
2008).
Out of that number, it is
estimated that blue whales are taken
least. It can, therefore, conservatively be
estimated that 20% of the harvest is blue
whales and that the rest of it is split
between minke and humpback whales.
This means that minke and humpback
whales are harvested at 40% each. For
humpback whales, that means that 40%
of harvest * 930 whales = 372 humpbacks.
Those 372 whales equal nearly 10% of the
area V humpback population, not 0.1% as
the Japanese claim.
Under these
assumptions, if whaling stopped, then P1
= 0.95*0.95 =0.91, P2 = 1.0 (becomes 1.0
with just a 7% average return), and P3 =
220
female and no whaling at the 10% level, λ
= 1.15. However, an average of 6 calves
per female is probably more accurate
than 9 calves per female, and perhaps the
10% estimate is closer to reality than the
1% calculations, especially because other
studies have found similar population
growth rates to be true (Noad et al.,
2007).
scientific research methods that do not
interfere with whale migratory paths,
feeding and breading grounds in order to
learn about the humpbacks and their
lifestyles. To accomplish this, scientists
must locate specific migratory routes of
area V humpback whales in nondisturbing ways, which can be done
through careful observation using
technological advances and satellite
imagery. Such routes could then be
sanctioned by the government and put
under satellite watch to ensure complete
specie protection. Sanctions would help
ensure that no invaders come near their
migratory paths without permission or
valid
excuses
validated
by
the
government. Australia has already taken
measures to ensure that humpback
whales are protected in established
sanctuaries and within parameters of the
Exclusive Economic Zone (EEZ), an area
over which a country has special rights,
but needs to ensure that such areas are
completely guarded at all times
(“Australian Government,” 2005, p. 5).
Furthermore, the Australian government
can impose policy limiting the activities
of whale watching industries and the
impacts they make on their tours.
Whales should be left undisturbed and
should be observed from a distance,
allowing tourists to see them while not
interfering with their lives.
Finally,
similar restrictions should be placed on
fishing boats and other fleets traveling
near migratory or sanctuary humpback
areas. Measures should be taken to
ensure that no whales get caught in
fishing nets, none are hit accidentally by
boats, and pollution to their waters is
minimized by reducing waste that is
discarded by all ships. Fines should be
It is clear that whaling of
humpbacks poses an apparent and
eminent danger to the animals but
numerous countries that depend on
whaling as part of their main industries
have been pushing for reform within the
IWC to allow commercial whaling to
legally
recommence
(“Australian
Government,” 2005, 6).
Humpback
whales are also threatened by many
other anthropogenic activities and
factors, including climate change,
pollution, whale watching, illegal
whaling, fish nets, and collisions with
boats (“NPCA,” n.d.). In addition, it is
very possible that the specie’s habitats
will be destroyed by climate change
factors. In the future, “humpback whale
migration, feeding, resting, and calving
site selection may be influenced by
factors such as ocean currents and water
temperature” which are influenced by
climate
change
(“Australian
Government,” 2005, p. 7). This could
impact humpback whale populations in
many negative ways, contributing to
their shrinking numbers.
Through careful monitoring of
populations and their habitats, the
Australian government can help ensure
that humpback whales do not become
extinct in the near future. This can be
done by implementing programs and
221
put in place for all who disturb the
regulations.
Japanese to stop their whaling and
urging the international community to
do the same, as non-lethal whale
research can provide all needed
information about whales.
The
Australian Minister for the Environment,
Peter Garrett has proclaimed that “nonlethal whale research can provide all the
information needed to understand and
conserve whales and so called ‘scientific
whaling’ is unnecessary” (Garret et al.,
2009).
Therefore, scientists need to
make it clear that humpback whales are a
threatened and endangered specie, that
survival of even 50 humpbacks can make
a difference on stock populations, and
that scientific whaling needs to come to
an
end.
Once such measures are taken,
humpback whales would be protected to
a certain degree.
However, the
Australian government needs to focus on
eliminating the scientific whaling done
by Japan and place a ban on their
practices.
Even though restrictions
already exist on the amount of
humpbacks that can be harvested per
year by the Japanese, Australia and other
countries need to work together towards
lowering such rates and monitoring how
many whales are truly harvested by
Japanese ships until a total ban is
established on whaling. This can be
accomplished by putting pressure on the
http://samiryadav.wordpress.com/2008/
02/
Works Cited
(2009). Killers in Eden.
Retrieved
October
10,
2009,
from
http://www.pbs.org/wnet/nature
/episodes/killers-in-eden/australiaswhaling-industry/1051/
(n.d.). Japan’s Whale Research in the
Antarctic—Backgrounder.
Retrieved
October
20,
2009,
from
http://www.icrwhale.org/eng/backgroun
d.pdf
Act Now. (2008). Whaling. Retrieved
October
27,
2009,
from
http://www.actnow.com.au/Issues/Whali
ng.aspx
(n.d.). Whale Watching Australia.
Retrieved October 8, 2009, from
http://www.totaltravel.com.au/au/travell
er/whale-watching-australia
(2005). Animal Info—Humpback Whale.
Retrieved October 8, 2009, from
http://www.animalinfo.org/species/cetac
ean/meganova.htm
Anderson, M., Baverstock, P., & Slade, R.
(2004). Historic Stock Structure and
Genetic Variation in Area V Humpback
Whales: Impacts on Modern Stock
Recovery. Retrieved October 20, 2009,
from
http://www.environment.gov.au/coasts/
publications/ballina-workshop
2004/pubs/anderson.pdf
(2008). Samir’s Weblog. Retrieved
November
10,
2009,
from
Australian Antarctic Division. (2009).
Humpback Whale. Retrieved October 8,
(1995). Australia’s Whaling History.
Retrieved October 8, 2009, from
http://www.whales.org.au/history/aust.h
tml
222
from
http://www.publish.csiro.au.ezproxy.uow
.edu.au/?act=view_file&file_id=MF95800
01.pdf
2009,
from
http://www.aad.gov.au/default.asp?casid
=2451
Australian delegation to the 60th annual
meeting of the International Whaling
Commission.
(2008).
International
Whaling Commission: History and
Future. Retrieved October 19, 2009, from
http://www.environment.gov.au/coasts
/publications/pubs/iwc-factsheethistory.pdf
Chittleborough, R.G. (1963, November).
Dynamics of Two Populations of the
Humpback
Whale,
Megaptera
Novaeangliae (Borowski).
Australian
Journal of Marine and Freshwater
Research,
16,
33-128,
from
http://www.publish.csiro.au.ezproxy.uow
.edu.au/?act=view_file&file_id=MF96500
33.pdf
Australian delegation to the 60th annual
meeting of the International Whaling
Commission. (2008).
International
Whaling Commission: Scientific Whaling.
Retrieved
October
19,
2009,
http://www.environment.gov.au
/coasts/publications/pubs/iwc-factsheetwhaling.pdf
Conforth Images. (2009). Retrieved
November
12,
2009,
from
http://www.cornforthimages.com/Galleri
es/Wildlife/Whales/Humpback.htm
Department of the Environment, Water,
Heritage, and the Arts. (2007, August).
Retrieved October 8, 2009, from
http://www.environment.gov.au/coasts/s
pecies/cetaceans/history.html
Australian Government: Department of
the Environment and Heritage. (2005).
Humpback Whale Recovery Plan 20052010. Retrieved October 27, 2009, from
http://www.environment.gov.au/biodive
rsity/threatened/publications/recovery/
m-novaeangliae/pubs/mnovaeangliae.pdf
Department of the Environment, Water,
Heritage and the Arts (2009). Megaptera
novaeangliae in Species Profile and
Threats Database, Department of the
Environment, Water, Heritage and the
Arts,
from:
http://www.environment.gov.au/sprat
Chaloupka, M., Osmond, M., & Kaufman,
G. (1999, July). Estimating seasonal
abundance
trends
and
survival
probabilities of humpback whales in
Hervey Bay (east coast Australia).
Marine Ecology Progress Series, 184, p.
291-301,
from
http://www.intres.com/articles/meps/184/m184p291.pdf
Egan, Max. (1995). Australian Whaling
History. Retrieved October 9, 2009, from
http://www.bigvolcano.com.au/human/
whaling.htm
Embassy of Japan. (2009). FAQ on
Whaling—Answering Questions we have
Received in Recent Times. Retrieved
October
27,
2009,
from
http://www.au.emb-japan.go.jp/e_web
/files/whaling.html
Chittleborough, R.G.
(1957, August).
The Breeding Cycle of the Female
Humpback Whale, Megaptera Nodosa
(Bonnaterre).
Australian Journal of
Marine and Freshwater Research, 9, 2-24,
223
Garrett, P., McCully, M., & Smith, S.
(2009, June). Australia and New Zealand
Launch the World’s Largest Non-lethal
Whale Research Partnership. Retrieved
November
9,
2009,
from
http://www.environment.gov.au/ministe
r/garrett/2009/mr20090618.html
October
8,
2009,
from
http://www.dfat.gov.au/aib/history.html
McNamara, Robert. (2009). A Brief
History of Whaling. Retrieved October 8,
2009,
from
http://history1800s.about.com/od/whalin
g/a/histwhaling.htm
Harrison, Molly. (2000). The Kid’s Times:
Humpback Whale. Retrieved October 27,
2009,
from
http://oceanservice.noaa.gov/education/
yos/resource/15E_kids_times_whale_hu
mpback.pdf
Paterson, Robert. (2000). Humpback
Whales. Retrieved October 8, 2009, from
http://www.qm.qld.gov.au/inquiry/facts
heets/leaflet0011.pdf
Perrin, William F. and Wursing, Bernd.
(2008).
Humpback
Whale.
In
Encyclopedia of Marine Mammals (p. 582585). Burlington: Academic Press.
Jackson, Jennifer. (2009). Southern
Hemisphere
humpback
whale
Comprehensive Assessment workshop,
Retrieved October 25, 2009, from
http://www.iwcoffice.org/_documents/sc
i_com/workshops/SC-F09-SH7.pdf
Rothwell, Donald R. (2007, July). Time
to End Loophole Scientific Whaling.
Retrieved October 20, 2009, from
http://law.anu.edu.au/cipl/Expert%20Op
inion/2007/Rothwell%20%2031%20July.pdf
Liebler, Robert A. (2003). Basic Matrix
Algebra
with
Algorithms
and
Applications. Florida: CRC Press.
Noad, M., et al. (2007). An Update of the
East Australian Humpback Whale
Population (E1) rate of increase.
Retrieved October 20, 2009, from
http://www.iwcoffice.org/_documents
/sci_com/workshops/MSYR/SC-60SH31.pdf
National Parks Conservation Association.
(n.d.). Humpback Whale (Megaptera
novaeangliae), Retrieved November 9,
2009,
from
http://www.npca.org/marine_and_
coastal/marine_wildlife/humpbackwhale
.html
McKernan, Michael. (n.d.). Ancient
Heritage, Modern Society.
Retrieved
224
Validating Foreign Language Issues: What are the Effects
of Non-English or Bilingual Population on an MSA?
Mary Liz Brady, Class of 2011
Increasing immigration, as well as globalization of the national economy, has put foreign
language issues in the popular media spotlight. ProEnglish advocates support establishing
English as the national language and diminishing bilingual education in favor of English
immersion programs. On the other hand, several school systems are emphasizing the
importance of foreign language classes in light of the increasing economic interdependence
and importance of foreign diplomacy. This paper examines the effects of a non-English
speaking population on the well-being of a metropolitan statistical area. Furthermore, it
explores what effect a bilingual population has on the MSA. Results indicate that both a
non-English and bilingual population are insignificant in determining average income
across a city.
Introduction
assimilate with the U.S. labor market.
The ProEnglish organization, one of the
major interest groups fighting to
establish English as the official national
language, argues that multilingualism is
creating a growing lower class, walled off
from the larger economy and segregated
into
enclaves
(ProEnglish
2010).
Advocates of this movement call for
decreased bilingual education, and
instead call for a replacement by an
English immersion curriculum, such as
the common English as a Second
Language or ESL program. Most U.S.
states have taken action to establish
English as their official language (See
Figure 1). Declaring English as the official
language
means
“that
official
government business at all levels must be
conducted solely in English. This
includes all public documents, records,
legislation and regulations, as well as
hearings, official ceremonies and public
meetings” (U.S. English Inc.).
This paper examines the effects of
an MSA’s collective English proficiency
and bilingual ability on average earnings
within the area. This question is
particularly prevalent as immigration
grows and international relationships
gain importance. In data collected by the
2000 U.S. Census, 47 million Americans,
or about 18%, reported speaking a
language other than English inside the
home. This figure has increased from
14% in the 1990 Census and further from
11% in the 1980 Census (Vistawide 2010).
Combined
with
the
continuing
globalization of the national economy,
the growing presence of foreign language
within U.S. borders has inspired political
and social issues that have captured
public attention. Popular media has
underscored two major, and in many
ways ideologically opposite, concerns
surrounding foreign language.
First, pro-English advocates assert
the importance of immigrants learning
proficient English in order to aptly
225
Figure 1
Research has demonstrated that
immersion into the American labor force
weakens mother tongue ability from first
to second generation immigrants (Fry
2003).
Certainly,
the
immersion
programs proposed by pro-English
supporters would exacerbate this issue.
The weakening of foreign language
ability among American immigrants falls
into direct conflict with the second major
issue emphasized by popular media: the
recognition of foreign language classes as
an essential piece of the American
education. In the face of the budget cut
backs, believers in foreign language
education point to the increasing
globalization of the American economy
as a key reason to maintain these
programs. Moreover, at a time when the
United States is involved in two wars and
other unstable foreign relationships, they
argue that intelligence gathering,
military planning, and diplomacy are as
essential as ever (Holman 2010). Thus, it
appears that while immigrants and
second generation Americans are often
encouraged to weaken ties with nonEnglish
languages,
the
American
education system is encouraging and
requiring students to pursue bi- or
multilingualism.
It is uncertain whether either of
these assertions is economically sound.
It seems logical to support the
assimilation of new immigrants into the
American labor market through learning
English. Empirical studies provide mixed
conclusions on the significance of
English proficiency on the earnings of
immigrants. However, in the study
conducted by Park (1999) he makes a
convincing argument suggesting a
positive correlation between English
fluency and earnings. Rather than
examine a direct correlation, Park
investigates the effect of English fluency
on the parameters of other variables in
the earning equation, (Park 1999). In
particular, he observes the effects of
English proficiency on immigrants’
226
return to education and experience
obtained before and after immigration.
racism are responsible for curbing
economic growth in these instances.
The results are largely in line with
Park’s hypothesis. Experience and
education received outside of the United
States are more valuable to English
speakers who are able to transfer the
benefits easily in to the American labor
market.
Furthermore,
non-English
speakers received a greater return to
experience and education obtained
within the United States as it is a
comparably more significant acquisition
of human capital. The non-English group
is divided into those who speak English
very well denoted “fluent” and all others
denoted “nonfluent.” In accordance with
Park’s hypothesis, the fluent group
receives greater returns to education and
experience abroad, while the non-fluent
group receives greater returns from
education and experience obtained
domestically.
However, other empirical studies
provide more optimistic results. In their
2005 investigation “Cities and Cultures,”
Ottaviano and Peri argue that linguistic
diversity is associated with higher wages
for natives in a city. This effect is
strongest for high skilled workers, but
remains present for the unskilled, as
well. In addition, they find a positive
association between employment and
linguistic diversity within a city.
Including several demographic controls
such as schooling, experience, share of
women, and share of blacks, they find
that moving from a linguistically
homogenous city like Pittsburgh, to a
more linguistically diverse city like Los
Angeles is associated with a 13% increase
in the average wages of U.S. born
workers. They contend that their results
indicate that with linguistic diversity
come distinct skills to production
(Ottaviano 2005). They address the
crucial issue of endogeneity by
constructing
instruments
involving
distance to borders and gateway cities, as
well as using a shift-share methodology.
These estimations support the proposed
causation.
Settling on the conclusion that
Americans are better off with increased
English proficiency, the question remains
whether this is a political or individual
matter. In other words, it is worth
exploring
whether
the
individual
immigrant’s decision to speak his or her
mother tongue will have a negative
impact on an MSA as a whole. Again,
empirical analysis points to mixed
conclusions about the benefits of a
homogenous civilization. Easterly and
Levine (1997) demonstrate a negative
relationship between per capital GDP
growth
and
ethnolinguistic
fractionalization across a large sample of
countries (Easterly 1997). It is possible
that preferential conflict, communication
obstacles, or outright prejudice and
Ratna et al. (2009) build on
Ottaviano and Peri’s research. While the
main argument in their study is the
negative relationship between racial
diversity and Gross State Product, their
regressions consistently produce positive
correlation between GSP and linguistic
diversity. Oddly, this study also finds a
positive relationship between the
percentage of the population with
insufficient English fluency and GSP but
227
they assume this is a simple endogeneity
problem resulting from more productive
states like California and Florida having a
large number of Spanish speakers and
non-native workers (Ratna et al 2009). In
their discussion, Ratna et al. provide
arguments against their counterhypothesis language result. They insist
that while enclaves such as “Chinatown”
or “Little Italy” provide a springboard for
newly arrived immigrants, the continual
use of mother tongue as a business
language is a huge barrier to economics
success (Ratna et al 2009).
may need to speak another language.
Finally, they argue that research finds
bilinguals tend to perform better
academically, and thus are probably
more productive professionally (Fry
2003).
In analyzing their data collected
from the National Adult Literacy Survey,
Fry and Lowell find that workers who
know a second language cluster at two
extremes of the education distribution.
Bilinguals are both more likely than
monolinguals to have not completed
high school and to have completed at
least a bachelor’s degree. Their initial
regression suggests that workers who
know a second language in addition to
English earn about 7% more than similar
workers who only know English.
However, once educational credentials
are controlled for, there are no
statistically significant wage gains from
knowledge of a second language. The
results are not altered when specific
language or occupational categories are
considered.
While both studies support
arguments for linguistic diversity, neither
refutes the importance of English
proficiency. In fact, Ottaviano and Peri
conclude that “benefits from immigrants
who have integrated (i.e. have been in
the US for a longer period of time and
speak English well) are larger than those
from new immigrants” (Ottaviano 2005).
Thus it is fair to conclude that the
cultural benefits captured in the
linguistic diversity studies may be
attributable to bilingual individuals who
maintain distinct cultural and linguistic
characteristics
still
overcoming
communication obstacles with proficient
English ability.
In summary, popular media topics
about foreign language remain unsettled.
While it is true that it is individually
beneficial for an immigrant to be English
proficient, it is not clear what indirect
effects English-only assimilation may
have on a particular labor market.
Furthermore, while one study refutes
popular opinion about the advantages to
being bilingual, this matter can certainly
be re-approached. Accordingly, I will
focus my research on a two-fold topic:
It
is
then
reasonable
to
hypothesize that bilingualism has a
significant value in the U.S. labor market.
Fry and Lowell (2003) explore this
avenue and reach unexpected results.
They rationalize that in an increasingly
global
economy,
multinational
corporations need workers with ability to
speak both English and another
language. Furthermore, workers in
supervisory service-provider positions
• Do non-fluent English speakers
lower average incomes within an
MSA?
228
are 567 total observations taken from 284
metropolitan areas in the 2000 US
Census and the 2008 American
Community Survey. In order to measure
an MSA’s general economic quality, I use
the average wage and salary income for
all residents over age 18 who have an
income greater than 0. While not totally
comprehensive, the income variable
captures the relative well-being of a
region for this analysis. In order to
measure elasticity and diminish the
effects of outliers, we use the natural
logarithm of wage and salary income as
the dependent variable in all regressions.
• How does the percentage of
bilingual residents of a city affect
average earnings?
How does
bilingualism impact cities with
relatively low English proficiency?
Data
In order to capture the effects of
non-English and bilingual speakers, this
paper will measure the economic
consequences they impose on a
Metropolitan Statistical Area.
I use
MSAs rather than cities because they
consider social and economic integration
as well as geographic boundaries. There
Table 1: General Summary Statistics
Variable
Wage & Salary Income
Years of Education
English-speaking Population
Bilingual Population
Population
Hispanic Population
Black Population
Foreign-born Population
N
567
567
567
567
567
567
567
567
Mean
$31,790
13.24
97.16%
10.71%
770,250
9.74%
9.26%
9.99%
To measure the English-speaking
ability of a given MSA, we use the
variable “ENG” that takes a value of 1
when the respondent claimed to speak
only English, speak English very well, or
speak English well. “ENG” takes a value
of 0 otherwise. When the variable is
averaged across a metropolitan area, it
ranges from 0 to 1 and represents the
population of proficient English speakers
in the given MSA. Table 1 indicates
“ENG” ranges from 76.81% to 100%.
Std. Dev.
$6,025
.52
0.04%
10.54%
1,588,433
14.37%
9.41%
9.12%
Min.
$19,741
10.87
76.81%
0.17%
96.493
0.27%
0.11%
0.69%
Max.
$89,160
14.68
100.00%
73.93%
17,500,000
95.43%
46.73%
61.97%
Another
major
variable
is
bilingual
ability
“BILING”
in
a
metropolitan area. It takes a value of 1
when the respondent spoke both
proficient English and a second language
inside the home. Like the “ENG” variable,
it is averaged across a metropolitan area
and represents the MSA’s population of
bilingual
speakers.
Although
it
successfully captures the foreign born
population that also speaks English, its
major shortfall is the failure to consider
those who speak English as their first
language but are proficient in another
229
language, as well. Although this skews
the data, this study focuses mainly on the
impacts of immigrants on a city.
Accordingly, this paper will consider the
group irrelevant.
Table 2: Summary Statistics for Significant MSAs (2008)
MSAs with highest average incomes
MSAs with lowest average incomes
Metropolitan
Area
Average
Income
Englishspeaking
Population
Bilingual
Population
Metropolitan
Area
Average
Income
Englishspeaking
Population
Bilingual
Population
Stamford, CT
$89,161
95.12%
21.43%
Laredo, TX
$22,663
76.82%
69.32%
Danbury, CT
$58,620
96.54%
17.51%
Flint, MI
$23,151
100.00%
3.11%
San Jose, CA
$55,451
92.23%
41.14%
Johnstown,
PA
$23,167
99.79%
1.37%
Washington,
DC/MD/VA
$51,388
96.19%
20.37%
Muncie, IN
$23,180
99.01%
2.50%
$23,284
80.51%
67.06%
San
FranciscoOaklandVallejo, CA
$49,423
93.03%
29.12%
McAllenEdinburgPharrMission, TX
Trenton, NJ
$49,249
95.51%
21.52%
Joplin, MO
$23,720
98.67%
3.52%
$23,799
84.88%
61.44%
$23,806
99.42%
1.94%
Bridgeport,
CT
New YorkNortheastern
$49,249
96.48%
19.15%
BrownsvilleHarlingen-San
Benito, TX
$47,053
93.21%
28.42%
Danville, VA
230
NJ
Boston, MANH
$47,008
96.76%
16.71%
Sharon, PA
$23,827
100.00%
2.34%
MonmouthOcean, NJ
$44,871
97.81%
11.30%
Yuma, AZ
$24,355
86.09%
3.44%
Educational attainment level
“YEARSED” is an essential control
variable. Education is arguably one of the
most significant factors in wage and
salary determination and should be
accounted for in all regressions.
Education level is likely correlated with
the principal independent variables,
English-speaking and bilingual ability, in
addition to other race controls capturing
black and Hispanic populations (See
Figures 2, 3). The exclusion of
“YEARSED” from any regression will
result in omitted variable bias.
link between minorities and higher levels
of poverty. General population is
included as a control variable, as well.
Scatterplots of the data do not
demonstrate an obvious relationship
between income and English ability nor
income and bilingual ability (See
Appendix). In examining both the top
ten richest and poorest MSAs in our
study, a few notable trends emerge
(Table 2). All of the rich cities have a
percentage of English-proficient speakers
between 92% and 98%. Their bilingual
populations are all above the mean
clustering in 20-30% range. These
statistics are not especially surprising as
one would equate a successful economy
with strong ability to communicate. The
data for the poor MSAs does not have the
same uniformity. Laredo, Texas, the city
with the lowest average income, has the
smallest English-speaking population
and one of the largest populations of
bilingual inhabitants. However, the
second poorest MSA, Flint, Michigan, has
largely opposite characteristics. Flint
boasts a population with 100% English
proficiency but a comparatively low level
of bilingual ability of 3.11%. Similar
variation exists in the other highlighted
poor MSAs. It is difficult to make any
generalization in regard to the linguistic
abilities of the lowest income cities in
our study.
Additional important variables
include the foreign born population
“FOREIGN,” and the percentage of a
population identifying as Hispanic
“HISP” in a specified MSA. Given the
characteristics of U.S. immigration, a
strong relationship exists between
language ability, foreign birthplace, and
the Hispanic ethnicity. It is likely that in
a location that attracts large scale
immigration, both significant bilingual
and non-English speaking populations
are prevalent. Perhaps a bilingual
individual lives among non-English
speaking family members or attracts new
immigrants that have not adopted the
language immediately. These controls
help evade any bias related to their
overlap. Similarly, I include a control for
the percentage of a population which is
black “BLACK,” due to the established
231
Hypothesis & Specification
time-related variation, and
represents an MSA’s population.
Although the raw data does not
demonstrate any clear correlation, this is
not a complete representation. I
hypothesize that English ability does not
have any significant relationship with
average income in a metropolitan area.
While previous studies have asserted that
English ability is an important factor in
an individual’s wage, I contend that there
is no negative effect resulting from a high
percentage of non-English speakers on
other members of an MSA (Park 1999). I
expect the inconvenience arising from
inability to communicate is not large
enough to impose a city-wide reduction
in income.
In order to test the
hypothesis
econometrically,
the
following regression is done:
“POP”
In regard to bilingualism, I predict
that a higher bilingual population will
result in a higher average income across
MSAs. According to the research
conducted by Fry (2003) bilingual
speakers
perform
superiorly
in
academics.
Their
strengths
likely
translate into work related activity.
Moreover, in a bilingual society,
constituents are better able to transfer
skills and ideas across cultures. Benefits
resulting from linguistic-- and thus
cultural-- diversity across cities are
established in the study of Ottaviano and
Peri (2005).
Bilingual ability will
certainly only facilitate the distribution
of these cultural elements. In order to
test this hypothesis econometrically, the
following regression is employed:
LWAGE = β0 + β1*ENG +
β2*POP + β3*YEARSED +
β4*FOREIGN + β5*DYEAR
+ β6*BLACK + β7*HISP
LWAGE = β0 + β1*BILING +
β2*POP + β3*YEARSED +
β4*FOREIGN + β5*DYEAR
+ β6*BLACK + β7*HISP
“LWAGE” represents the log of
wage and salary income, “ENG” is the
percentage of proficient English speakers
within the MSA, “FRACT,” is a measure
of linguistic fractionalization within the
MSA, and “YEARSED” accounts for the
city-wide average education level.
“FOREIGN,” “BLACK” and “HISP”
measure the percentage of foreign-born,
black,
and
Hispanic
populates,
respectively. “DYEAR” is a dummy
variable for 2000 data that control for
“BILING”
represents
the
percentage of the population with
bilingual ability. I also include “ENG” and
“BILING” in the same specification, along
with an interaction of the terms. The
utility of bilingualism for translation and
communication likely depends on the
magnitude of non-English speakers
within
a
given
MSA.
232
Results
Table 3: Estimated Income Returns of English & Bilingual Ability
Specification:
Constant
English-speaking
population
Bilingual population
English/bilingual
interaction
Average years of
education
Foreign-born population
Year 2000 dummy
1
6.53***
(0.31)
1.14***
(0.41)
2
7.84***
(0.39)
-0.14
(0.47)
0.19***
(0.01)
1.05***
(0.15)
0.05***
(0.01)
0.19***
(0.01)
1.04***
(0.15)
0.04***
(0.01)
0.02
(0.05)
-0.35***
(0.07)
0.00***
(0.00)
0.58
Black population
Hispanic population
Population
Adjusted R2
***1% significance level
**5% significance level
0.00***
(0.00)
0.56
Table 3 displays the results of an
econometric analysis measuring the
effect
of
an
English
proficient
population, as well as a bilingual
population on average income of a
metropolitan area. In a general OLS
regression, population of proficient
English speakers appeared to have a
positive and significant relationship with
income; however, once controls are
added
for
black
and
Hispanic
populations,
the
coefficient
loses
significance (See columns 1&2). Initially,
the English- speaking population variable
Standard OLS
3
4
7.75***
7.96***
(0.16)
(0.46)
-0.26
(0.52)
0.06
0.09
(0.16)
(0.18)
0.19***
(0.01)
1.05***
(0.10)
0.04***
(0.01)
0.01
(0.05)
-0.38***
(0.11)
0.00***
(0.00)
0.58
0.19***
(0.01)
0.98***
(0.18)
0.04***
(0.01)
0.02
(0.01)
-0.42***
(0.14)
0.00***
(0.00)
0.58
5
9.93***
(0.52)
-2.31***
(0.58)
-4.41***
(0.65)
5.12***
(0.72)
0.20***
(0.01)
0.48***
(0.19)
0.04***
(0.01)
0.07
(0.05)
-0.49***
(0.13)
0.00***
(0.00)
0.61
was capturing the effect of Hispanic
residents in a city. According to column
2, a 1% increase in the Hispanic makeup
of an MSA is associated with a 35%
decrease in the average income in this
city. An explanation of general
Hispanic/Latino
demographic
characteristics
may
explain
this
relationship. According to a study
conducted by the US Census Bureau in
2002, Hispanics are more likely to have
dropped out of high school and to be
below the poverty line than non-
233
Hispanic
Report).
whites
(2000
US
Census
when a metropolitan area has abundant
English proficient speakers, the setbacks
of failure to assimilate are offset. In this
environment, it is possible that diversity
benefits are better able to emerge.
The specification in column 3
examines the effect of a bilingual
population on the earnings in a
metropolitan area. According to the
regressions, it is an insignificant
determinant of average earnings. The
final column considers the interaction
effect of an English speaking and
bilingual population
on a
city.
Intuitively, an area with low-English
proficiency would benefit from having a
large bilingual population in order to
overcome
communication
barriers;
however, the regression points to
different results. The English-speaking
population has a significant and negative
relationship with average earnings.
Likely this result captures some
endogeneity due to greater immigration
in cities that have relative economic
success as concluded by Ratna et al
(2009). It may also indicate that cities
with less diversity are worse off, a finding
established by Ottaviano and Peri (2005).
In either case, it does not lend itself well
to the Pro-English argument. The
bilingual variable also has a negative and
significant relationship with income.
This result is also counterintuitive and
probably reveals a lack of assimilation in
cities that have a lot of people who
speaks their native tongue with their
families inside the home. In addition,
bilingual speakers generally cluster at
extreme ends of the education spectrum
(Fry 2003). It is possible that by
eliminating native born bilinguals I am
capturing only the less-educated. The
interaction term is positive and large in
magnitude.
Again, this outcome is
unexpected. It may demonstrate that
In all regressions, average years of
education, foreign-born population, and
the year 2000 dummy are all significant.
According to the results, a 1 year increase
in the average education level across a
city is associated with a 0.19% increase in
average income.
The foreign-born
percentage of the population is positive
and captures the endogeneity effects of a
successful city attracting foreign-born
immigrants.
The fixed effects regressions
employed in Table 4 aim to reduce
omitted variable bias by controlling for
random variation across metropolitan
areas. The results do not vary greatly.
Both English-speaking ability and
bilingual
ability
prove
largely
insignificant in determining average
earnings within a city. Again, the English
language variable probably captures
some
endogeneity
as
well
as
disadvantages from a homogenous
community. The negative sign on the
bilingual variable may indicate a
resistance to assimilation by people
maintaining their language within the
home or living in isolated enclaves. In
column 5, both variables as well as their
interaction term are highly insignificant.
Nonetheless, the education, foreignborn, and year 2000 dummy variables are
all positive and significant across
specifications. According to these results,
a 1 year increase in the city-wide average
education is associated with a 0.13%
increase in average earnings.
234
economic well-being of an MSA. Despite
convincing literature on the subject, the
results are counter-intuitive. Rather than
indicate communication facilitation and
cultural exchange, the variable may
capture resistance to assimilation
through preservation of one’s native
tongue.
In accord with my hypothesis, the
aggregate English speaking ability of a
city has no effect on average earnings
Any private or externality effect imposed
by barriers to communication are not
large enough to affect an MSA as a
whole. On the other hand, bilingual
ability does not appear to increase the
Table 4: Estimated Income Returns of English & Bilingual Ability
Specification:
Constant
English-speaking
population
Bilingual population
English/bilingual
interaction
Average years of
education
Foreign-born population
Year 2000 dummy
1
7.77***
(0.58)
0.72
(0.55)
0.13***
(0.03)
0.78***
(0.29)
0.02***
(0.01)
Black population
Hispanic population
Population
Adjusted R2
***1% significance level
**5% significance level
0.00
(0.00)
0.54
Fixed Effects
2
3
4
7.71*** 8.32***
7.29***
(0.60)
(0.35)
(0.63)
0.76
1.19
(0.57)
(0.61)
-0.35
-0.58
(0.28)
(0.18)
0.13***
(0.03)
0.70***
(0.29)
0.03***
(0.01)
0.07
(0.22)
0.20
(0.30)
0.00
(0.00)
0.47
Conclusion
0.15***
(0.03)
0.68***
(0.29)
0.03***
(0.01)
0.14
(0.21)
0.25
(0.31)
0.00
(0.00)
0.47
0.13***
(0.03)
1.02***
(0.34)
0.03***
(0.01)
0.07
(0.22)
0.45
(0.33)
0.00
(0.00)
0.48
5
7.46***
(0.88)
1.02
(0.89)
-1.19
(2.39)
0.67
(2.47)
0.13***
(0.03)
1.01***
(0.34)
0.03***
(0.01)
0.08
(0.22)
0.45
(0.33)
0.00
(0.00)
0.50
globalization
that
encourages
multilingual ability. This paper aims to
uncover any economic legitimacy of
these issues. In examining the effect of
an MSA’s population of English
proficient speakers, I find that there is no
Foreign language issues have
pervaded popular media spotlight due to
concerns about immigrant assimilation
through learning English and increased
235
clear link to average earnings within that
MSA. Once fixed effects were accounted
for,
any
association
disappeared.
Furthermore, this paper finds that the
bilingual-able population of an MSA has
no effect on average income. Some
specifications provided contrary results,
but once fixed effects were eliminated,
the coefficients are largely insignificant.
does not account for native born
bilinguals, it is possible that the variable
represents
a
largely
uneducated
population (2003).
In light of this paper’s results, it is
possible to reexamine political concerns.
While other issues certainly factor into
legislation, these findings indicate that
while promoting English may be
beneficial on the individual level, setting
a national language is unnecessary as the
majority of the population is largely
unaffected by immigrants’ language
proficiency.
Furthermore,
foreign
language education may be readdressed.
Although
this
study
does
not
demonstrate any clear benefits of
bilingual ability, it is possible that by
eliminating American born bilinguals the
results have been skewed. In addition,
most of the bilinguals likely speak
Spanish and English. Perhaps knowledge
of a more obscure language has stronger
advantages. Despite the regression
results, it is probable that there is
something to be gained from bilingual
ability.
Language proficiency is a major
barrier to success for immigrants hailing
from non-English speaking countries.
Previous literature has established a link
between English fluency and higher
wages for immigrants. However, if an
individual is unable or choose not to
become
proficient,
any
resulting
disadvantage is exclusive to the personal
level. This result is largely expected as it
is difficult to imagine a non-English
speaking population becoming so
pronounced that trading and economic
relations are strained. Furthermore, it is
possible that despite communication
barriers, the general public benefits from
cultural diversity and new skills
introduced by a non-native (Ottaviano &
Peri 2005).
The results of the bilingual ability
analysis are much less intuitive. All
research indicates that bilingual ability
should increase income. However, this
was proven incorrect on the individual
and now MSA level. This result may be
explained by the skill quality of most new
settlers in the United States. As most
immigrants are considered to be of the
“low skill” quality, becoming bilingual
may not have the capability of moving
them to a higher income bracket. Fry’s
findings
about
education
levels
underscore this issue. Since the study
236
Bibliography
Easterly, W., and R. Levine. (1997).
“Africa’s Growth Tragedy: Politics and
Ethnic Divisions,” Quarterly Journal of
Economics 111(4), 1203-1250.
Ottaviano, G.I.P., and G. Peri. “Cities and
Culture.” Journal of Urban Economics 58,
2005.304-337.
Park, Jin Heum. “The Earnings of
Immigrants in the United States: The
Effect of English Speaking Ability,”
American Journal of Economics and
Sociology 58. 1999. 43-56
Fry, Richard & Lowell, B. Lindsay. 2003.
"The value of bilingualism in the U.S.
labor market," Industrial and Labor
Relations Review, ILR Review, ILR School,
Cornell University, vol. 57(1), pages 128141, October.
Ratna, Nazmun N. & Graton, R. Quentin
& Kompas, Tom. “Is Diversity Bad for
Economic Growth? Evidence from Statelevel Data in the US.” Journal of SocioEconomics 38, 2009. 858-870.
Holman, Rhonda. “Don’t leave foreign
languages left behind.” The Wichita
Eagle.
Feb
09
2010.
<http://www.kansas.com/2010/02/09/117
2748/dont-leave-foreignlanguagesbehind.html>
“States with Official
English,
Inc.
english.org/view/13>
“Languages spoken in the United States”
Vistawide World Languages & Cultures.
Feb
20
2010.
<http://www.vistawide.com/languages/u
s_languages2.htm>
English.” U.S
<http://us-
“Why Official English?” ProEnglish:
English Language Advocates. Feb 20 2010.
http://www.proenglish.org/issues/offeng
/index.html
237
Appendix
238
www.youarewhatyoueat.com: The Religious Nature of
Healthy Eating Blogs
Kate Gundersen, Class of 2010
Introduction
some healthy late-night
tomorrow evening.
Hunger strikes, and I am sitting in
the library. Not having packed a
nutritious apple like my smug neighbor
in the next cubicle, I begin the search for
nourishment. The library’s café offers a
wide array of choices, and an intricate
decision-making process ensues. A
banana, perhaps? Healthy, for sure. My
snack-toting neighbor might smile upon
me like we belong to the same
confraternity
of
fruit-worshippers,
although I can almost hear the scoff of
the environmental studies major to my
left at my insensitivity to carbon
emissions. Cherry Danish? No fiber or
protein, or anything, really, but butter
and sugar, so I’ll be hungry again in an
hour. The glowing vending machine
lures me to its glassy façade, and it
appears
I’ve
internalized
those
anthropomorphic
Peanut
M&Ms
characters: “You’re working so hard! Go
ahead and indulge in this tasty treat.
Candy makes you feel good!” I shove a
dollar bill in the machine and the yellow
package comes writhing down the glass. I
do deserve these forbidden balls of highfructose corn syrup goodness! I do! I
return to my desk, and apple girl’s look is
one of undeniable disapproval and scorn.
I should have never given in to that
vending machine. I should have bought
the banana. I vow to avoid those M&Ms
for the foreseeable future, and to pack
my own peanut butter sandwich for
munching
Food choices, preferences, and
habits are incredibly complex, and
influenced by biological, social, and
cultural factors. The way we eat is crucial
to understanding our selves and our
society, and food has long been a focus of
sociological and anthropological inquiry
in various capacities (Mintz & DuBois
2002). With this perspective on food, the
simple midnight snack in the library
mentioned above is not as trivial as it
appears: a closer look at the agonizing
choice of Peanut M&Ms, for example,
reveals how concerns of environmental
justice, popular nutritional discourse,
media
and
advertisement,
peer
judgments, and a myriad of other factors
influence a seemingly “simple” food
choice.
As with nearly every aspect of
American
life,
the
internet
is
revolutionizing the way we eat. The
increasing popularity of blogging
(maintaining a “web log,” a personal
website updated sporadically by its
author, the “blogger”) has generated
online forums for the discussion of many
aspects of food, drink, cooking, dining,
and eating. A particular niche of food
blogs has emerged, devoted to the
blogger’s documentation of her food
intake (the bloggers are overwhelmingly
female).
Daily
posts
consist
of
239
prominent, eye-catching photos of meals
and snacks alongside detailed text
narration of the day; these bloggers
create the story of their lives with food as
the central, organizing structure. The
self-proclaimed purpose of the “healthy
living blogs” is not for the blogger to
engage in or promote extreme weight
loss or calorie restriction, but to serve as
a place to document her commitment to
leading a “healthy,” “balanced” life, and
to connect with others with similar goals.
And others are certainly interested and
reading. A true community has emerged
around these blogs. Although nearly all
interaction is internet-based, bloggers
form relationships with each other and
with their readers, who frequently
contribute to the conversation via the
“Comments” section at the end of each
post.
This paper will interpret these
healthy-living blogs as religious in
nature. The virtual community of
bloggers and readers unite and rally
around the belief in “healthy,” “balanced”
living, asserting the Durkheimian
“unified system of belief” that lies at the
foundation of
collective
religious
behavior (Thompson 1982:129). The
bloggers explicitly state these values
online then enact them, in a Geertzian
sense, through daily “ritual” practices.
The language and images that the
bloggers publish on their daily posts will
be analyzed in order to uncover how
food is understood and practiced in the
lives of these women, and to
demonstrate the religious character of
their communal journey towards
“health.”
With national obesity rates on the
rise (CDC 2009) and a big-business food
industry whose practices only exacerbate
the problem, these bloggers appear to
avoid
this
widespread
American
condition by staying consistently
devoted to their pillars of nutrition and
health. Delving into these blogs - these
“extremely small matters,” (Geertz
1973:21) - offers perspective on the
devotion and effort that are perhaps
necessary to practice “good health” in
this country. Furthermore, the public
forum of the Internet has brought to
light the typically “backstage” behaviors
and emotions concerning everyday food
consumption (Goffman 1956). The
accessibility of the internet and the
ability to blog – to connect and
communicate with like-minded others
throughout the nation and the world –
has created and fostered the bloggers’
food-based religion and its surrounding
This philosophy of food as a
means to health, life balance, and
fulfillment is common across all 20 of the
blogs analyzed in this study. Food is
about more than a slim body for these
women; although they certainly exhibit a
concern for staying trim, eating “well,”
according to their definition, is linked to
a
deeper
sense
of
personal
accomplishment and empowerment.
These blogs reflect a larger, societal
perception
that
consumption
is
indicative of one’s moral character – that
how we eat truly reflects who we are
(Griffin & Berry 2003:45, Madden &
Chamberlain 2004:585). The online
community that has assembled around
these blogs exhibits a strong devotion to
the goal of health. They are steadfastly
united to embody their common values,
and to attain the virtue and satisfaction
of a “healthy” lifestyle.
240
community. Through these healthyliving,
this
paper
explores
the
burgeoning role of the internet in the
personal negotiation of food choices and
habits. Online activity has completely
shaped and transformed the bloggers’
relationship with and understanding of
food, and this influence is not isolated.
The overwhelming ubiquity of the
internet in everyday life implicates that it
does - and will increasingly continue to affect the way we eat.
can connote animalistic vitality, while a
vegetarian diet can symbolize the
consumer’s purity (through eating
“whole” foods) and compassion for living
creatures (Twigg 1983:28-29). Foods hold
various and changing social meanings
and connotations: drinking milk, for
example, brings to mind both the purity
and innocence of a small child, and the
strong, calcium-rich body of an athlete
(Atkinson 1983:10-11). In the United
States and throughout the West,
advertisements play a key role in
enforcing the virtues attributed to foods
– qualities that can be embodied by the
consumer if only they purchase and eat
the product. Atkinson (1983) discusses
how health food marketing in particular
draws on the myth of the pastoral,
“natural” “good life,” luring customers to
buy “farmhouse” bread or cheese that
might transport them, too, back to the
purity of the farm (15-16).
Discussion of Relevant Literature
No sociological or anthropological
literature has been published regarding
the specific blogs of this study, as these
blogs are a recent and not particularly
publicized phenomenon. However, the
role of food in society has been
exhaustively studied in the fields of
sociology and anthropology. Food
consumption patterns are driven by
more than simple nutritional need
(Mintz & DuBois 2002), and numerous
theoretical frameworks have been
created (and critiqued) to attempt to
codify and understand the vital human
act of eating, and its social implications
(Wood 1995). This paper’s analysis is
based upon the general understanding
that food choice and habits can
communicate information about the
consumer’s
social
context
and
surrounding community (Wood 1995:1423; Anderson 2005:109-139; Bourdieu
1984:ch. 3; Atkinson 1983). Especially
important is the assertion that culturallyascribed meanings and properties of food
items themselves are thought to be
transferred to the consumer when foods
are eaten (Madden & Chamberlain
2004:584-586). Eating meat, for example,
Madden and Chamberlain (2004)
argue that especially as food choices are
increasingly equated with physical health
concerns such as obesity and heart
disease, the pressure is high to avoid the
“wrong” foods – foods that are culturally
perceived as unhealthy or detrimental to
the body (584-585). One who does
“indulge” in M&Ms or French fries and
subsequently exhibits an “unhealthy”
(read:
large)
body
is
deemed
irresponsible, lazy, lacking in selfcontrol, and a moral failure (Madden &
Chamberlain 2004:585). It is the personal
responsibility, and even moral duty, of
the individual to effectively control the
quantity and quality of food intake.
McGuire (1988) uncovered this belief in
her study of alternative, holistic health
practice among suburban, middle-class
241
Americans. These communities often
considered health to be a truly spiritual
or religious endeavor; living free of
illness implied physical and moral
accomplishment (McGuire 1988:247249). The failure to achieve the “desired
state defined as health” through personal
control and discipline of behavior was
considered
“morally
reprehensible”
(McGuire 1988:249).
The steadfast pursuit of this “morality” of
food is certainly represented in the
healthy-living blogs of this study. The
internet has emerged as a forum to
discuss and negotiate conflicting,
confusing
messages
conveyed
in
American society about health, food
choices, and the ideal body type.
Increased access to the internet has
allowed individuals to project their
opinions and identities through the fairly
new medium of personal home pages
and, more recently, blogs (Harp &
Tremayne 2006:248; Stern 2007:160-161).
Blogging is a means of narrating and,
thus, making sense of one’s life, identity,
and situation in society (Bell 2007:102;
Somers & Gibson 1994:58-59). Blogs and
internet participation also foster social
bonds and the support of community
without any face-to-face interaction
(Rettberg 2008:60-61).
Griffin and Berry (2003) point to
the transfer of Calvinist asceticism from
the church to the world of capitalism as
the reason why “consumption has come
to be taken for granted as a morally
correct ethos… a sense of social and
moral standing come from what we
consume and normative assumptions of
what should be consumed” (45). Foods
are thus deemed “good” or “bad” based
on how they fit on the path towards
health and, thus, morality (Madden &
Chamberlain
2004:588-589).
Advertisements especially exploit the
notion that food choice dictates moral
worth and frequently use religious
rhetoric to do so: messages in food
advertisement alternate between the call
for restraint from “bad” foods, and the
urge to indulge in the “sin” of forbidden
ice cream bars or cheesecake (Griffin &
Berry 2003; Madden & Chamberlain
2004:589). These advertisements suggest
the serious implications of consuming
any food product.
Literature concerning the role of
the internet in food preferences and
practices is relatively sparse. One online
niche that has been the focus of recent
sociological study is that of pro-anorexia
and pro-bulimia (“pro-ana” and “promia”) web sites and message boards.
Contributors to these sites work to
promote and encourage the selfdiscipline,
control,
and
restraint
necessary to achieve their unifying goal
of the “thin body ideal” (Malson,
2009:357; Day & Keys 2008:7-8). The
literature pertaining to these sites is
most relevant in unpacking the
sociological significance of online
discourse of food consumption. The
basic function of the pro-ana websites
seems surprisingly similar to the food
blogs’: to assert and perpetuate a
common belief related to eating, and to
When
cultural
influences
communicate that the choice between an
apple and a package of Peanut M&Ms
will determine one’s inner worth, an
enormous
amount
of
discipline,
restraint, and self-control seem necessary
at all times when interacting with food.
242
Theory
connect with others who share that belief
(Day & Keys 2008, Malson 2009).
Healthy-living blogs and pro-ana sites
provide similar features (the ability to
write about oneself or to respond to
another contributor), and the users
reflect a comparably regimented,
methodical approach to food (Day &
Keys 2008:6, 8). What is so interestingly
different about these virtual spaces,
though, is that the pro-ana sites are
commonly viewed as deviant, disturbing,
and
undoubtedly
unhealthy.
The
bloggers of this study, despite a similar
preoccupation
with
food
as
demonstrated through the internet, are
frequently considered to be pillars of
health and, one could presume, morality.
The conception of food as a
means to inner goodness and an
enlightened life is prevalent in the blogs
of this research. Bloggers and readers
conflate certain food choices and habits
with the larger motivation of “good
health” – the sign of true personal
achievement and worth. The bloggers’
daily behaviors are regimented, timeconsuming, and driven by a common
understanding that health is of utmost
importance in life and existence – and
that their behaviors are the means to
achieve “good health.” In some ways, and
with all due respect, the bloggers’ daily
practices echo those of a Catholic who
attends
weekly
confession,
understanding this habit as a way to
prove and enact her abstract belief in a
loving and forgiving God. Similarly, a Jew
who keeps a kosher kitchen practices
these eating habits because he believes
that following God’s law (as stated in
biblical text) is the path to a religiously
fulfilling life and, eventually, heaven. The
healthy-living bloggers and formal
religious believers share the perception
of an inextricable link between collective,
abstract, intangible belief and personal,
active practices as a way to enact those
beliefs. Although the bloggers do not
physically unite for sporadic services or
meetings – as followers of formal religion
might – they use the internet to publicly
affirm and discuss their beliefs and
practices on a daily (or even thrice-daily)
basis. These comparisons, however
informal, suggest the typically religious
intersection of strong belief and devoted
practice within this cohesive community
of bloggers. It should be noted that these
bloggers do not comprise a subculture.
Studying this new venue for the
presentation of food practices fills a
serious void in current literature
concerning the role of the internet in
constructing food-based identity and
community. In the United States, being
online has become as routine and
habitual as gathering around the dinner
table – in some cases, perhaps even more
so. The healthy-living blogs of this study
represent one of many ways that these
two ubiquitous actions of everyday life –
eating and internet use – intersect. The
public nature of the internet, and
particularly of blogs, has been widely
studied (Harp & Tremayne 2006; Hevem
2004; Loft 2005; Rettberg 2008; Weber &
Dixon 2007; Pauwels 2008). Investigation
of these public virtual spaces as a means
to understand current food practices,
choice, and habits, and their societal
implications, however, is largely absent
in the existing literature.
243
Their aspiration for health, and
approaches to reach this goal, are rooted
firmly in the mainstream; in fact, their
behaviors and philosophy often echo
popular health-related discourse and
trends, such as support for local, organic
foods and the USDA’s food pyramid.
Their behaviors, though embedded in
the secular world, markedly embody the
traits of a religion.
their theory in order to convey the
religious nature of the blogs of study.
1. Durkheim: A Unified System of Beliefs
Emile Durkheim’s definition of
religion notes that the existence of “a
unified system of beliefs” is the
foundation for the creation of a group of
religious adherents (Thompson 1982:129).
Specific beliefs are not inherently
religious, though; rather, Durkheim
asserts that, “once a goal is pursued by a
whole people, it acquires, as a result of
this unanimous adherence, a sort of
moral supremacy which raises it far
above private goals and thereby gives it a
religious character” (Giddens 1972:23). It
is the collectivity of belief in shared
“sentiments
and
ideas”
(Giddens
1972:243)
that
intensifies
their
significance and imbues them with
religiosity. The healthy-living blogs of
this study publicly affirm the shared
value of “good health”; the discrete ways
in which they enact this overarching goal
can be interpreted through the second
aspect of theoretical framework:
In this paper, the niche of the
healthy-living blogs will be interpreted as
religious in nature using two aspects of
major sociological frameworks regarding
religion: Emile Durkheim’s conception of
the church as a group with an explicit,
unified system of common beliefs and
values; and Clifford Geertz’s emphasis on
ritual acts as the way in which church
members concretize and manifest these
shared beliefs.
According to Durkheim, religion
extends beyond the conventional
conception of formal and organized
religious practice (Giddens 1972:25).
Rather, “religion is something eminently
social”; the beliefs and practices of
religion
represent
the
“collective
realities” of society itself (Thompson
1982:125). Religious practice or belief
does not mandate a “God” to worship, a
priest to preside over a service, or even a
physical structure in which to gather
(Giddens
1972:223).
The
virtual
community surrounding these healthyliving blogs is certainly reflective of
Durkheim’s and Geertz’s recognition of
the religious nature of secular, social life.
These theorists’ work on the subject of
religion is extensive; this paper will draw
on and discuss only a small portion of
2. Geertz: Ritual Practices as Models of
and for Belief
Clifford Geertz, in his essay
“Religion as a Cultural System” (1993),
asserts that collectively practiced
behaviors or actions (“rituals”) are a
means to concretize a devotion to
abstract, intangible values (112). Carrying
out a ritual action is a way to
demonstrate religious beliefs to others:
to use Twigg’s (1983) example of
vegetarianism, abstaining from eating
meat is a public exhibition of the
vegetarian’s conviction that animal
cruelty is immoral. Simultaneously,
244
though, Geertz explains that for the
individual participant, religious ritual is
an active step towards achieving the
collective religious goal; rituals are “not
only models of what they (the religious
group] believe, but also models for the
believing of it…men attain their faith as
they portray it” (1993:114). In the
performance of ritual, the system of
belief is physically materialized and “the
world as lived and the world
imagined…become the same world”
(Geertz 1993:112). The bloggers’ rituals
include the meticulous preparation and
consumption of meals, in addition to the
reflection and discussion of their food
choices through the act of updating the
blog (“blogging”). The healthy-living
bloggers’ active documentation of meals
exhibits their belief system to an outside
(online) viewer. At the same time, these
ritual practices allow the bloggers to
concretize and then perpetuate their
values, allowing them to personally
progress towards their wider goal of
health.
and concretize the abstract goal of
a “healthy” lifestyle and allow the
bloggers to move towards that
goal.
Methodology
The starting point for the study of
these blogs was Kath Eats Real Food,1 one
of the most popular sites of this blogging
niche. I investigated other healthy-living
blogs through Kath’s “blogroll,” a feature
wherein the blogger posts links to related
blogs of interest. A list of 35 blogs was
compiled (see Appendix 1, 2) that fit
certain criteria: 1) the blog is updated
nearly daily; 2) the purpose of the blog is
to highlight a healthy, “normal” (in the
blogger’s words) lifestyle, as opposed to
attempts at extreme weight loss or
dieting; 3) the blog records all meals of
the blogger in a diary-style format, as
opposed to sporadic recipes, or posts
unrelated to food intake. This initial
survey was meant to investigate the
prevalence of these blogs, and supports
the fact that this blogging is not an
isolated activity, but that it is thriving and growing. While some of the blogs
have been maintained for over a year
(Kath Eats Real Food recently celebrated
its second anniversary), several bloggers
of the sample specifically credited the
inspiration of other healthy-living blogs
for the more recent creation of their own
blog. It should be noted that females
authored 34 of the 35 blogs on this list,
and that bloggers hail from all across the
United States and Canada, as do readers
who contributed comments.
This paper will prove that these
healthy-living blogs do not just faintly
echo
the
religiosity
and
moral
implications often associated with food
choices. Through the sociological lens of
Durkheim and Geertz, the community
surrounding these blogs functions as
religious. Data analysis will demonstrate
that:
1. The bloggers assert a “unified
system of beliefs” pertaining to
food and health.
2. Their eating habits, and the public
commentary concerning those
habits (the act of blogging),
function as “rituals” that reinforce
1
245
http://katheats.com.
From this master list of 35 blogs of
similar structure and content, 20 were
selected in order to study the “My Story”
or “About Me” pages (see Appendix 1).
These sections typically communicate
the blogger’s general philosophy of food
and health. The list was narrowed to 20
based on my assessment of which blogs
best represented this community; for
example, one blog, although it met the
preliminary criteria, focused in much
more detail on exercise habits than on
food. This blog – and others similar to it were eliminated when forming the list of
20. The singular blog authored by a male
was also eliminated, but not on the basis
of the blogger’s gender. Admittedly, this
selection process involves some level of
personal bias. However, I was confident
in my ability to choose the blogs that
would provide the most food-related
content and, thus, the most analyzable
data.
photos in these sections (see Data
Sources). The 20 blogs were then
codified by the life stage of the blogger:
college life (18-22, attending an
undergraduate
institution);
early
adulthood (20s and early 30s, living
independently or with a significant
other); and middle age (mid- to late-30s,
usually married and or with children).
The blogs were also codified by
popularity, determined by the average
number of user comments per post: low
popularity (0 – 20 comments per post);
middle popularity (20 – 40 comments per
post); high popularity (40+ comments
per post). High-popularity bloggers
could be considered the pseudocelebrities
of
this
community.
Corporations sometimes provide them
with food items to review on the blog, or
sponsor their attendance at nationwide
blogging conferences because of their
high site traffic. These high-popularity
bloggers often make a small income from
their blogging through advertisement
and offers for freelance work.
Data was collected from the
“About Me” or similarly named sections
of the 20 blogs, including all text and
Table 1. Demographic Breakdown of 20-Blog Sample
College Life
Low Popularity
Middle Popularity
High Popularity
(0-20 comments/post)
(20-40 comments/ post)
(40+ comments/post)
3
1
0
4
4
5
3
0
0
(ages 18-22)
Early Adulthood
(20s – early 30s)
Middle Age
(late 30s+)
246
With these classifications, the list
of 20 blogs was then narrowed to a
representative sample of 8 (see Appendix
3). A smaller sample allowed for a more
detailed analysis of blog posts, while still
ensuring that the sample represented the
themes and trends of the wider
community. The blogs of the smaller
sample were chosen to reflect the
frequency
of
the
demographic
codifications of the sample of 20. Again,
this introduced my personal selection
bias, but at this point in the research
process I was following the 20 blogs
almost daily, and had become familiar
with the bloggers and their posting-style.
Thus, I chose a sample that both fit the
distribution of the original sample and
reflected what I believed to be a diversity
of perspective and voice. The majority of
the blogs of this nature are authored by
females in Early Adulthood, and the 8blog
sample
reflects
this
fact.
Table 2. Demographic Breakdown of 8 Blogs Chosen for Detailed Analysis
College Life
Low Popularity
Middle Popularity
High Popularity
(0-20 comments/post)
(20-40 comments/ post)
(40+ comments/post)
NYCollegEats
K’s Good Eats
_
(ages 18-22)
Early Adulthood
(20s – early 30s)
Kath Eats Real Food
Danica’s Daily
Healthy and Sane
Carrots ‘n’ Cake
Middle Age
Running with a Recipe
(late 30s+)
_
_
Krista’s Kravings
ordinary series of events. If the blog posts
were consistent throughout the week,
the three days with the most text and
photo content were chosen. Again, this
introduces my personal selection bias,
but posts were chosen solely based on
my perception of the quantity – as
opposed to targeted quality – of data. 34
posts were chosen in all (see Data
Three days of blog posts were
selected from each these 8 blogs,
between the dates of 6 October 2009 and
13 October 2009 (see Data Sources). This
range of dates was chosen for its
“neutrality” from holiday vacations or
feasts. The week-long window allowed
for flexibility if the blog had not been
recently updated for some reason, or if a
post reflected a particularly out-of-the247
Sources), as some bloggers post more
than once a day.
exploration of the latent themes and
similarities of these women’s experiences
than would the quantitative results of a
survey or a similar, more wide-reaching
method.
The data collected from the
selected posts included all text, photos,
and user comments published on the
selected dates. Drawing on Keats’ (2009)
methodology, an initial reading of the
textual “About Me” data was conducted
in order to become familiar with the
context of each blog and to detect
overarching themes and similarities.
After the initial appraisal, a more specific
reading was conducted to address these
themes. During this second reading, data
was codified into four subcategories that
showed the daily “ritual” practices of the
bloggers as enactments of their wider
belief system: language used to explicitly
indicate food as “good” or “bad”;
language used to describe or reflect on
indulgence; indications of control and
agency; and mental/ emotional reactions
to food intake.
Research Limitations
The greatest limitation posed by
this research is the homogeny of the
sampled bloggers. All were female,
between the ages of 18 and 45, mostly
Caucasian, and with socio-economic
means that allowed for nutritious, fresh
food on their plates on any given day. In
terms of financial ability, this depiction
of eating is far from the reality that many
Americans face, but healthy-living blogs
capturing a lower-income experience are
essentially non-existent. Additionally, a
male perspective was largely absent from
these blogs, save the sporadic mention of
a husband or boyfriend. This mirrors
previous literature showing that men are
more likely to generate “filter-type” blogs
(those focused on news, politics, and the
spreading of information) than women,
who tend towards communicative,
journal-type blogs like the ones of this
study (Bell 2007:99-100; Harp &
Tremayne 2006:250). Research exploring
the pressure surrounding food choices
for people of various social locations and
backgrounds – and the role of the
internet in those choices - would reveal
whether the same implications of food
appear across a wider spectrum.
Qualitative content analysis is an
effective way to understand these
bloggers’ nuanced perspectives on food.
The blog is a method of personal
narration, and because these healthyliving bloggers are comfortable and open
with their online community, the blogs
are a space for an honest presentation of
feelings, beliefs, and experiences, in
addition to more objective information
like food intake. The blogs are a lens
through which to view the complex,
usually private personal negotiation of
eating and health. They demonstrate the
ways in which food structures these
women’s lives, and the influence of
online activity in their everyday
practices. Close reading of the bloggergenerated text allows for a deeper
Analysis of Data
An analysis of the textual content
of these healthy-living blogs proves the
blogging community to be a religious
one, using the key aspects of Durkheim’s
and Geertz’s definitions.
248
attitude towards food consumption is
overwhelmingly “anti-diet”; the bloggers
understand dieting as an unsustainable
means of food choice that “deprives” the
eater of cravings and pleasures. The
pattern of eating that the bloggers
support and believe in is considered a
“lifestyle” as opposed to the temporality
of a “diet.” They believe in eating a
balance of different food groups (to use
the term so popularized in nutrition
discourse), emphasizing “healthy” foods
and only occasionally eating “bad,”
“unhealthy” foods. 50% of the blogs
explicitly state this belief by using the
words “indulgence” “moderation” or
“deprive.” There is a statement of faith
that “it will all even out in the end”
between healthy food and treats.1 Jenn
affirms that: “the occasional pigout or
rich treat will (not] alter your entire
weight loss diet or health for the
duration of your life”.2 The bloggers
believe in indulging one’s cravings from
time to time, and that a healthy lifestyle
is liberating as opposed to limiting and
oppressive, like dieting.
1. A Unified System of Beliefs
The bloggers of this online
religion designate a specific section of
their blogs to clearly and candidly state
their core values concerning food and
health. 20 blogs (see Appendix 2) were
analyzed for this section of analysis to
determine the “unified system of beliefs”
(Thompson 1972:129) of the community.
All 20 included mini-autobiographies of
the bloggers (sections entitled “About
Me,” “My Story,” “My Journey”), and all
20 mention, in varying levels of detail,
the
bloggers’
food
attitudes or
philosophy. These sections of the blogs
function as explicit manifestation of the
bloggers’ collective, guiding beliefs, as
there is a great deal of overlap among
these philosophies related to health. The
bloggers reveal additional information
about their beliefs throughout their daily
posts, but this section of analysis will
focus specifically on the About Me
sections.
The overarching theme that
emerges from the analysis is that of
“balance.” 50% of the sample (n=20)
specifically use the word “balance” in the
About Me section, and 2 include the
word in their blog’s subtitle (e.g. “Carrots
‘n’ Cake: A Blog Dedicated to Healthy
Balance”). Even when bloggers did not
explicitly use the word “balance,” the
data addresses this overarching theme in
three ways:
1.a. Balance
Moderation
of
Diet:
Indulging
However, striking the balance
between indulgence and eating “healthy”
is complex and nuanced, and seems to
require a great deal of vigilance. Many
bloggers explicitly state which foods
should be part of a daily diet in the
About Me section. Foods mentioned as
healthy and good include: vegetables;
fresh produce; products that are natural,
unprocessed,
organic,
free-range,
“whole,” or local. Usually, though, the
in
The most common food-related
belief of the bloggers is that eating
healthily does not entail restriction of
any food, no matter its caloric value. The
1
Eat Like Me. 1 October 2007. Let’s Take a Step
Back and Review.
2
Jenn’s Menu and Lifestyle Blog. About.
http://www.j3nn.net/about/.
249
moderation. (Emphasis present in
4
the original)
term “healthy food” is used without
specification, as though the reader
should immediately understand what
this categorization entails. Bloggers are
much more explicit in noting indulgence
foods to avoid and eat in moderation:
Snickers, pizza, full-fat ice cream,
birthday cake, fast food, Little Debbie
bars, “treats” or “dessert” in general,
chocolate, candy, and French fries were
all mentioned by name. Although the
bloggers encourage the inclusion of all
foods into one’s diet, they are more
acutely aware and expressive of “bad”
foods to be avoided. It is clear that
striking the balance between different
varieties of food is complex, and that the
call to indulge is tempered by a wider
cognizance of health:
This philosophy of a delicate
balance of eating that requires
surveillance, mindfulness, and careful
control is internalized throughout the
blog community. It is manifest in the
daily food choices of the bloggers, as
demonstrated in the posts examined in
section 2 of this paper’s analysis.
1.b. Balance of Food and Exercise: Calories
In vs. Calories Out
Every blog in the sample
acknowledges the necessity of physical
activity in achieving a healthy lifestyle.
Food consumption and exercise are seen
as interdependent, and are often referred
to as an inseparable entity (“Eating
Healthy and Staying Fit”; “eating and
exercise”). Biologically, maintaining an
equal
balance
between
calories
consumed and calories burned ensures
weight stability, and the bloggers
recognize this reality. Although none of
the bloggers are self-proclaimed “calorie
counters,” comparing food intake to
levels of physical activity is an accepted
duty of daily life for nearly all the
bloggers - especially because they believe
in occasional indulgence in caloric treats.
Maintaining the right balance ensures
that one will “never feel deprived”5
because calories eaten will not exceed
calories expended. Exercising more
If you love chocolate, sausages, ice
cream,
candy,
French
fries,
wine/beer or something else less
healthful then by all means have it
and don’t deprive yourself. I do think
we all should practice some form of
restraint to best benefit our longterm health, as these “unhealthy” or
“bad” foods…do not provide the body
3
with…nutrients.
If you want to lose weight once and
for all, commit to a new lifestyle
that includes a “diet” filled with
healthy, nutritious foods 80% of the
time, but allow yourself controlled
portions of the “fun” foods for the
remaining 20%. Don’t think you
can’t splurge once in a while. Let
yourself go crazy once a week, but in
4
Carrots ‘n’ Cake. Weight Loss.
http://carrotsncake.com/weight-loss.
5
Kath Eats Real Food. Kath.
http://www.katheats.com/kath-2/.
3
Eat Like Me. 1 October 2007. Let’s Take a Step
Back and Review.
250
means that one can “afford” a little more
“wiggle room” in caloric intake.6
and general well-being are all equated
with “proper” health-related behaviors;
these feelings are especially clear in the
stories of the bloggers who have
transformed previously-unhealthy eating
habits and now claim to be reaping the
physical and emotional benefits of their
knack for balance. Elina states that, “I am
fit, healthy and happy, and that’s what
truly matters” as opposed to the number
on her scale: “I’m much more in tune
with what I want, physically and
mentally.”9 After a long journey towards
weight loss and health, Kath claims, “I
feel radiant every day…I have never felt
better about myself or been happier
about where I am in life, and I believe it
all stems from a boost in self
confidence.”10 Clearly, the bloggers
believe in the power of physical health to
greatly influence their emotional wellbeing. How and what they eat extends
beyond physical implications and affects
the whole self. Similar to McGuire’s
(1988) findings about suburban, middleclass practitioners of alternative healing,
physical health for these bloggers also
indicates
moral,
virtuous
accomplishment,
and
they
wholeheartedly
believe
in
this
correlation.
Four bloggers point to exercise as
the turning point in the “conversion” to a
healthy lifestyle. These About Me
sections
narrated
the
bloggers’
transformations from, “a beer-guzzling,
anti-gym college student”7 to a runner or
spinning fanatic. 14 of the bloggers
express a true enjoyment of exercise,
even a “passion,” and associate
happiness, positive self-esteem, and
stress relief to physical activity.
The constant attention needed to
maintain this balance can be taxing and
difficult, though: Monica confesses, “I
love running, but I also love eating and
unfortunately I eat more than I burn.”8
Bloggers recommend pursuing a variety
of exercise routines to avoid boredom,
and acknowledge the difficulty of finding
time to run or bike when working at a 9to-5 desk job. Regardless, the blogging
community believes in the virtue and
necessity of exercise, and in monitoring
and maintaining a balance between
physical activity and food.
1.c. Balance of Life in General: Healthy
Habits Impact Everything
Many bloggers declare that
balance in the personal spheres of food
and health extend to wider aspects of
their lives, which can be hectic and full
of commitments to work, significant
others, and children. The titles and selfproclaimed purposes of the blogs
The
bloggers
believe
that
combining healthy eating with regular
exercise – maintaining these first two
aspects of balance – influences much
more than the shape of their bodies.
Happiness, self-confidence, self-esteem,
6
Eat Like Me. 1 October 2007. Let’s Take a Step
Back and Review.
7
Healthy Tipping Point. Tips.
http://www.healthytippingpoint.com/tips.
8
Run, Eat, Repeat. About.
http://runeatrepeat.com/?page_id=2.
9
Healthy and Sane. About.
http://healthyandsane.com/about/.
10
Kath Eats Real Food. Kath.
http://www.katheats.com/kath-2/.
251
my focus is on healthy eating,
12
wellness, fitness, and happiness.
indicate that a devotion to health
provides structure for the bloggers’ lives.
Personal histories are narrated in terms
of food habits and weight, and one blog
(Healthy and Sane) is subtitled, “My Life
in Food.” Additionally, food-related
metaphors are used in abundance: “I
hope you enjoy the journey with me as I
fulfill all of my life’s cravings!”11;
“Running with a Recipe: Savoring the
Ingredients of My Life.” When it is clear
that food is the critical, structuring
aspect of life for these bloggers, the
encompassing belief of the blogging
community becomes present: that eating
“right” and maintaining these forms of
balance bring about general happiness,
fulfillment, and a sense of inner virtue.
Because the attainment of “health” is so
linked to individual food practices, the
bloggers believe that achieving this goal
is always possible for the individual, if
only she takes the initiative to make the
correct choices.
I was…very unhappy with the
relationship I had with food and
exercise, and I knew I was the only
13
person who could change it.
The bottom line is, you can be
healthy if you want to. Make the
choice. There’s no secret. Move
more, eat healthy. If you put your
health first, weight loss comes
naturally. (Emphasis present in
14
original)
It is clear from an analysis of
About Me sections of these blogs that the
bloggers indeed exhibit Durkheim’s
“unified system of beliefs.” The bloggers
are dedicated to the attainment of health
as a constant negotiation of balance, and
they assert their responsibility and
capacity to “take control” and achieve
these goals.
No matter how long the work day
or how tempting the cupcake, these
bloggers prioritize their goals of balanced
eating and exercise; they believe that
these behaviors are the key to overall
well-being and purpose, and that they
are in control. Maintaining balance on so
many levels does indeed require control,
but the bloggers are encouraged and
empowered to exhibit agency in their
own lives:
The following section of analysis
results from a detailed reading of a
sample of 8 blogs, chosen from the
previous group of 20 (see Appendix 3).
Three days of posting were selected from
these blogs, resulting in an analysis of 34
total posts (some bloggers posted more
than once a day). While a wider array of
blogs was helpful in determining the
overall principles and beliefs of this
With a free gym membership and a
new outlook on life, I was finally able
to take control of my weight…Today,
12
Carrots ‘n’ Cake. My Journey.
http://carrotsncake.com/weight-loss/my-journey.
13
Healthy Tipping Point. Caitlin.
http://www.healthytippingpoint.com/about.
14
Kath Eats Real Food. Kath.
http://www.katheats.com/kath-2/.
11
Krista’s Kravings. Krista – About Me.
http://www.blogger.com/profile/14597868827300
760281.
252
community, closely investigating this
smaller sample allowed for a deeper
understanding
of
the
day-to-day
practices (“rituals”) of these bloggers.
blogs. Through daily food behaviors, the
bloggers emphasize this value of balance
by clearly demarcating “good” foods from
“bad,” and discussing the ways that they
navigate equilibrium between the two.
2. Ritual Practices as Models of and for
Belief
2.a.i. “Good food”
For the sake of this analysis, the
“rituals” of the bloggers were defined as
the meals and snacks that are
documented daily on their blogs, in
addition to the act of blogging itself – the
generation
of
discussion
and
commentary regarding their own actions.
Because food is the central organizing
structure of these bloggers’ lives, the
blogs are rife not only with photos and
objective descriptions of what was eaten,
but also with subjective commentary and
reflections on these “rituals.” Analysis of
the discourse surrounding food choices
and habits shows that eating is indeed
ritual for these women, in the Geertzian
sense of the term. In publicly publishing
their food habits online, the bloggers
demonstrate their actions as a guide for
the navigation of health (a model of
belief). Additionally, the way that these
women eat and reflect on their eating is
an active means of embodying the selfprofessed goals and values of the About
Me sections (a model for belief). The
bloggers’ documentation of their rituals
of everyday life concretizes and enacts
their belief system of balance, and the
internet is an indispensable tool in their
ritual.
Analysis of the daily blog posts
reveals both explicit and subtle ways that
the bloggers denote what they believe to
be “good” foods. There is often a
concerted effort to highlight certain
traits of a food product when it especially
follows collective food preferences. Krista
notes that her slice of banana bread is,
“wholesome and homemade”15 and
describes a photo as “a gorgeous,
delicious, organic honey crisp apple”16.
Ava deems Chobani yogurt to be “the
ideal study fuel since it contains twice
the protein of regular yogurt and tons of
calcium”17. Foods that are extolled as
“good” frequently possess inherent traits
such as nutritional content and
provenance (organic, from a Farmer’s
Market, in season, etc.). And although
“processed” foods are largely discouraged
by the bloggers, certain characteristics
can redeem them: “without added
sugar”18 and “a short list of ingredients
that
are
all
pronouncable
and
recognizable,”19 for example. “Good”
foods, though, are not only inherently
nutritional - they are also assessed in
15
Krista’s Kravings. 6 October 2009. Dinner
Pancakes.
16
Krista’s Kravings. 13 October 2009. Couch
Potato.
17
NYCollegEats. 7 October 2009. Greek Yogurt
Gods.
18
Running With a Recipe. 9 October 2009.
Versatile Veggie.
19
NYCollegEats. 7 October 2009. Greek Yogurt
Gods.
2.a. Balance of Diet
The bloggers commonly affirmed
the need for balance between “healthy,”
virtuous food and occasional indulgence
foods in the About Me sections of their
253
terms of the satiety they can deliver. The
“best” foods in the eyes of the bloggers
are nutritionally sound while also being,
“Delish! And filling!”20 The significance of
satiety will be discussed further in the
following section of analysis (2.b.i.).
Although this exacting devotion
to oatmeal is common across so many of
the healthy-living blogs, Kath Eats Real
Food sets the standard for the “worship”
of oatmeal. Her site includes an entire
tabbed section entitled “Tribute to
Oatmeal,” dedicated to “the world’s
greatest
breakfast!”22
The
section
includes
123
high-quality,
almost
pornographically detailed photographs of
various oats combinations, with links to
the meticulous preparation directions to
each, and an embedded video clip of her
cooking technique.23 This dedication
towards the daily ritual of oatmeal
proves it to be the pinnacle of “good”
food – oatmeal is simultaneously
nutritious, satiating, and tasty, and it
allows for personal agency and control in
its concoction.
Oatmeal stands out across the
blogs as the epitome of a “good” food. 6
of the 8 bloggers ate oatmeal at least
once during the three-day sampling
period, and all take pride in their
customizations of the stereotypically
plain, boring food. Granola, nut butter,
fruits, and infinite other toppings are
added to the oatmeal in different
variations each day, and recorded in
painstaking detail. The blogs rarely
include precise food measurements (even
when the bloggers publish “recipes” they
usually encourage approximation) but
one’s morning oatmeal is frequently
described in depth:
2.a.ii. “Bad food”
By frequently including “good”
foods in their daily meals and snacks,
and by commenting on the virtues of
these food products, the bloggers
positively enact their abstract beliefs
through ritual. According to the About
Me data, “bad,” indulgent foods are also
permitted for the bloggers’ - but only in
specific
manner
that
maintains
acceptable balance. In the same way that
eating and narrating a nutritious meal of
“good” food is ritual for these bloggers,
so is a day of eating that includes the
“bad.” After all, the bloggers clearly
communicate
that
they
oppose
restriction and dieting, and that “it will
• 1/3 cup oats, milk, water, pinch
salt
• 1 small banana
• 1 tbsp raisins
• ½ tbsp chia seeds
• 1/3 cup cooked wheatberries,
stirred in at end
• Splash vanilla
• Sprinkle Mix My Granola
• A few toasty almonds
• Packet Barney Butter that didn’t
come out of the packet as nicely as I
had envisioned21
20
22
Carrots ‘n’ Cake. 10 October 2009. Errands to
the Max.
21
Kath Eats Real Food. 6 October 2009. WBO.
Kath Eats Real Food. 19 February 2008. Tribute
to Oatmeal.
23
Ibid.
254
all even out in the end.”24 Just as the
bloggers’ comments and descriptions of
“healthy” food manifest their system of
beliefs, the discourse surrounding
indulgence is equally part of the process
of “attain(ing] their faith as they portray
it” (Geertz 1993:114).
of the 8 blogs in this sample utilized the
emoticon in conjunction with a “bad”
eating habit. I came to interpret this
playful, end-of-sentence smiley face to
translate in layman’s terms to: “I know I
really shouldn’t be doing this, but…I am.”
The emoticon followed confessions of an
extra sip of wine, a free cupcake sample,
a few dips into the bowl of candy corn, a
square of chocolate during class, and
numerous other “indulgences.” Smiley
faces were also used in conjunction with
condemned practices regarding portion
control: “I didn’t even bother with a
plate! ;)”28 The emoticon in itself truly
functions as a micro-level representation
of the bloggers’ ritual. It is a “model of
belief” because it signifies a “bad”
behavior to others. Simultaneously, the
smiley allows the blogger to actively
document a transgression, contemplate
it in her striving for balance, and
continue on the path toward “balance”
and “health.” It is an enactment of the
religious process – a “model for belief.”
In some cases, the bloggers
overtly address the fact that a food is
“unhealthy” or “bad.” Regarding a
restaurant’s jalapeño poppers, Danica
states, “I am pretty sure there isn’t
anything healthy about them, but,
sometimes you just have to go for it!”25
This acknowledgement, though, is
usually accompanied by some statement
of justification for eating the “bad” food,
even if the treat seems to be innocuous
and to fit in with an otherwise healthy
diet. An after-dinner shmear of Nutella is
okay because its vehicle is whole-grain
toast26; frozen yogurt with sprinkles is
allowed because it is only available at a
certain faraway shop – “when in Rome,
right!?”27 This explicit admittance of
“bad” eating clarifies the line between
“good” and “bad” foods and denotes
accepted reasons for indulgence in the
“bad.”
2.a.iii. The Best of Both Worlds
It is more common, however, for
the bloggers to narrate their indulgences
or
other
deplorable
food-related
behaviors in more subtle ways. The
unexpected player in this enactment of
“bad” is the emoticon: the online version
of the smiley face, formed by typing a
colon adjacent to a close parenthesis :). 6
The foods of highest veneration
on these blogs, though, are at the
convergence of the “good” and “bad”
extremes. There is constant comparison
of “healthy” foods to their indulgent
counterparts. A “good” food that tastes
like something more decadent and “bad”
is valued and praised: squash roasted in
the oven “tastes just like ‘fries’ this
way”29; “These bagels seriously reminded
me of croissants”30; a granola bar
24
28
Eat Like Me. 1 October 2007. Let’s Take a Step
Back and Review.
25
Danica’s Daily. 17 October 2009. Rolly Polly.
26
Healthy and Sane. 6 October 2009. Good Girl.
27
K’s Good Eats. 13 October 2009. Welcome Back.
Carrots ‘n’ Cake. 12 October 2009. Not a Happy
Camper.
29
Danica’s Daily. 17 October 2009. Rolly Polly.
30
Krist’as Kravings. 6 October 2009. Dinner
Pancakes.
255
“reminded me of a peanut butter
chocolate chip oatmeal cookie.”31 Food
products that are nutritionally virtuous
and embody the taste qualities of
indulgence are highly appreciated.
Oatmeal often verges on the “bad,”
which explains its esteemed status: with
apples and vanilla yogurt, it transforms
into apple crisp; with peanut butter and
chocolate chips, a Reese’s cup; with
banana and strawberry, a banana split.
These foods are the most significant in
the bloggers’ diets because they
singlehandedly
facilitate
the
maintenance of balance between good
and bad: they are nutritious and
“healthy” while simultaneously satisfying
taste cravings for the “bad.”
control their own health; and the idea
that food habits affect emotional, mental,
and physical well-being. These ideas
frame and explain the intense desire to
maintain a balanced diet. Because the
balance between “good” and “bad”
influences nearly every aspect of the
blogger’s life, serious control of the self
and of outside situations is critical to
consistently reach this balance. This is
the essential underpinning of the
bloggers’ conception of health, and it is
demonstrated and enacted through their
“ritual.”
2.b.i. Agency and Control
Eating, for the bloggers, is in some
respects an extremely internalized and
personal experience. Although they
frequently share meals with spouses or
friends and subsequently publish them
online, the bloggers demonstrate that
making “correct” food choices is possible
in any context, so long as they are in
control and responsible for their own
desires and actions. There is a collective
assertion among the bloggers that eating
should only occur when one is truly
hungry; eating out of boredom or
sadness happens occasionally on the
blogs,
but
these
behaviors
are
acknowledged and chided more seriously
than with the playful emoticon: “And I
wish this was it. I was full and satisfied.
Did I stop eating? Nope.”32 Consequently,
satiety is a ubiquitous topic of discussion;
each of the 8 bloggers in this sample
discussed their “fullness” at least once a
day (usually much more frequently)
within the three days of post analysis.
A collective goal of these women
is to strike balance between the realms of
“good” and “bad” food. Accordingly, their
ritual practice consists of the physical act
of eating, but it also comprises
monitoring, discussion, and commentary
concerning which foods to extol and
which to avoid. The bloggers’ ritual
practices mark the separation between
“good” and “bad” foods and demonstrate
their navigation of these extremes. In
posting their habits and thoughts online,
the bloggers manifest their beliefs for
others (a model of religion), and to
progress through their own journey
towards the higher goals of their religion
(a model for religion).
2.b. Balance of Life in General
Two further tenets of the bloggers’
food-centered
belief
system
are
represented in their daily rituals: the
power and agency of the individual to
32
31
Healthy and Sane. 13 October 2009. My Head is
Getting to Me.
K’s Good Eats. 9 October 2009. Replay.
256
These bloggers “listen” to their hunger
levels – they analyze them closely, and
are aware of the best practices to “solve”
a certain intensity of hunger or to
maintain satiety over long periods of
time. Specific combinations of oatmeal
and toppings, for example, are praised as
“delicious and tummy filling”33 and “the
perfect pre-run fuel.”34 One afternoon,
Kath states that, “I can’t believe I wasn’t
hungrier (at 2:00PM] but whenever I
have coffee mid-morning I’m less hungry
for lunch.”35 This constant awareness and
monitoring of hunger levels often
informs eating decisions. If breakfast was
surprisingly filling, the blogger chooses a
smaller lunch; if preparing dinner will
take half an hour, the blogger eats a
snack that she knows will perform well in
the given situation. Asserting control
begins for the bloggers with a sharp
cognizance of their own bodies, and
continues with the invention of the
perfect “antidote” to hunger. The
bloggers state their firm belief in the
power of the responsible individual to
make positive food-related changes in
the About Me section, and their
preoccupation with satiety is a
fundamental enactment of this belief.
a considerable amount of effort to plan
ahead. Meals are often packed from
home to avoid inadequate food offerings
in offices, airports, school cafeterias, or
on the road. Kath devotes an entire post
(13 photos) to document a day’s lunch,
snacks, and dinner, packed on her
kitchen counter for a lengthy evening of
nutrition courses: “a good example of me
eating on the go!”36 When the bloggers’
food intentions are thwarted, as Tina’s
were during an unexpected 10-hour visit
”
to the dentist, they are “cranky!
37
(picture is an angry devil-emoticon).
These practices of self-surveillance and
extensive planning not only manifest and
enact the bloggers’ system of beliefs; they
also serve as a means to assert control in
an overwhelming, chaotic world. In this
way, the bloggers’ rituals function
similarly to those of the suburban
holistic health groups of McGuire’s
(1988) research. These groups practiced
yoga, meditation, and alternative
medicine during times of illness to
exhibit their personal power and control
during
frightening
circumstances
(McGuire 1988:229). The bloggers’
consistent health-related habits and
practices are similarly empowering and
grounding.
Beyond individual, bodily cues for
eating, the bloggers’ food habits are also
structured around external factors of life:
work, children, spouses, and other
obligations. In order to skirt these
impediments to eating how they want
and when they want, the bloggers exhibit
2.b.ii. Mental and Emotional Impact of
Eating
Food clearly structures the lives of
the bloggers: the narration of their days
is based on the retelling of meals; they
devote extensive energy and effort to
33
Running With a Recipe. 7 October 2009. Not a
Day Goes By.
34
NYCollegEats. 7 October 2009. Greek Yogurt
Gods.
35
Kath Eats Real Food. 13 October 2009. Fun Run
+ Sun.
36
Kath Eats Real Food. 7 October 2009. Prepped,
Packed, Cleaned.
37
Carrots ‘n’ Cake. 12 October 2009. Not a Happy
Camper.
257
fulfilling their belief of “healthy” eating;
most importantly, they take the time to
broadcast these habits and feelings in an
online public forum. The ritual practice
of blogging attests to the bloggers’ most
intangible food belief that “good” eating
brings about a good, happy, fulfilling life
– and proves the eater to be inherently
good herself.
Tina, consuming too many beers and
tailgate snacks taints her memory of an
otherwise enjoyable day with friends.39
Danica feels, “like a rolly polly – so full
and sorta greased out” after eating rich
steaks and dessert at a restaurant.40 Elina
perhaps expresses these feelings most
overtly:
I’ve been feeling kind of down
for the past few days because of
my snacking, so today when the
urge to buy a cookie hit, I
ignored it. Typically I’m a
proponent of actually satisfying
my cravings…but when it gets
out of control, I need to ‘detox’ a
bit with cleaner eats and more
nutritious foods. It’s all about
balance! As you’ll see below, I
wasn’t eating carrot sticks all day
– my food was still satisfying,
41
mentally and physically.
In analyzing the textual data, it is
clear that food plays a significant role in
the bloggers’ feeling of overall wellbeing. Food has the power to alter
moods: oatmeal is rejuvenating on a cold
morning, and a new package of Greek
yogurt can brighten an otherwise
“dismal” day38. Conversely, feeling like
one has overindulged or made the wrong
choices can be depressing and even
debilitating. Just as the characteristics of
the warm, comforting, wholesome bowl
of oatmeal seem transferred to the
blogger upon consumption, the same can
be said for “bad” foods. The effect of
“bad” food is perhaps even heightened
for the bloggers because (as discussed in
section 2.a.) they have such an acute
awareness of what ought to be eaten in
moderation. Additionally (as discussed in
section 2.b.i.) the bloggers feel an
enormous
sense
of
personal
accountability for their food-related
choices. Overindulgence does not simply
hold physical implications of weight
gain; a food-related falter is a failure of
the blogger to assert her control and
fulfill her personal obligation to be
healthy. The blogger has transgressed a
tenet of her religion, and the discourse in
the posts of this analysis proves this. For
Food truly affects all aspects of
these bloggers’ lives. They expressed this
sentiment hypothetically in the About
Me sections of the blogs, and they
embody
it
through
their
daily
eating/blogging ritual actions. The
incredible power that the bloggers have
ascribed to food - as a means to
happiness, fulfillment, and personal
success and worth – impels their
devotion to balance the “good” and “bad”
through self-control. The practices are so
reliably and meticulously carried out
because the bloggers believe them to be
inextricably linked to something larger.
Like so many religious followers, the
bloggers truly have faith in the greater
ramifications of their everyday actions. It
39
Carrots ‘n’ Cake. 11 October 2009. Chili +
Football.
40
Danica’s Daily. 17 October 2009. Rolly Polly.
41
Healthy and Sane. 6 October 2009. Good Girl.
38
NYCollegEats. 7 October 2009. Greek Yogurt
Gods.
258
is this ethereal relationship between
practice and belief that proves this group
as religious.
these women’s religion could not be
publicly shared or perpetuated. The
religion would cease to exist, as would
the structuring rituals of the bloggers’
everyday lives. The ability to blog has
undoubtedly impacted these women’s
experience of eating, a fundamental
practice of physical and social life.
Concurrently, their blogging allows for
an unprecedented opportunity to explore
the nuances of food-related behavior and
philosophy.
Conclusion
When first introduced to these
blogs, some readers might be taken
aback: the diligent dietary updates,
extensive photography, and scrutinizing
self-commentary and policing appear to
be compulsive and bizarre behaviors.
Why do these people think we care about
what they eat? Where do they find the
time to do all this? Most importantly,
why are they so obsessed with food? In
many ways, the new reader is correctly
appraising the situation; as the blogs
posts show, these women demonstrate a
visibly dogged devotion to specific eating
habits. However, textual analysis of the
blogs illustrates the logical reasoning
behind these practices: the bloggers
steadfastly perform their “rituals”
because they sincerely believe the
practices will lead to the venerated
condition of “health.” Being healthy
connotes
personal
achievement,
righteousness, and value of the self.
Understanding
this
overarching,
collective belief shows the bloggers’ daily
practices (extensive tinkering with
oatmeal; pursuit of organic produce; or a
guilt-infused emoticon, for example) as
completely valid within their ideological,
religious context. The bloggers’ faith in
the posited outcome of their rituals,
coupled with their devotion to maintain
the practices necessary to the definitive
goodness of health, proves them to
indeed be religious through the
theoretical
definitions
given
by
Durkheim and Geertz. Additionally,
without the online medium of blogging,
Discussion
Because of the demographic
make-up of their authors, these blogs
could easily be dismissed as selfindulgent excuses for privileged women
to flaunt their food preferences. But to
combat this interpretation, I return to
the scenario presented at the opening of
this paper involving the midnight Peanut
M&Ms in the library. Although that
situation is dramatized, the daily
decisions we make concerning food are
complex and often daunting. In the
United States, consumers are constantly
bombarded with the cultural messages of
advertisement encouraging them to eat a
certain way; with promises to “lose
weight quick” endorsed by rail-thin or
muscle-man celebrities; with the notion
that eating “how you’re supposed to”
produces a good person, while everyone
else is a fat, irresponsible failure. The
bloggers’
monitoring
and
close
documentation of their food intake,
then, is not unreasonable. If food choice
is accepted as indicative of one’s moral,
personal worth and identity, it deserves
the time, energy, attention, and care that
the bloggers demonstrate. And the
internalization of the deep moral
259
Bafflement, suffering, and a
sense of intractable ethical
paradox
are
all…radical
challenges to the proposition
that life is comprehensible and
that we can…orient ourselves
effectively within it – challenges
which any religion, however
“primitive,” which hopes to
persist must attempt somehow
to cope. (100)
implications of food consumption does
not simply affect the women of these
blogs. After all, “you are what you eat” is
a ubiquitous adage throughout the
United States, and that rhetoric is used
throughout
public
and
personal
discourse.
In his recent New York Times
article “Rules to Eat By,” acclaimed food
writer Michael Pollan (2009) addressed
the question plaguing the bloggers and
so many Americans: “If we can’t rely on
the marketers or the government or even
the nutritionists to guide us through the
supermarket woods, then who can we
rely on?” Continually escalating national
obesity rates prove that current societal
food messages are far from curative
(CDC 2009). Pollan (2009) suggests that
navigating the treacherous task of eating
right mandates a reliance on others we
know
and
trust:
“an
informal,
unsanctioned way of negotiating our
eating lives that becomes indispensable
at a time when official modes of talking
about food have suffered a serious loss of
credibility.” I would argue that this is the
function of the healthy-living blogs. In a
food climate where the stakes seem high
(“Do you even know what non-organic
produce does to your/your child’s/your
spouse’s body?”) and conflicting paths to
“health” battle for our attention and our
dollars, the bloggers hold fast to their
religious beliefs and practices, using the
internet as their essential tool. Geertz
(1993) further explains a purpose of
religion in his essay Religion as a Cultural
System:
Religion is a source of solace and
order in the face of chaos and
uncertainty.
Through
the
public,
collective assertion of their food-related
beliefs, and unwavering practices that
enact and embody these values, the
bloggers rally around their religion of
health. In making the private negotiation
of eating public and communal, the
bloggers answer Pollan’s (2009) call for
“informal” eating commandments. The
blogs of this study provide a model for
this necessary food-based support
network. Through their religion of
health, the bloggers effectively maintain
their beliefs through communal, online
assertion and encouragement of accepted
practices. The fundamental traits of their
community could inform nutrition
initiatives in any social circumstance.
These healthy-living blogs demonstrate
the deep entrenchment of the internet in
everyday life and, concurrently, its power
to religiously unite any community to
declare their beliefs, give meaning to
their experiences, and attain their goals –
even if that goal is simply to resist the
beckoning glow of the vending machine.
260
Works Cited
Geertz, C. (1993). Chapter 4: Religion as a
Cultural System. In The Interpretation of
Cultures (pp. 87-125). New York: Fontana
Press.
Anderson, E.N. (2005). Everyone Eats:
Understanding Food and Culture. New
York: New York University Press.
Giddens, A. (Ed., Trans.). (1972). Emile
Durkheim: Selected Writings. Cambridge,
England: Cambridge University Press.
Atkinson, P. (1983). Eating Virtue. In A.
Murcott (Ed.), The Sociology of Food and
Eating (pp. 9-17). Hampshire, England:
Gower Publishing Company Limited.
Goffman, E. (1956). The Presentation of
Self in Everyday Life. New York:
Doubleday.
Bell, B. (2007). Private Writing in Public
Spaces: Girls’ Blogs and Shifting
Boundaries. In S. Weber and S. Dixon,
(Ed.s), Growing Up Online: Young People
and Digital Technologies (pp. 95-112). New
York: Palgrave MacMillan.
Griffin, J., & Berry, E.M. (2003). A Modern
Day Holy Anorexia? Religious Language in
Advertising and Anorexia Nervosa in the
West. European Journal of Clinical
Nutrition, 57, 43 – 51. Retrieved from
Academic Search Premier.
Bourdieu, P. (1984). Distinction: A Social
Critique of the Judgment of Taste (R. Nice,
Trans.). Cambridge, MA: Harvard
University Press. (Original work published
1979).
Gusfield, J.R., & Michalowicz, J. (1984).
Secular Symbolism: Studies of Ritual,
Ceremony, and the Symbolic Order in
Modern Life. Annual Review of Sociology,
10, 417 – 435. Retrieved from JSTOR.
Day, K. & Keys, T. (2008). Starving in
Cyberspace: a Discourse Analysis of ProEating-Disorder Websites. Journal of
Gender Studies, 17 (1), 1 – 15. Retrieved
from Academic Search Premier.
Harp, D. & Tremayne, M. (2006). The
Gendered Blogosphere: Examining
Inequality Using Network and Feminist
Theory. Journalism & Mass
Communication Quarterly, 83 (2), 247 –
264. Retrieved from ProQuest Platinum.
Franzosi, R. (1998). Narrative Analysis – Or
Why (And How) Sociologists Should be
Interested in Narrative. Annual Review of
Sociology, 24, 517 – 554. Retrieved from
Academic Search Premier.
Hevem, V.W. (2004). Threaded Identity in
Cyberspace: Weblogs & Positioning in the
Dialogical Self. Identity, 4 (4), 321 – 335.
Retrieved from Academic Search Premier.
Frost, N. (2009). ‘Do You Know What I
Mean?’: the Use of a Pluralistic Narrative
Analysis Approach in the Interpretation of
an Interview. Qualitative Research, 9 (9), 9
– 29. Retrieved from SAGE Publications.
Hsu, C. L. (2001). Political Narratives and
the Production of Legitimacy: The Case of
Corruption in Post-Mao China. Qualitative
Sociology, 24 (1), 25 – 54.
Geertz, C. (1973). Chapter 1: Thick
Description: Toward an Interpretive Theory
of Culture. In The Interpretation of Cultures
(pp. 3-30). New York: Basic Books.
Keats, P.A. (2009). Multiple Text Analysis
in Narrative Research: Visual, Written, and
Spoken Stories of Experience. Qualitative
Research, 9 (2), 181 – 195. Retrieved from
SAGE Premier 2009.
261
Loft, J. (2005). Understanding Community
Weblogs. Dissertation Abstracts
International, A: The Humanities and Social
Sciences, 65 (7), 2790. Retrieved from
Sociological Abstracts.
Rettberg, J. W. (2008). Blogging: Digital
Media and Society Series, Malden, MA,
Polity Press.
Seaton, E.E. (2008). Common Knowledge:
Reflections on Narratives in Community.
Qualitative Research, 8 (3), 293 – 305.
Retrieved from SAGE Premier 2009.
Madden, H., & Chamberlain, K. (2004).
Nutritional Health Messages in Women’s
Magazines: A Conflicted Space for
Women’s Readers. Journal of Health
Psychology, 9(4), 583 – 597. Retrieved from
SAGE Premier 2009.
Somers, M.R. & Gibson, G.D. (1994).
Reclaiming the Epistemological ‘Other’:
Narrative and the Social Construction of
Identity. In Social Theory and the Politics of
Identity, ed. Craig Calhoun, Cambridge,
MA: Blackwell.
Malson, H. (2009). Doing Weight: Pro-ana
and Recovery Identities in Cyberspace.
Journal of Community & Applied
Psychology, 19 (5), 348 – 359. Retrieved
from Wiley InterScience Online.
Stern, S. R. (2007). Adolescent Girls’
Expression on Web Home Pages: Spirited,
Somber, and Self-Conscious Sites. In S.
Weber and S. Dixon, (Ed.s), Growing Up
Online: Young People and Digital
Technologies (pp. 159-179). New York:
Palgrave MacMillan.
McGuire, M. (1988). Ritual Healing in
Suburban America. New Brunswick, NJ:
Rutgers University Press.
Mintz, S. & DuBois, C. (2002). The
Anthropology of Food and Eating. Annual
Review of Anthropology, 31, 99 – 119.
Retrieved from JSTOR.
Thompson, K. (1982). Emile Durkheim.
New York: Tavistock Publications.
Turner, V. (1974). Dramas, Fields, and
Metaphors: Symbolic Action in Human
Society. Ithaca, NY: Cornell University
Press.
Pietrykowski, B. (2004). You Are What You
Eat: The Social Economy of the Slow Food
Movement. Review of Social Economy, 62
(3), 307 – 321. Retrieved from Business
Source Premier.
Twigg, J. (1983). Vegetarianism and the
Meanings of Meat. In A. Murcott (Ed.), The
Sociology of Food and Eating (pp. 18 - 30).
Hampshire, England: Gower Publishing
Company Limited.
Polak, M. (2007). Girls Creating Identity
and Forming Community in Pro-Ana/Mia
Websites. In S. Weber and S. Dixon, (Ed.s),
Growing Up Online: Young People and
Digital Technologies (pp. 81-94). New
York: Palgrave MacMillan.
Vo-Jutabha, E.D., Khanh, T. D., McHale,
J.P., & Valsiner, J. (2009). A Qualitative
Analysis of Vietnamese Adolescent Identity
Exploration Within and Outside an Ethnic
Enclave. Journal of Youth Adolescence, 38,
672 – 690.
Pollan, M. (2009, October 6). Rules to Eat
By. New York Times. Retrieved December 9,
2009 from http://www.nytimes.com/.
262
Walker, K. (2000). ‘It’s Difficult to Hide It’:
The Presentation of Self on Internet Home
Pages. Qualitative Sociology, 23 (1), 99 –
120. Retrieved from Academic Search
Premier.
Cultural Omnivorousness, Social Distinction
and Dining Out. Sociology, 33 (1), 105 –
127. Retrieved from SAGE Publications.
Wood, R. (1995). Chapter 1: Food and
Social Theory: States of the Art. In The
Sociology of the Meal (pp. 1-45). Edinburgh,
Scotland: Edinburgh University Press.
Warde, A., Martens, L., & Olsen, W. (1999).
Consumption and the Problem of Variety:
Data Sources
Carrots ‘n’ Cake. (2009, October 10).
Errands to the Max. Message posted to
http://carrotsncake.com/2009/10/errand
s-to-the-max.html.
Carrots ‘n’ Cake. Weight Loss. Message
posted to
http://carrotsncake.com/weight-loss.
Danica’s Daily. About. Message posted to
http://danicasdaily.com/about/.
Carrots ‘n’ Cake. (2009, October 11). Chili
+ Football. Message posted to
http://carrotsncake.com/2009/10/chilifootball.html.
Danica’s Daily. (2009, October 12). Role
Reversal. Message posted to
http://danicasdaily.com/role-reversal/.
Carrots ‘n’ Cake. (2009, October 11).
Sweet Sunday Morning. Message posted
to
http://carrotsncake.com/2009/10/sweetsunday-morning.html.
Danica’s Daily. (2009, October 13).
Terrific Tortellini Toss. Message posted
to http://danicasdaily.com/terrifictortellini-toss/.
Carrots ‘n’ Cake. (2009, October 12). First
of the Season. Message posted to
http://carrotsncake.com/2009/10/firstof-the-season.html.
Danica’s Daily. (2009, October 17). Rolly
Polly. Message posted to
http://danicasdaily.com/rolly-polly/.
Carrots ‘n’ Cake. (2009, October 12). Not
a Happy Camper. Message posted to
http://carrotsncake.com/2009/10/not-ahappy-camper.html.
Eat, Drink, & Be Aware. About Me.
Message posted to
http://www.blogger.com/profile/0553626
0830375831583.
Carrots ‘n’ Cake. About. Message posted
to http://carrotsncake.com/about.
Eat Like Me. (2007, October 1). Let’s Take
a Step Back and Review. Message posted
to
http://www.self.com/fooddiet/blogs/eatli
keme/welcome-to-eat-like-me/.
Carrots ‘n’ Cake. My Journey. Message
posted to
http://carrotsncake.com/weightloss/my-journey.
Eat, Live, Run. My Food Philosophy.
Message posted to
263
http://eatliverun.com/my-foodphilosophy/.
Healthy Tipping Point. Tips. Message
posted to
http://www.healthytippingpoint.com/tip
s.
Eating Bender. About. Message posted to
http://eatingbender.com/about/.
Jenn’s Menu and Lifetsyle Blog. About
Jenn. Message posted to
http://www.j3nn.net/about/.
Eating Bird Food. About Me. Message
posted to
http://www.eatingbirdfood.com/about2/.
Kath Eats Real Food. (2008, February 19).
Tribute to Oatmeal. Message posted to
http://www.katheats.com/kaths-tributeto-oatmeal/.
Hangry Pants. Meet Hangry Pants.
Message posted to
http://www.hangrypants.com/meethangry-pants/.
Kath Eats Real Food. (2009, February 3).
Tangerine Sunshine. Message posted to
http://www.katheats.com/?p=15272.
Hangry Pants. Meet Heather. Message
posted to
http://www.hangrypants.com/2008/07/
meet-heather/.
Kath Eats Real Food. (2009, October 6).
WBO. Message posted to
http://www.katheats.com/wbo/.
Healthy and Sane. About Me. Message
posted to
http://healthyandsane.com/about/.
Kath Eats Real Food. (2009, October 6).
A Good Day. Message posted to
http://www.katheats.com/a-good-day/.
Healthy and Sane. (2009, October 6).
Good Girl. Message posted to
http://healthyandsane.com/2009/10/goo
d-girl/.
Kath Eats Real Food. (2009, October 6).
Haircuts + Limericks. Message posted to
http://www.katheats.com/haircutslimericks/.
Healthy and Sane. (2009, October 8). My
First Blog Anniversary. Message posted
to
http://healthyandsane.com/2009/10/myfirst-blog-anniversary/.
Kath Eats Real Food. (2009, October 7).
Decisions, Decisions. Message posted to
http://www.katheats.com/decisionsdecisions/.
Healthy and Sane. (2009, October 13). My
Head is Getting to Me. Message posted
to
http://healthyandsane.com/2009/10/myhead-is-getting-to-me/.
Kath Eats Real Food. (2009, October 7).
Prepped, Packed, Cleaned. Message
posted to
http://www.katheats.com/preppedpacked-cleaned/.
Healthy Tipping Point. Caitlin. Message
posted to
http://www.healthytippingpoint.com/ab
out.
Kath Eats Real Food. (2009, October 7).
A Full Gas Tank. Message posted to
http://www.katheats.com/full-gas-tank/.
264
Kath Eats Real Food. (2009, October 13).
Morning Gossip. Message posted to
http://www.katheats.com/morninggossip/.
http://ksgoodeats.wordpress.com/2009/1
0/09/replay/.
K’s Good Eats. (2009, October 13).
Welcome Back. Message posted to
http://ksgoodeats.wordpress.com/2009/1
0/13/welcome-back/.
Kath Eats Real Food. (2009, October 13).
Fun Run + Sun. Message posted to
http://www.katheats.com/fun-run-sun/.
K’s Good Eats. About Me. Message
posted to
http://ksgoodeats.wordpress.com/aboutme/.
Kath Eats Real Food. (2009, October 13).
Sing Along. Message posted to
http://www.katheats.com/sing-along/.
Kath Eats Real Food. Kath. Message
posted to
http://www.katheats.com/kath-2/.
NYCollegEats. (2009, October 7). Greek
Yogurt Gods. Message posted to
http://collegaleats.blogspot.com/2009/10
/greek-yogurt-gods.html.
Krista’s Kravings. (2009, October 6).
Dinner Pancakes. Message posted to
http://kristaskravings.blogspot.com/200
9/10/dinner-pancakes.html.
NYCollegEats. (2009, October 9). Bio
Exam. Message posted to
http://collegaleats.blogspot.com/2009/10
/bio-exam.html.
Krista’s Kravings. (2009, October 9). Wet
Vacation Day. Message posted to
http://kristaskravings.blogspot.com/200
9/10/wet-vacation-day.html.
NYCollegEats. (2009, October 13). Study
Strategies. Message posted to
http://collegaleats.blogspot.com/2009/10
/study-strategies.html.
Krista’s Kravings. (2009, October 13).
Couch Potato. Message posted to
http://kristaskravings.blogspot.com/200
9/10/couch-potato.html.
NYCollegeEats. Ada. Message posted to
http://www.blogger.com/profile/1597839
0392818633848.
Krista’s Kravings. Krista – About Me.
Message posted to
http://www.blogger.com/profile/1459786
8827300760281.
Run, Eat, Repeat. About. Message posted
to http://runeatrepeat.com/?page_id=2.
Running With a Recipe. (2009, October
7). She Recommended It. Message posted
to
http://runningwitharecipe.blogspot.com/
2009/10/she-recommended-it.html.
K’s Good Eats. (2009, October 8). All
These Things that I’ve Done. Message
posted to
http://ksgoodeats.wordpress.com/2009/1
0/08/all-these-things-that-ive-done/.
Running With a Recipe. (2009, October
7). Not a Day Goes By. Message posted to
http://runningwitharecipe.blogspot.com/
2009/10/not-day-goes-by.html.
K’s Good Eats. (2009, October 9). Replay.
Message posted to
265
Running With a Recipe. (2009, October
9). Versatile Veggie. Message posted to
http://runningwitharecipe.blogspot.com/
2009/10/versatile-veggie.html.
to
http://runningwitharecipe.blogspot.com/
2009/10/warm-by-oven.html.
Running With a Recipe. Why I Do This.
Message posted to
http://runningwitharecipe.blogspot.com/
Running With a Recipe. (2009, October
13). US Wins for Best Porridge. Message
posted to
http://runningwitharecipe.blogspot.com/
2009/10/us-wins-for-best-porridge.html.
SnackFace. About. Message posted to
http://snackface.wordpress.com/about/.
The Chic Life. About. Message posted to
http://thechiclife.typepad.com/about.ht
ml.
Running With a Recipe. (2009, October
13). Warm by the Oven. Message posted
Appendix
1) 20 Blogs Included in “About Me”
Analysis:
http://carrotsncake.com/ (Carrots ‘n’
Cake)
http://www.j3nn.net/ (Jenn’s Menu &
Lifestyle Blog)
http://collegaleats.blogspot.com/ (NY
Collegeats)
http://www.katheats.com (Kath Eats
Real Food)
http://danicasdaily.com/ (Danica’s Daily)
http://eatdrinkandbeaware.blogspot.com
/ (Eat, Drink, & Be Aware)
http://kristaskravings.blogspot.com/
(Krista’s Kravings)
http://ksgoodeats.wordpress.com/ (K’s
Good Eats)
http://eatingbender.com/ (Eating
Bender)
http://runeatrepeat.com/ (Run, Eat,
Repeat)
http://runningwitharecipe.blogspot.com/
(Running With a Recipe)
http://eatingbirdfood.com/ (Eating Bird
Food)
http://eatliverun.com/ (Eat, Live, Run)
http://graduatemeghann.com/ (The
Inner Workings of a College Graduate)
http://www.self.com/fooddiet/blogs/eatli
keme (Eat Like Me)
http://hangrypants.com/ (HangryPants)
http://snackface.wordpress.com/
(SnackFace)
http:healthyandsane.com/ (Healthy and
Sane)
http://thechiclife.typepad.com/ (The
Chic Life)
http://healthytippingpoint.com/
(Healthy Tipping Point)
266
2) 15 Blogs Omitted from Initial List for
Analysis:
http://www.melissanibbles.com/
(Melissa Nibbles)
http://actorsdiet.blogspot.com/ (The
Actors’ Diet)
http://onalobsterplacemat.com/ (On a
Lobster Placemat)
http://applecrumbles.com/ (Apple
Crumbles)
http://pastriesandbacon.wordpress.com/
(Pastries & Bacon)
http://blueberryhil.wordpress.com/
(Blueberry Hill)
3) 8 Blogs Included in Daily Post
Analysis:
http://burpandslurp.wordpress.com/
(Burp and Slurp)
http://carrotsncake.com/ (Carrots ‘n’
Cake)
http://caretoeat.net/ (Care To Eat)
http://collegaleats.blogspot.com/ (NY
Collegeats)
http://foodmakesfunfuel.com/ (Food
Makes Fun Fuel)
http://danicasdaily.com/ (Danica’s Daily)
http://happyherbivore.com/about/
(Happy Herbivore)
http:healthyandsane.com/ (Healthy and
Sane)
http://healthyashley.blogspot.com/
(Healthy Ashley)
http://heathereatsalmondbutter.com/
(Heather Eats Almond Butter)
http://www.katheats.com (Kath Eats
Real Food)
http://kristaskravings.blogspot.com/
(Krista’s Kravings)
http://www.honormyhealth.com/
(Honor My Health)
http://ksgoodeats.wordpress.com/ (K’s
Good Eats)
http://lengslog.wordpress.com/ (A Fat
Cat and a Vegetarian)
http://runningwitharecipe.blogspot.com/
(Running With a Recipe)
http://luckytastebuds.wordpress.com/
(Lucky Taste Buds)
267
268
269
270
271