Read the article - cohpa - University of Central Florida

Transcription

Law Enforcement Executive
Vol. 15, No. 1 • March 2015
Law Enforcement Executive Forum
Editor-in-Chief
Thomas J. Jurkanin, PhD
Senior Editors
Vladimir A. Sergevnin, PhD
Susan C. Nichols, MS Ed.
Joanne Kurt-Hilditch, PhD
Associate Editors
Jennifer M. Allen, PhD
Department of Criminal Justice, University of Northern Georgia
Barry Anderson, JD
Professor, School of Law Enforcement and Justice Administration, Western Illinois University
Tony A. Barringer, EdD
Division of Justice Studies, Florida Gulf Coast University
Michael J. Bolton, PhD
Chair, Department of Criminal Justice and Sociology, Marymount University
Dennis Bowman, PhD
Professor, School of Law Enforcement and Justice Administration, Western Illinois University
Becky K. da Cruz
Criminal Justice and Law and Society, Armstrong Atlantic State University
Jose de Arimateia da Cruz
Political Science and Comparative Politics, Armstrong Atlantic State University
Larry Hoover, PhD
Director, Police Research Center, Sam Houston State University
Steven M. Hougland, PhD
Criminal Justice, Bainbridge State College
William Lewinski, PhD
Director, Force Science Research Center, Minnesota State University
Hyeyoung Lim, PhD
Department of Justice Sciences, University of Alabama at Birmingham
Stephen A. Morreale, DPA
School of Public Policy and Administration/Criminal Justice, Walden University
Gregory Boyce Morrison
Department of Criminal Justice and Criminology, Ball State University
Deborah W. Newman, JD, EdD
Professor and Chair, Department of Criminal Justice, Middle Tennessee State University
Michael J. Palmiotto
Professor, Criminal Justice, Wichita State University
Wayne Schmidt, LL.M., JD
Director, Americans for Effective Law Enforcement
Aaron Thompson, PhD
Department of Sociology, Eastern Kentucky University
Qing Tian, PhD
Illinois Law Enforcement Training and Standards Board Executive Institute
Brian N. Williams, PhD
Department of Public Administration & Policy, School of Public & International Affairs,
University of Georgia
The Law Enforcement Executive Forum is published electronically four times per year by the Illinois
Law Enforcement Training and Standards Board Executive Institute (ILETSBEI) located at Western
Illinois University in Macomb, Illinois.
The Law Enforcement Executive Forum is a peer-reviewed publication intended to present articles
relevant to today’s law enforcement leadership environment; it provides the criminal justice
community with best practices and emerging technology for law enforcement leaders at all levels.
The Forum is written for and by criminal justice professionals and scholars. The Forum is an
environment for criminal justice professionals and scholars to share new ideas, updated research
and information, and success with others.
Copyright
© Illinois Law Enforcement Training and Standards Board Executive Institute. All rights reserved.
No part of this publication may be reproduced, stored, or transmitted in any form or by any
means without the prior permission in writing from the copyright holder. Special requests should
be addressed to ILETSBEI at [email protected].
Subscription
To subscribe, please visit the journal website: http://iletsbeiforumjournal.com/index.php/
subscribe. You can also subscribe via email at [email protected] or via phone at (309) 298-2646.
Back Issues
Back issues and individual articles are available for purchase at the journal website:
http://iletsbeiforumjournal.com.
Other publications
Monographs and research reports published by the Illinois Law Enforcement Training
and Standards Board Executive Institute are available for free download or purchase at
www.iletsbei.com/publications.php.
ISSN: 1552-9908
Disclaimer
The publisher, Illinois Law Enforcement Training and Standard Board Executive Institute, and
the editors cannot be held responsible for errors or any consequences arising from the use of
information contained in this journal. The views and opinions expressed do not necessarily
reflect those of the publisher and the editors. Please address questions about individual articles
to their respective author(s).
Submission Guidelines
General Information
The acceptance criteria for all articles are the
quality and originality of the research and its
significance to our readership. The subject
matter must provide a clear and consistent
message related to law enforcement issues
and their relation to those in leadership roles,
whether formal or informal.
Manuscripts are considered provided they
have not been published previously or concurrently submitted for publication elsewhere
and its publication has been approved by all
co-authors, if any, as well as by the responsible authorities. In the case that a manuscript
is found to have already been published, the
manuscript will be denied and the author(s)
of that work will be banned from submitting
manuscripts to the Forum in the future.
Authorship must be limited to those who have
contributed substantially to the work. All collaborators should have in place an appropriate process for reviewing the accuracy of the
reported results. The publisher will not be
held legally responsible should there be any
claims for compensation.
Book reviews and research notes will be considered for publication. No submission will
be published until recommended by referees,
who will review blind copies.
Copyright
Authors are required to secure permission
for the reproduction of any figure, table, or
extensive (more than 50 words) extract from
the text of a source that is copyrighted or
owned by a party other than the Forum or the
author. This applies to direct reproduction as
well as “derivative reproduction,” wherein
the contributor has created a new figure
or table which derives substantially from
a copyrighted source. Contributors should
write to both the publisher and author of
such material requesting nonexclusive world
rights in all languages for use in the article
and in all future editions of it.
Article Format
Manuscripts should be in English, typed double-spaced, and should not exceed 25 pages,
inclusive of the narrative, tables, references,
appendices, and figures. Pages of each copy
must be sequentially numbered.
The journal follows the Publication Manual
of the American Psychological Association
(6th ed., 2009) for article format and citations.
Please refer to the References sections in past
issues for examples. Webster’s Third New
International Dictionary (3rd ed., 1993) is the
standard reference for spelling.
Manuscripts should be compiled in the following order: title page, abstract, key words,
main text, acknowledgments, references,
appendices (as appropriate), table(s) with
caption(s) (on separate pages), and figure(s)
with caption(s) (on separate pages). Material
which is not essential to the continuity of the
text (e.g., proofs, derivations, or calculations)
should be placed in Appendices.
Two separate electronic files will be needed to
complete the submission:
1. The title page – The title page should
include a running head (no more than 35
characters), the manuscript title, and the
author’s name and contact information
(including e-mail, phone number, and
mailing address). A brief biographical
statement describing each author’s current affiliation and rank should also be
included in the title page. In addition, one
author should be designated as the corresponding author on the title page.
2. The main document – The main document
should not contain any identifying information of the author(s). The first page of
the main document must be a blinded
title page, including the title only. The
second page should include an abstract of
no more than 200 words. Authors are also
asked to provide up to six key words or
phrases that characterize the contents of
the paper. These will be used for indexing and data retrieval purposes. The main
text of the manuscript should start from
the third page of the main document.
Online Submission
Manuscripts should be submitted electronically via Scholastica (https://scholasticahq.
com/law-enforcement-executive-forum/for_
authors). Authors should click “Submit now”
on the “For Authors” page to submit a manuscript. Authors who are not current users of
Scholastica will be required to fill out a simple
form to create a new account.
On the submission page, authors should
upload the main document as the “Primary
Manuscript File” and submit the title page
as well as other supporting documents as
“Supporting Files.” The acceptable file formats for the main document include doc,
docx, and pdf. Authors will receive an e-mail
confirmation once a manuscript is successfully submitted.
All subsequent correspondence, including
notification of decisions reached by editors or
reviewers, takes place via e-mail. Authors will
also be able to track the manuscript submitted
and view editors or reviewers’ responses in
their Scholastica account. For more detailed
submission instructions, please refer to
http://help.scholasticahq.com/customer/
portal/articles/1218626.
Table of Contents
Vol. 15 ▪ No. 1 ▪ 2015
Editorial ............................................................................................................................................... i
Thomas J. Jurkanin
Ambushes Leading Cause of Officer Fatalities – When Every Second Counts:
Analysis of Officer Movement from Trained Ready Tactical Positions ............................... 1
William J. Lewinski
Jennifer L. Dysterheft
Jacob M. Bushey
Nathan D. Dicks
Women and SWAT: Making Entry into Police Tactical Teams .............................................. 16
Thorvald O. Dahle
Keeping Kids Out of Corrections: Lowering Recidivism by Strengthening Teamwork
and Collaboration Between Law Enforcement Officers and Transition Coordinators
in Juvenile Correctional Facilities .............................................................................................. 29
Theresa A. Ochoa
Lawrence J. Levy
Kelly M. Spegel
Yanua F. Ovares
Factoring Fatigue into Police Deadly Force Encounters: Decision-Making and
Reaction Times ............................................................................................................................... 44
David M. Blake
Edward Cumella
Criminal Justice Practitioner Attitudes Toward Risk Assessments on Response to
Domestic Violence ........................................................................................................................ 66
Lee E. Ross
Detection and Prevention of Racial Profiling Practices: Case Study of a Medium-Sized
City in Texas ................................................................................................................................... 79
Won-Jae Lee
Shawn S. Morrow
Seungmug (Zech) Lee
Editorial
On December 18, 2014, President Obama
approved an Executive Order establishing
the Task Force on 21st Century Policing. The
Executive Order was responsive to growing
public concern regarding incidents during
which police officers shot and killed young
black men during confrontations on the
streets. The Task Force was composed of individuals representing the broad interests of
law enforcement, community leaders, advocacy groups, and experts on race relations
and civil rights. The Task Force was given a
90-day window to complete their work and
to issue a report of their findings and recommendations. The draft report was released in
March 2015.
Policing represent more of a refocusing on
things we already knew but from which we
may have drifted. The Task Force recommendations ask us to refocus on such important
areas as building partnerships/collaboration, honoring the public trust, ethics, policy
development and review, supervision/oversight/accountability, research, and education
and training. I am reminded of the familiar
expression, “What is old is new again.”
Given the short timeline allotted for consideration of this matter, the Task Force followed
an expedited, methodical process for gathering research and commentary from numerous
subject-matter experts, community leaders,
and concerned citizens. The Task Force held
seven “listening sessions” before deliberating
and summarizing their recommendations.
In the 76-page “Draft” report, the Task Force
structured their recommendations around
six, themed “Pillars”: Pillar One: Building
Trust & Legitimacy, Pillar Two: Policy and
Oversight, Pillar Three: Technology & Social
Media, Pillar Four: Community Policing &
Crime Reduction, Pillar Five: Training &
Education (for the police), and Pillar Six:
Officer Wellness & Safety. The final section
dealt with implementation. In total, there
were 62 recommendations. Concomitant to
each recommendation, the Task Force provided an outline of Action Items.
First, the Task Force provided valuable
insight to emphasize the point that matters of
crime and police response are symptomatic of
larger issues involving inequalities, blocked
opportunities, endemic social problems, and
racial tensions within American society. The
fact that President Obama’s Executive Order
is entitled, “Task Force on 21st Century
Policing” might suggest that the police are the
only government institution that is in need
of improvement. The Task Force recognized
the limited scope of their charge and recommends that “The President should promote
programs that take a comprehensive look at
community-based initiatives that address
the core issues of poverty, education, health
and safety.” In addition, I would emphasize
the need to examine police shootings, citizen shootings, and incidents of mass killings
related to failed programs and services provided to persons suffering from mental illness
and their interconnections.
It appears that the Task Force did the work
required of them. Like many previous
national task force reports issued related to
law enforcement, crime, and justice, the recommendations presented in 21st Century
Second, it must be acknowledged that the
world is seen through different frames of
reference, depending upon one’s role in society. The Task Force developed their recommendations from a professional perspective,
As we await the final printing of the Task Force
on 21st Century Policing, it seems to me that
three important points must not be missed if
the recommendations are going to have an
impact.
Law Enforcement Executive Forum • 2015 • 15(1)
i
emphasizing structural and organizational
reform. The report is now placed in the hands
of politicians, whose frame of reference may
be on what is politically expedient rather than
what is required/needed.
Third, the federal government must be willing to invest the necessary financial resources
to ensure that the Task Force report does not
end up on the desks of policymakers and law
enforcement officials collecting dust. Time
is of the essence if we truly seek to improve
public safety while concurrently protecting
the civil rights of our citizenry.
Thomas J. Jurkanin, PhD
Editor-in-Chief
ii
Ambushes Leading Cause of Officer Fatalities –
When Every Second Counts: Analysis of Officer
Movement from Trained Ready Tactical Positions
William J. Lewinski, PhD, Force Science Institute
Jennifer L. Dysterheft, MS, Force Science Institute
Jacob M. Bushey, Minnesota State University, Mankato
Nathan D. Dicks, MS, Minnesota State University, Mankato
Abstract
Recently, the threat of ambush assaults to police officers has dangerously increased. These assaults can
occur very rapidly, and to be better prepared to respond, it is important to understand the speed of officer
responses and any advantages officers may gain from various tactical techniques. Therefore, the purpose of
this study was to understand and examine officer movement times from various finger-indexing positions
as well as the speed at which officers can fire their weapons from various starting positions. In the first
experiment, officers (n = 52) fired their weapons from four trained finger-index positions to measure their
time to fire. In the second experiment, officers (n = 68) fired their weapons from various starting, or tactically ready, positions to measure the speed of movement to weapon discharge. Results of Part One showed
that contrary to training, all indexing positions were similar in time to contact the trigger, except indexing
high on the slide. Part Two revealed that point shooting was significantly faster than aimed shooting as well
as that the Low-Ready position was the fastest from which to fire, and the High-Guard ready position was
the slowest. These results may provide analytical and training implications to improve the safety of officers.
Introduction
Law enforcement officers are continuously
reminded of the risks and dangers they
face while working on patrol; however, just
recently, the International Association of
Chiefs of Police (IACP) (2014) has brought
attention to the threat of ambush assaults.
An ambush assault is considered to be an
attack on an officer that contains the element
of surprise, concealment of the assailant,
suddenness to the attack, and lack of provocation. From 2003 through 2012, 115 officers
were killed and 267 officers were injured
as a result of these types of attacks (Federal
Bureau of Investigation [FBI], 2014). While
ambush assaults may be premeditated, according to Law Enforcement Officers Killed and
Assaulted (LEOKA) reports, over 68% of the
ambushes that have occurred since 1990 have
been spontaneous and unprovoked (FBI, 2014).
Additionally, a vast majority (82%) of officers caught in ambush situations were alone
at the time. Although in 35% of cases, officers
were attacked with an assailant’s hands, in
over 36% of ambush assaults, officers were
attacked with a firearm, greatly increasing their risk of injury or death and rapidly
increasing the speed at which the attack
occurs (IACP, 2014). Alarmingly, the survival
rate for officers caught in an ambush situation is only 46% (IACP, 2014). With such a
low survival rate, the threat of officers being
hunted and attacked without notice gives one
more, of many, reasons to emphasize the need
for officers to be tactically ready, aware at all
times, and able to effectively respond.
1
To ensure tactical awareness, officers and law
enforcement trainers should understand the
rapid speed at which these ambush assaults
can occur to be better prepared to respond.
Analysis of deadly traffic stops has demonstrated that a suspect in the driver’s seat
can draw a weapon and fire at an officer in
as little as 0.23 seconds (s), with an average
time of 0.53 s (Lewinski, Dysterheft, Seefeldt,
& Pettitt, 2013). Research examining sprinting
mechanics has shown that the average individual, in their early 20s, is able to cover a distance of greater than 15 feet in just over 1 s
and slash or stab an officer with an edged
weapon (Dysterheft, Lewinski, Seefeldt, &
Pettitt, 2013; Lewinski, Hudson, & Dysterheft,
2014). If an individual attacking an officer had
their finger on the trigger of a handgun and
the handgun aimed, he or she would be able
to fire once in 0.06 s (the actual travel time of
the trigger to break point) and then fire an
additional round in just another approximate
quarter of a second (0.28 s) (Lewinski et al.,
2014). All the while, an officer faced with
a complex decision-making process, comprised of movement pattern recognition and
a choice response task, will take an average
of anywhere from 0.46 to 0.70 s to begin their
response (Lewinski et al., 2014; Ripoll et al.,
1995; Vickers, 2007). With the addition of
movement time to bring the weapon on target
and then time to return fire, unprepared officers are immediately placed at a tactical disadvantage during an assault. As officer survival rates during ambush situations nearly
double when officers take cover and are able
to return fire (IACP, 2014), it is pertinent that
officers be tactically and mentally prepared
to respond at the earliest possible moment.
Along with tactical movement training, early
threat detection and pattern recognition can
help to ensure officers stay ahead of the reaction curve.
As previously mentioned, it is known how
quickly an officer can fire a handgun once his
or her finger is on the trigger and even when
he or she is faced with a complex decision
from that position (see Lewinski et al., 2014).
2
Unfortunately, one key piece of information that is missing in the analysis of an officer-involved shooting is the amount of time
it takes officers to move their weapon from
whichever location it may be in to a firing
position. While officers are taught numerous
ready positions and finger-indexing positions, little to no research has examined the
amount of time it takes officers to react and
move from them. Therefore, it is unknown
what positions may most benefit officers with
the quickest responses during deadly use-offorce situations.
The first and seemingly most basic position
officers learn during their firearms training is
where to index, or place, their finger outside
of the trigger well when handling their gun
to minimize the risk of unintentional finger
movement and accidental discharge. It is theorized by law enforcement officers that placement of the index finger on the handgun has
a direct influence on finger movement time
and then the speed of trigger pull completion or weapon discharge time (DT). The DT
of a trigger pull is considered to be the time
from the initial movement of the index finger,
from its safe position outside of the trigger
well, to the time when the trigger is pulled
completely, resulting in weapon discharge
(Lewinski, 2003). Based off of author experience and observation of law enforcement firearms training, there are four finger indexing
positions that are predominantly taught and
practiced by officers: (1) the index finger is
placed straight ahead, resting on the trigger
guard; (2) the index finger is placed straight
ahead, the same as position a, however, the
finger has a slight bend, resembling a c-curve;
(3) the trigger finger is placed with the tip of
the finger on the frame of the weapon; and
(4) the trigger finger is placed resting with the
tip of the finger on the slide of the weapon
(see Figures 1a-d). While it is argued by some
law enforcement professionals that positions
a and b are considerably faster for shooting, it is speculated that the risk of unintentional discharge may be greatly increased
(Enoka, 2003; Heim et al., 2006). However, it
Figure 1. Commonly Trained Finger Indexing Positions
a
b
c
d
From upper left to lower right: (a) finger-indexed straight ahead, (b) finger-indexed straight
ahead with c-curve, (c) finger-indexed on frame, and (d) finger-indexed on slide
is still unknown whether any difference exists
between each position and if positions a or b
have any significant benefit of speed.
The position officers take when approaching a
potentially threatening situation is one of the
most arguably influential aspects to the speed
of their reaction and their ability to defend
themselves. According to Adams, McTernan,
and Remsberg (2009), officers should draw
their weapon if they have reason to believe
a deadly force situation may occur as it is
implied that drawing from a holster is likely
to take the longest time of any of the reactions
officers could have. Once their weapon is
drawn out of the holster, officers are trained
to take any one of numerous ready positions
to tactically prepare for a possible deadly
threat. Some positions, such as the Bootleg
position (Figure 2a), slightly conceal the
weapon from the suspect; while others, such
as the Belt Tuck position (Figure 2b), allow
officers to have their weapons directly in
front of them and ready for a possible deadly
encounter (Remsberg, 2001). Additional positions in which officers hold their firearms in
front of them at the ready are the High-Ready,
Low-Ready, and Sul (Figures 2c-e) positions
(Taubert, 2012). Officers are most commonly
trained to use these positions as they are
considered to be some of the safest and tactically ready positions to take when entering a threatening situation (Remsberg, 2001).
It should be noted that not all departments
train or endorse the use of all of the positions
studied.
3
Figure 2. Handgun Tactical Ready Positions
a
b
c
d
e
f
From upper left to lower right: (a) Bootleg, (b) Belt Tuck, (c) High-Ready, (d) Low-Ready, (e) CloseReady/Sul, and (f) High-Guard
While not widely practiced in the U.S., but
frequently seen, the High-Guard position
(Figure 2f), commonly used by Hollywood to
depict officers approaching threatening situations or doing building searches, is trained
and used in the United Kingdom. The HighGuard position is generally used as a safer
position than High- or Low-Ready when
officers are surrounded by others, such as in
a crowd situation, in order to prevent unintentional or accidental discharges and ricochets. It can be a very fast and accurate position from which to shoot. However, as it is
unpracticed in North America, it is believed
that an officer’s movement from this position
is very awkward, relatively slow, and likely
inaccurate. In North America, there is also a
stated concern that a High-Guard position
may result in unintended discharges upward
into multi-story buildings.
While handguns are the most commonly
used firearms by officers, a growing number
of departments have considered increasing
the use of rifles by their patrol officers (IACP
National Law Enforcement Policy Center
4
Board, 2007). The primary reason for this
shift is that officers are being outgunned by
deadly assailants, and the officer’s traditional
sidearm does not match the firepower of a
rifle or the myriad of other more powerful
weapons with which officers are confronted.
With the increasing use of rifles and shotguns
by officers, along with the growing number
of long barrel firearm assaults on officers, it
is important to start investigating the movement and timing differences while using
these weapons (for more information on rifle
use in law enforcement, see IACP National
Law Enforcement Policy Center Board, 2007;
see Figures 3a-c for shotgun positions).
In addition to ready positions, to improve
officer response time even more, some law
enforcement training experts also suggest that
when in deadly, close quarters combat, officers should use a variation of instinct, or point,
shooting relying on a visual fixation if possible and then kinesthetic alignment or pointing
(Adams et al., 2009; Applegate & Janich, 1998;
Conti, 2006, Vickers & Lewinski, 2012). When
using instinct shooting, officers align their
Figure 3. Shotgun Tactical Ready Positions
a
b
c
From upper left to lower right: (a) Port, (b) High-Ready/Modified Port, and (c) Low-Ready
gaze and often their body to point the muzzle
of their gun in the direction of the target and fire
without using traditional aiming or reference
to the sights of the weapon. Although some
may argue against this technique, previous
literature on officer shooting ability and gaze
patterns have observed that average officers
may spend too much time focusing on their
sights rather than their target, particularly in
close and fast-moving encounters, thus compromising speed and even shooting accuracy
(for more information, see Lewinski et al.,
in review; Vickers & Lewinski, 2012).
While all of the aforementioned positions
offer an advantage in various situations, it
is still unknown whether some might allow
officers to move and respond faster during a
deadly force situation. It is generally hypothesized that positions in which an officer’s gun
is held closer to the final firing position will
result in much quicker responses; however,
the degree to which these positions are faster
is still unknown. In general, understanding
all of the motor components of an officer’s
response, including the finger-indexing and
ready positions that offer officers the quickest reactions and best responses, are necessary to help better train and prepare officers
for unanticipated attacks. Therefore, the purpose of this study was twofold: (1) to examine the average amount of time it takes officers to fire their weapons, beginning from
various finger-indexing positions; and (2) the
speed of movement from commonly trained
starting positions to weapon discharge. These
two types of movement times were examined to determine and better understand officer movement during high-risk situations.
Additionally, in order to address the growth
of long barrel firearms use in officers, a small
pilot sample investigating shotgun movement times was also examined.
Methods: Part One
Participants
An original sample of 52 participants (94%
male) from a participating national government law enforcement agency volunteered for
the study. Participants were recruited through
information distributed by their supervisors, and they responded anonymously to
the testing site during their arranged firearms training session, as well as individually
on personal time. All participants were told
that the purpose of the study was to “better
understand trigger finger placement and the
influence of that on time to fire a weapon.”
All participants completed informed consent
waivers and demographic data information
forms before beginning any trials.
Equipment
All data collection took place at the participating government law enforcement agency’s firing range. The participants used their
own service pistols; therefore, no practice for
acclimation to the testing weapon used was
5
needed. The following weapon models were
used during testing: 9 mm Beretta (n = 1),
.45 caliber Colt (n = 1), .40 and .45 caliber
Glock (n = 40), .40 and .45 caliber SIG Sauer
(n = 11), and .45 caliber Springfield 1911
(n = 1). For digital movement analysis, the
trials were video-recorded using high-speed
digital cameras (Cannon Powershot s120,
Cannon U.S.A., Melville, NY, USA), filmed
at a rate of 120 Hz. The cameras were positioned on a tripod at weapon height and
located on the participant’s dominant hand
side to record all trigger finger movement.
Participant videos were digitized on a frameby-frame basis using commercial digital analysis software (Dartfish Prosuite 6.0, Dartfish,
Alpharetta, GA, USA). For data analysis, the
time for the initial movement of the trigger
finger to the time of contact with the trigger,
as well as the time the weapon was fired, was
recorded.
Procedures
To examine movement action time, the four
previously discussed finger-indexed positions were chosen for testing: (1) straight
ahead on trigger guard, (2) straight ahead on
trigger guard but with a c-curve, (3) at a slight
15° angle on the frame, and (4) at about a 30°
angle placed on the slide (see Figures 1a-d).
These finger-indexing positions were chosen
based on the current methods used for training purposes by law enforcement and military firearms instructors.
Prior to arriving at the range facility, all officers
were instructed to bring their service weapons. A participating researcher instructed
all officers on the procedure prior to testing.
Participants were told they would be completing a total of four trials, each beginning
from a different finger-indexing position, and
they would fire their weapon three times from
each position for a total of 12 rounds fired
overall. Prior to testing, all finger-indexing
positions were described and demonstrated
for the participant. Once participants were
cleared to enter the range, following their regulation range protocol, they were instructed
6
to approach the firing line with their service
weapons and four magazines.
After the range supervisor declared the range
hot, as instructed, participants drew their
weapons, inserted a magazine, and charged
the weapons while in their natural firing
stance. With the weapon pointed downrange,
participants were asked to take the first of
the randomized physical finger-indexed
positions as demonstrated by the researcher.
Participants were then instructed that they
would fire a total of three rounds beginning
from this indexing position, pausing at least
5 s between each round to ensure proper
indexing. The researcher gave participants
a signal when they were allowed to fire the
next round. Therefore, all of the movements
studied in Part One of this study were self-initiated, and the time recorded reflects only
movement time and not reaction and motor
movement time.
Although officers were not required to
respond immediately after the signal, they
were instructed by the researcher that they
were required to complete the trigger pull
as quickly as possible, without focusing on
weapon aim or accuracy. After completing
the first trial, the researcher instructed and
demonstrated the remaining three finger-indexed positions for participants to complete
in a randomized order. Upon completion of
all four finger-indexing trials, participants
were asked to clear their weapons and were
taken off the range by the range supervisor.
Once off the range, participants were allowed
to ask questions pertaining to the study and
given researcher contact information.
Data Analysis
Due to the observational nature of the study,
only descriptive and comparative statistical
analyses were performed on the variables.
The dependent variables measured for analysis were the movement action times for participants both (1) making contact with the
trigger (Contact Time) and (2) DT for each
of the defined finger-indexed positions. Both
variables were measured from the first initiation of movement observed in the trigger
finger. During video analysis, a fifth position was added for analysis, based on officer
positioning during testing (position e). This
position includes the index finger held low
on the trigger guard at a downward angle.
Data (25%) was analyzed for inter-rater reliability for DT time using intraclass correlation
coefficient (ICC) and coefficient of variations
(CV) (Hopkins, 2000). Additionally, all variances in Contact Time and Fire Time between
finger safety positions were analyzed using
an ANOVA with repeated measures and
Bonferroni-adjusted post-hoc testing.
to their department and supervisors and
responded randomly to researchers to
schedule testing times either while on duty,
with permission from supervisors, or when
off-duty. Upon arrival, participants were
informed that the purpose of the study was
to “better understand and examine the speed
of movement and how quickly officers can
fire their weapon from various starting positions.” All participants completed informed
consent waivers and were informed of all
details of the study prior to testing.
During data analysis, a discrepancy in performance effort and techniques was observed.
Although researchers emphasized that participants should fire as quickly as possible
and shoot without aiming, some participants,
likely due to habit, moved cautiously and
took the time to aim their weapons or keep
their weapons aligned downrange. Therefore,
officers were divided into two groups: (1) No
Aim and (2) Aim. Participants in the No Aim
group had no pause in their movement and
moved as quickly as possible; whereas participants in the Aim group distinctively aimed,
paused, or chose to perform movements very
slowly for precision. An independent samples t-test was used to compare the results of
these groups for both Contact Time and Fire
Time. As a result of the number of participants who aimed, Contact Time was used for
primary analysis for the various finger-indexing positions. The criterion to reject the null
hypothesis was p < 0.05. All descriptive statistics for both study parts are reported as mean
(M) ± standard deviation (SD), and change is
reported as Δ.
All data collection took place at the department’s training facility. Targets used for testing were provided by the facility. Participants
used their own service pistol (9 mm Beretta
[n = 30] and .45 caliber Smith and Wesson
[n = 38]); therefore, no practice was needed
for acclimation to a testing weapon. Of the 68
officers tested, nine also had Remington 870
shotguns and, therefore, were measured for
movement times from the various shotgun
ready positions. Participants were asked to
use their own weapon holster for testing to
prevent the need to practice. This also ensured
participants could move as quickly as possible without needing to adjust to an unfamiliar
new holster type or fit, which may have had
slower movement times and biased the data.
Methods: Part Two
Participants
An original sample of 68 participants (95%
male) from a large metropolitan police department participated in the study. Participants
were recruited through information provided
Equipment
A PACT shot timer was used to signal participants to begin movement, as well as to record
time to fire (Club Timer, PACT Inc., Grand
Prairie, TX, USA). A PACT shot timer creates
an auditory stimulus, most often a loud beep
or buzz sound, which signals the shooter to
fire. As the shooter fires his or her weapon,
the vibration caused by the weapon discharge
triggers an internal diaphragm, which then
timestamps the discharge time accurately to
within 0.01 s. The PACT shot timer is able to
record the times for three rounds fired after
the auditory stimulus. These times were
recorded by researchers in a separate format
immediately following each trial.
7
Because the officers were reacting to the
simple auditory stimulus of the PACT timer,
all of the times recorded for Part Two of this
study are inclusive of both a motor movement
time and an auditory reaction time. For reference purposes only, an average reaction time
to an auditory cue is just under 0.20 s (Vickers,
2007).
Procedures
6. Drawing weapon from holster unsnapped
(if applicable), raising weapon, sighting,
and firing
7. Drawing weapon from holster snapped
(if applicable), raising weapon into CloseContact/Combat Tuck position, and firing
8. Beginning in Low-Ready position with
finger on frame, raising weapon, sighting,
and firing
All officers reported to the designated testing site at the training facility for their scheduled testing time. One researcher greeted and
took all officers through the study procedure,
allowing them to use their own guns and holsters for testing. Officers were informed they
would be performing a number of trials, discharging their weapon at a target, beginning
from various positions, in reaction to a PACT
shot timer.
9. Beginning in Low-Ready position with
finger on frame, raising weapon, and
point shooting
Once participants were cleared to enter the
range, a researcher instructed them to stand at
a line 4.5 meters from a target placed directly
ahead of them. Officers were randomly
assigned to randomly complete 10 of the 20
shooting tasks below. A researcher guided
officers through each position and demonstrated the movement (if necessary). These
tasks included the following:
12. Beginning in Close-Ready position with
finger on frame, raising weapon, and
point shooting
1. Point shooting, from Weapon on Target,
Finger Indexed on Frame position
2. Point shooting, from Weapon on Target,
Finger on Trigger position
3. Firing a three round burst, sighted, from
Weapon on Target, Finger Indexed on
Frame position
4. Firing a three round burst, sighted, from
Weapon on Target, Finger on Trigger
position
5. Drawing weapon from holster snapped (if
applicable), raising weapon, sighting, and
firing
8
10. Beginning in High-Ready position with
and firing
11. Beginning in Close-Ready position with
and firing
13. Beginning in Belt-Tuck position with
and firing
14. Beginning in High-Guard position with
finger on frame, lowering weapon, sighting, and firing
15. Beginning in High-Guard position with
finger on frame, lowering weapon, and
point shooting
16. Beginning in Bootleg position with finger
on frame, raising weapon, sighting, and
firing
17. Beginning in Bootleg position with finger
on frame, raising weapon into CloseContact/Combat Tuck position, and firing
(Shotgun Pilot Research)
18. Beginning in Port position, bringing
shotgun down, sighting, and firing (if
applicable)
19. Beginning in Low-Ready position, bringing shotgun up, sighting, and firing (if
applicable)
20. Beginning in High-Ready position, bringing shotgun down, sighting, and firing (if
applicable)
Participants were instructed to complete each
task as quickly as possible in reaction to the
PACT timer’s signal. Researchers reminded
participants they should not take the time
to focus on aiming or accuracy but should
generally get a glimpse of their front sight
on the target and fire as rapidly as possible.
Only participants who had an on-duty shotgun were tested for the Port, Low-Ready, and
High-Ready shotgun tasks. Once participants
had completed each task, they were asked to
reload their weapons, if necessary, and given
instructions on the next task. After all of the
tasks were completed, the participants were
asked to clear their weapons and were escorted
from the firing range.
Data Analysis
The primary values measured in Part Two of
the study were times to react and complete
all of the aforementioned movement tasks
(1 to 17 in “Procedures” in the “Methods: Part
Two” section). Comparative analysis was performed on similar movement tasks to examine whether significant variance occurred. A
one-way ANOVA with Bonferroni adjusted
post-hoc testing was used to compare movement times for the Weapon on Target positions (Indexed Finger vs. Finger on Trigger,
both for the single and three shots fired). To
examine the effects of aim vs. point shooting
from the weapon in High-Guard, Close-Ready,
and Low-Ready positions on shooting time, a
two-way ANOVA was used with Bonferroni
adjusted post-hoc testing. Finally, a paired
t-test was used to compare movement times
from the Bootleg position, into sighted firing,
and into firing from the Combat Tuck position.
Due to the small number of officers who were
tested using shotguns, no comparative analysis was performed between the positions. The
criterion to reject the null hypothesis was p <
0.05. All descriptive statistics for both study
parts are reported as mean (M) ± standard
deviation (SD) and change is reported as Δ.
Results
Part One
Inter-rater reliability for the analysis of
DT for 25% of each indexing position was
extremely high (ICC = 0.96 and coefficient
of variation = 2.05%). All descriptive statistics are reported in Table 1. For Trigger
Contact Time, participants in the Aim group
(0.22 ± 0.12 s) were significantly slower than
those in the No Aim group (0.11 ± 0.06 s)
(p < 0.01). Likewise, participants in the Aim
group (0.55 ± 0.23 s) were significantly
slower to weapon discharge than those in
the No Aim group (0.20 ± 0.08 s) (p < 0.01).
Results from the one-way ANOVA demonstrated there was heterogeneity of variances
as assessed by Levene’s test (p = 0.01). There
was a significant effect of finger position on
Fire Time, Welch’s F(4, 82.12) = 7.92, p < 0.01,
and Contact Time, Welch’s F(4, 47.94) = 7.83.
Results of the Games-Howell post-hoc analysis indicated position d, or high on the slide,
was significantly slower for Contact Time
than positions a (p < 0.01), b (p < 0.05), and c
(p < 0.05). No other significant main effects for
finger-indexing positions were found.
Part Two
All descriptive statistics are reported in
Table 2. In the repeated measures ANOVA,
Mauchly’s test of sphericity indicated that
the assumption of sphericity had been violated, X2(2) = 65.43, p < 0.01. Epsilon (ε) was
used (ε = 0.74), as calculated according to
Greenhouse and Geisser (1959), to correct the
repeated measures ANOVA. Results of the
repeated measures ANOVA demonstrated
significant changes in movement time from
position, F(2.22, 150.67) = 42.43, p < 0.01.
From the Weapon on Target position, indexing on the frame was significantly slower
(p < 0.01) than indexing on the trigger. Starting
9
Table 1. Finger-Index Position Results for Contact and Fire Time
Group
No Aim
Aim
Overall
Position
Contact Time
Fire Time
A
B
C
D
E
Average
0.10 (0.06)
0.08 (0.05)
0.12 (0.05)
0.15 (0.05)*
0.12 (0.09)
0.11 (0.06)
0.18 (0.07)
0.17 (0.08)
0.19 (0.07)
0.25 (0.09)
0.25 (0.11)
0.20 (0.08)
A
B
C
D
E
Average
0.20 (0.08)
0.18 (0.06)
0.27 (0.12)
0.23 (0.09)*
0.10 (0.01)
0.22 (0.12)
0.50 (0.17)
0.47 (0.14)
0.61 (0.27)
0.57 (0.26)
0.60 (0.27)
0.55 (0.23)
A
B
C
D
E
Average
0.13 (0.08)
0.11 (0.07)
0.16 (0.10)
0.19 (0.09)*
0.12 (0.08)
0.15 (0.09)
0.30 (0.19)
0.26 (0.17)
0.32 (0.25)
0.42 (0.25)
0.39 (0.25)
0.33 (0.23)
* p < 0.05
Table 2. Movement Time Results for Tactical Ready Positions
Handgun Position
(1) Weapon on Target, Indexed Finger
(2) Weapon on Target, Finger on Trigger
(3) Weapon on Target, Indexed Finger, 3 Round Burst
(4) Weapon on Target, Finger on Trigger, 3 Round Burst
(5) Weapon in Holster, Snapped
(6) Weapon in Holster, Unsnapped
(7) Weapon in Holster into Combat Tuck
(8) Low-Ready, Indexed Finger, Aim
(9) Low-Ready, Indexed Finger, Point
(10) High Ready, Aim
(11) Close Ready, Aim
(12) Close Ready, Point
(13) Belt Tuck, Aim
(14) Weapon in High-Guard, Aim
(15) Weapon in High-Guard, Point
(16) Weapon in Bootleg, Aim
(17) Weapon in Bootleg, Combat Tuck
Shotgun Position
(18) Port
(19) Low-Ready
(20) High-Ready/Modified Port
*p < 0.05
**p < 0.01
10
Mean (SD)
0.51 (0.15)
0.37 (0.09)**
0.44 (0.15)
0.38 (0.13)*
1.82 (0.31)
1.68 (0.27)**
1.44 (0.31)**
0.97 (0.19)
0.64 (0.10)**
0.83 (0.20)
1.03 (0.20)
0.74 (0.11)**
1.02 (0.21)
1.13 (0.23)
0.73 (0.12)**
1.32 (0.20)
0.93 (0.19)**
Mean (SD)
1.28 (0.48)
0.99 (0.20)
0.84 (0.17)
Max
1.36
0.96
1.43
1.24
2.93
2.61
2.77
1.71
1.02
1.46
1.72
0.87
1.75
2.22
1.07
1.88
1.54
Max
2.88
1.35
1.15
Min
0.25
0.20
0.15
0.10
1.29
1.17
0.73
0.50
0.42
0.44
0.64
0.52
0.68
0.62
0.49
0.87
0.55
Min
0.79
0.63
0.65
from the Weapon in Holster position and
drawing into a Combat Tuck position was
significantly faster (p < 0.01) than aiming after
drawing from both a snapped and unsnapped
holster. Drawing from an unsnapped holster
was also significantly faster than drawing
from a snapped holster (p < 0.01).
Results from the two-way ANOVA demonstrated there was a significant effect of position, F(2, 247) = 20.48, p < 0.01, and aiming,
F(1, 247) = 46.58, p < 0.01. For the Weapon
on Target (both single and three shots fired),
Low-Ready, Close-Ready, and Weapon in
High-Guard starting positions, point shooting
was significantly faster (p < 0.01) than aiming
down the sights. The High-Guard position
was the slowest to fire from (p < 0.01), and the
Low-Ready position was the fastest (p < 0.01).
When participants began in the Bootleg position, firing from the Combat Tuck position
was significantly faster (p < 0.01) than raising the weapon to eye level and firing after
acquiring a sight alignment.
Discussion
The primary purpose of this study was to
understand and examine movement speed
from various holster types and finger-indexing positions, as well as how quickly an officer can fire his or her weapon from various
starting positions. By better understanding
the influence of these factors on the speed at
which officers can fire their weapons, officers
and law enforcement trainers may be able to
improve rapid response techniques to deadly
force situations.
The results from Part One of the study
demonstrated two important concepts. The
first is that, contrary to what many officers are
commonly taught, there is no significant difference in contact time found between the finger-indexing positions, except for position d.
When indexing their finger high on the slide,
officers were roughly 0.08 s slower to making
contact with the trigger and over 0.10 s slower
to fire than all other positions, except position e, low on the trigger guard. While many
law enforcement officers argue that indexing the finger on the trigger guard, curved
or straight, is faster than on the frame, the
difference in mean time to trigger contact in
comparison to the other positions (a, b, c, and
e) is less than 0.04 s. Therefore, when training
officers in which finger-indexing positions to
use, it might be more important to consider
implications of grasping reflexes, unintentional discharges, and effects of maintaining
weapon alignment. It is highly suggested that
further investigation into finger-indexing
placement and possible risk of unintentional
discharge take place.
The second important concept demonstrated
by these results is the average time experts
might expect officers to take to make contact
with the trigger and fire. Because positions
a, b, c, and e were all very similar in contact
and fire time, it should be generally accepted
during analysis that movement time to contact
with the trigger, from any of the faster positions (a, b, c, and e) will be an overall average
of 0.13 s. Additionally, an overall average for
officers with their finger indexed from any of
the faster positions and who quickly aim or
not are likely to fire their weapon in 0.32 s.
If officers move as quickly as possible, this
average time is decreased to 0.11 s to contact
time and 0.20 s to DT, respectively. These
values are supported by previous literature
examining the time to trigger pull completion
from trigger contact to weapon fire (Lewinski
et al., 2014).
One of the primary findings of Part Two was
the average amount of time it takes officers
to move from various ready positions using
both a handgun and a shotgun. With the use
of previously measured data, the movement
times collected in this study can help to create
a timeline of events from the time a stimulus is presented, such as a suspect drawing a
weapon, to the anticipated time of response
based on an officer’s positioning. As demonstrated in Table 2, officers beginning from some
of the most recommended ready positions of
Low-Ready, Close-Ready, and High-Ready,
may take anywhere from less than 0.50 s
11
to over 1.70 s to fire their weapon. Without
aiming, officers moving from the Low-Ready
position were fastest overall, firing in an
average time of 0.64 s. For tactical positions
involving aiming, the High-Ready position
was the quickest to fire from at 0.83 s. Other,
less recommended and perhaps not endorsed
by departments, but still highly used, positions, such as the Bootleg (aim: 1.32 ± 0.20 s)
and the High-Guard positions (aim: 1.13 ±
0.23 s), were similar in firing speed. While the
High-Guard position is very similar to HighReady, the large movement speed difference
is likely due to the lack of practice of the position in the U.S. Further research examining
this position movement time in officers from
the UK may shed more light on this variance
and the effects of practice on position speed.
Similarly, results of the pilot data from the
shotgun trial demonstrate that officers were
fastest when firing from the High-Ready or
Modified Port position. Contrary to what
researchers had expected, the fire time from
each of the shotgun positions was very close
to handgun times. Some of the fastest officers
firing with a shotgun were able to fire in just
over 0.60 s from the Low-Ready and HighReady positions (averages of 0.99 ± 0.20 s and
0.84 ± 0.17 s, respectively), and as quickly as
0.79 s from the Port position (1.28 ± 0.48 s).
Unfortunately, some officers took well over
1.0 s to fire from each of the shotgun positions, leaving far too much opportunity for an
assailant to attack. As with any skill, regular,
high amounts of repetition in practice at high
speeds will greatly benefit officers in reaction and moving as quickly as possible. With
the rise of assailant use of long barrel weapons, it is highly recommended that officers
who intend to use rifles or shotguns while
on patrol regularly practice each ready position and how they would move with speed
to an accurately aimed discharge. With the
immense threat posed by assault rifles used
against officers, minimal practice and slower
response times are likely to only result in
severe injury or death to the officer. Further
research is necessary to better understand
12
and improve officer training with long barrel
firearms, particularly their use in tactical situations such as an active assailant situation.
Not surprisingly, officers using point or
instinct shooting were significantly faster in
firing from each position (p < 0.01 for all positions). As supported by the results of Part One
of the study, point shooting was observed to
save officers over 0.30 s from each position
used in the comparison. This type of shooting can be effective at distances of 6.5 meters
or less and, with regular practice, up to
20 meters or more (Applegate & Janich,
1998). In a close-range confrontation, an officer taking the time to align and acquire their
sights will only delay their response time,
lessening their ability to neutralize a threat
and increasing their risk of injury or death.
Similar research supports this, finding that
law enforcement officers who use a point
shooting technique of driving their weapon
through their line of gaze instead of fixing
their focus on their sights have increased
levels of speed and accuracy when shooting
(Vickers & Lewinski, 2012), particularly at an
intermediate distance of conflict.
Another outcome anticipated by researchers was that individuals drawing from an
unsnapped holster were significantly faster
than those drawing from a snapped holster (1.68 ± 0.27 s vs. 1.82 ± 0.31 s). Although
these data demonstrate a benefit of officers
unsnapping their holster while approaching
a threatening situation, researchers observed
that many officers who had frequently practiced drawing quickly from their holsters
were actually slower and less accurate in their
movements when grasping their weapon
from an unsnapped holster. When the holster was unsnapped, each officer’s weapon
became slightly unstable within the holster,
and thus, as officers went in to grasp their
weapon, they needed to adjust their grip in
order to comfortably and automatically draw.
This was noted as a large disadvantage as
officers often rely on a familiar hand position
and automatic motor programs to quickly
and effectively draw and fire their weapons.
Additionally, the adjustment that took place
for this instability and the conscious attentional focus on the draw lengthened the time
it took officers to fire their weapons.
It is recommended that future research examine drawing times from unsnapped and
snapped holsters while officers maintain a
hold on their weapon while it is in the holster
in addition to the influence of other forms of
weapon retention and levels of retention in
police holsters. In this study, all holsters used
by officers had one or more forms of active
restraints, resulting in discharge times, in
reaction to simple stimuli with no additional
movements, of well over 1.5 s. However, the
holster type and the frequency with which an
officer practices drawing rapidly may play a
large role in the time it takes officers to return
fire. In a recent study of officer responses to
threatening traffic stop situations, officers
were required to respond to complex stimuli,
retreat from a deadly threat, and, either simultaneously to or after retreat, draw their weapons and return fire (Lewinski et al., 2013). It
was observed by researchers that some officers (n = 10) who were using more modern,
level two thermoplastic holsters were able
to perform all of the aforementioned movements and return fire in an average of 1.21 s
(Lewinski et al., 2013). Some officers were able
to go from grasping the weapon in the holster
to discharging in a time of 0.75 s. As officers
using the modern holsters were able to fire
their weapon during a high stress situation in
over 0.30 s less time than traditional holsters
using active restraints, further investigation
into training and holster type is highly recommended to determine possible speed and
retention benefits of thermoplastic holsters.
Limitations
As the current study was one of the first
investigative studies of its kind, there are limitations that exist in the research. The largest
limitation was that not all tests were performed in Part Two due to time constraints
within the participating officers’ schedules.
Additionally, only one police department
was tested for this research in Part Two.
While the experience and specific training
offered by that department may have influenced the abilities of officers, follow-up
investigation with multiple departments may
be used to aid in the verification of the results
of this study. The lack of information on shooting accuracy in correlation to movement times
and information on draw times from specific
holsters are also limitations; further research
and investigation can be done to examine
these components. Future studies could also
expand upon the long barrel firearm data
by investigating differences in weapon type,
training, and officer performance.
Implications
Overall, the values observed in this study are
key components in better understanding the
total response time of officers during a highly
stressful, and possibly life threatening, situation such as an ambush. These times not
only help to break down and analyze officer
response times, but also demonstrate how
quickly a deadly situation can unfold. Most
significantly, this research emphasizes the
drastic need for officers to be prepared to
respond as quickly as possible to potentially
deadly situations such as ambush assaults.
Some tactical ready positions allow on average for faster response times for officers; however, it is highly recommended that officers
train from the positions that are most comfortable and quickest for them. Additionally, it is
recognized that the positions which are most
practiced by officers will result in the quickest
reaction time; therefore, the positions found
to be the most tactically advantageous in this
study should be practiced and trained with as
often as possible to give officers the ability to
rapidly respond and to increase their chances
of safety and survival.
Acknowledgments
We would like to give our thanks to Patricia
Thiem, Bill Spence, Scott Buhrmaster from
13
Force Science, Jamie Borden from Henderson
Police Department, and Ron Libby, U.S.
Department of State, Diplomatic Security
Project Manager, for coordinating the participants through the trials and for all of their
help through data collection. Also, we thank
Lou Salcedo and Neil Goldberg from the
Los Angeles Police Department for their help
in research coordination and range management during data collection.
References
Adams, R. J., McTernan, T. M., & Remsberg, C.
(2009). Street survival: Tactics for armed
encounters. Northbrook, IL: Calibre Press.
Applegate, R., & Janich, M. D. (1998). Bullseyes
don’t shoot back. Boulder, CO: Paladin Press.
Hopkins, W. G. (2000). Measures of reliability
in sports medicine and science. Sports Medicine, 30(1), 1-15.
International Association of Chiefs of Police
(IACP). (2014). Ambush fact sheet (IACP Cooperative Agreement No. #2013-CK-WX-K022).
Retrieved from www.theiacp.org/AmbushProject.
IACP National Law Enforcement Policy
Center Board. (2007). The patrol rifle: Considerations for adoption and use. The Police
Chief, 74(2), 68-71.
Lewinski, W. J., Avery, R., Dysterheft, J. L.,
Dicks, N. D., & Bushey, J. (Under review).
The naïve shooter from a law enforcement
perspective: Hit probability.
Conti, M. E. (2006). Police pistolcraft: The reality-based new paradigm of police firearms training. North Reading, MA: Saber Press.
Lewinski, W. J., Dysterheft, J. L., Seefeldt, D. A.,
& Pettitt, R. W. (2013). The influence of
officer positioning on movement during
a threatening traffic stop scenario. Law
Enforcement Executive Forum, 13(1), 98-109.
Dysterheft, J. L., Lewinski, W. J., Seefeldt, D. A.,
& Pettitt, R. W. (2013). The influence of
start position, initial step type, and usage
of a focal point on sprinting performance.
International Journal of Exercise Science, 6(4),
320-327.
Lewinski, W. J., & Hudson, B. (2003). Time to
start shooting? Time to stop shooting? The
Tempe study. The Police Marksman, 28(5),
26-29.
Enoka, R. M. (2003). Involuntary muscle contractions and the unintentional discharge of
a firearm. Law Enforcement Executive Forum,
3(2), 27-40.
Lewinski, W. J., Hudson, B, & Dysterheft, J. L.
(2014). Police officer reaction time to start
and stop shooting: The influence of decision-making and pattern recognition. Law
Enforcement Executive Forum, 14(2), 1-16.
Greenhouse, S. W., & Geisser, S. (1959). On
methods in the analysis of profile data.
Psychometrika, 24(2), 95-112.
Remsberg, C. (2001). The tactical edge: Surviving high-risk patrol. Northbrook, IL: Calibre
Press.
Heim, C., Schmidtbleicher, D., & Niebergall,
E. (2006). The risk of involuntary firearms
discharge. Human Factors: The Journal of the
Human Factors and Ergonomics Society, 48(3),
413-421.
Ripoll, H., Kerlirzin, Y., Stein, J. F., & Reine, B.
(1995). Analysis of information processing,
decision making, and visual strategies in
complex problem solving sport situations.
Human Movement Science, 14(3), 325-349.
14
Taubert, R. (2012). Rattenkrieg!: The art and
science of close quarters battle pistol (1st ed.).
North Reading, MA: Saber Press.
Vickers, J. N. (2007). Perception, cognition,
and decision training: The quiet eye in action.
Champaign, IL: Human Kinetics.
Vickers, J. N., & Lewinski, W. (2012). Performing under pressure: Gaze control, decision
making and shooting performance of elite
and rookie police officers. Human Movement
Science, 31(1), 101-117.
Contact Information
*Dr. William Lewinski, PhD
Force Science Institute
124 E. Walnut Street, Suite 120
Mankato, MN 56001
[email protected]
Office Phone: (507) 389-1290
Office Fax: (507) 387-1291
Jennifer L. Dysterheft, MS
Force Science Institute
[email protected]
Jacob M. Bushey
Minnesota State University, Mankato
[email protected]
Nathan D. Dicks, MS
Minnesota State University, Mankato
[email protected]
* Corresponding author
15
Women and SWAT: Making Entry into
Police Tactical Teams
Thorvald O. Dahle, Department of Criminal Justice and Political Science,
North Dakota State University
Abstract
Since the late 1960s, women have made some progress in entering the policing profession; however, this is
not necessarily the case with all subunits within policing. SWAT units remain largely the domain of men.
Obstacles continue to deter female officers from applying for SWAT membership or succeeding when they
do apply. The current study examines the presence of female police officers on the SWAT teams for the 50
largest law enforcement agencies in the United States. Using interviews with SWAT team supervisors,
the testing processes for these teams are also examined to identify potential obstacles for women. Results of
this study find that women are rarely represented on SWAT teams and that SWAT testing processes may
be a contributing factor.
It has been more than a century since the
first woman police officer was hired. In 1905,
Lola Baldwin was hired in Portland, Oregon,
to patrol in street clothes to protect women
during the Lewis and Clark Exposition
(Schulz, 1993). By 1916, women were working in police departments in 30 cities; and by
1925, 417 women were working in 210 agencies (Garcia, 2003).
The 1964 Civil Rights Act created an equal
treatment standard for women in policing,
but it took several years for any real impact
to occur. In 1972, government agencies were
included in the provisions of the Civil Rights
Act, but the real incentive to change came
with the Crime Control Act of 1973, which
affected funding for agencies found to have
discriminatory hiring practices (Archbold &
Schulz, 2012). This change becomes apparent when, in 1971, only a handful of women
were on patrol with their male counterparts;
and by 1974, this number approached 1,000
(Milton, 1978).
Over the next two decades, the representation
of women in policing continued to increase. In
16
1987, women in large agencies made up 9.3%
of sworn positions; by 1990, it was 12.1%; and
by 2000, it was 16.3% of all sworn positions
in agencies serving populations of 250,000
or more (Bureau of Justice Statistics, 1991,
2002). As the 21st century began, the number
of women in policing appears to have hit a
plateau (Cordner & Cordner, 2011). Women
comprised 12.6% of all sworn positions in
2001 in agencies of more than 100 officers;
and by 2007, it had risen to about 15% in these
agencies (Bureau of Justice Statistics, 2010).
The most recent statistics show that women
make up 11.9% of all sworn positions (Bureau
of Justice Statistics, 2010).
Women have also faced resistance when seeking appointments into special assignments
like homicide and narcotics units, and SWAT
teams. Most of the opposition to women entering these specialized units stems from the
hyper-masculine subculture that exists within
such specialized groups. The research presented herein examines female representation
in SWAT teams in some of the largest police
agencies in the United States. In addition, this
article examines the entrance requirements
for membership in SWAT units to determine
if those requirements act as a barrier to female
police officers.
History of SWAT
SWAT (Special Weapons And Tactics) is a
general term that is used to describe specially
trained teams used by municipal, county, and
state police agencies for situations that are considered high risk. Depending on the region of
the country or particular agency, these teams
have many different names such as Special
Response Team (SRT), Hostage Barricade and
Terrorist (HBT) team, and Special Operations
Response Team (SORT). The specific responsibilities of these teams varies to some degree
from agency to agency; however, what is
consistent among these teams is an advanced
capability to handle a diverse set of weapons,
explosive entry devices, and specialized tactics to deal with situations regular patrol officers are not prepared to handle.
The origin of SWAT is often attributed to
the Los Angeles Police Department (LAPD)
(Weber, 1999). The turbulent social changes
that occurred in the 1960s presented several
challenges for which many police agencies
were unprepared. The Watts riots of 1966
taking place in and around the Los Angeles
area prompted the LAPD to take action.
Officer John Nelson presented the concept of
creating a special weapons and tactics unit to
then Inspector Darryl Gates to deal with the
riots, which led to the development of a small
group of officers who would receive the special tactical training (LAPD, 2012). The concept of this specialized team spread rapidly
to large police departments. By 1980, approximately 55% of the large police agencies in the
United States had adopted a SWAT team; and
by the late 1990s, this number increased to
90% (Kraska, 1999).
To create SWAT teams, police departments
drew from members of their agencies who
had previous military experience. The teams
were provided with training which was
above that found in typical police academies,
and they were held to a higher performance
standard than officers assigned to patrol.
Many of the SWAT teams formed at this time
were comprised of patrol officers who served
in this position on a part-time basis. In 1971,
the LAPD SWAT team became a full-time
assignment, which meant that its members
no longer worked in positions on patrol or
in investigations (LAPD, 2012). It was at this
time that SWAT teams became viewed as specialized, elite groups within police agencies
(Weber, 1999).
SWAT: A Gendered Institution
Within a Gendered Institution
SWAT teams have long been considered
a “boys club” that has kept women out by
having entry requirements based heavily
on physical agility and upper body strength
(Del Barco, 2008). The lack of female participation is evident when you look at the statistics. In 2001, only 17 of the 40,000 members
of the National Tactical Officers Association
(NTOA) were women (Prussel, 2001).
While the national percentage of women
working in sworn policing positions hovers
around 12% (Bureau of Justice Statistics, 2010),
the representation of women on SWAT teams
is far less than that (Prussel, 2001). Joan Acker
(1992) provides a framework for understanding issues related to gender in policing, particularly when considering the hyper-masculine
world of SWAT teams. In Acker’s theory of
“gendered institutions” (p. 567), she describes
how gender is present in processes, practices,
and images of social life. Certain institutions,
such as police agencies, have come to be
defined by the absence of women. The first
of Acker’s processes describes the decisions
and procedures used to control and exclude
members from institutions and groups based
on gender. Given the lack of female representation on SWAT teams, the practices used in
SWAT team personnel selection processes
may contribute to the exclusion of women.
17
The second gendered process considers
how images and ideologies are constructed
to legitimize institutions (Acker, 1992). The
images and ideologies of SWAT teams mirror
the elite military units in the mimicry of tactics, the value placed on masculine traits,
the manner of dress, and the use of a special
vocabulary. Military resistance to the admission of women is similar to that in policing.
It was not until 1976 that federal legislation
allowed women to be appointed to the service
academies (Boldry, Wood, & Kashy, 2001).
A U.S. Department of Defense (2010) report
suggested that the integration of women into
the military is also very similar to policing as
women held only 14.4% of the positions in the
active duty military.
Acker’s (1992) third and fourth processes
of gender involve how people “do gender”
(West & Zimmerman, 1987, p. 126) and their
internal process for constructing a persona
that is appropriate for the setting within the
institution. All SWAT teams are built on a
militaristic model emphasizing the concept
of team in their operation. As with specialized
teams in the military, these units are looking
for similar qualities in their team members—
specifically, members that possess traits that
are associated with masculinity (such as physical strength and aggressive behavior).
The process of “doing gender” and constructing a hyper-masculine persona has historically been a problem for women entering
policing. The construction of a masculine persona begins when cadets enter the police academy (Prokos & Padavic, 2002) and continues
on SWAT teams. For many, to be a member of
a SWAT team is to be considered among the
elite of an organization, similar to a member
of a Navy SEAL team or an Army Ranger unit.
As such, some SWAT members may view
this group as something worthy of protection from outsiders (or people who do not fit
the group’s “type”). The presence of women
in a hyper-masculine setting/group could
affect the work culture and public image of
the group by suggesting that masculine traits
18
of physical strength and aggressive behavior
are no longer relevant to the group (Martin,
1996). One way of protecting this masculine
image is to stop women from entering this
predominantly male subcultural bastion. This
perspective of hyper-masculinity is not only
consistent with the portrayal of SWAT teams
in popular media, but it is also supported by
research showing that it is the attitude shared
among its members (Kraska, 2001).
Many female police officers consider SWAT
teams the last area of male privilege in policing (Prussel, 2001). In the Dodge, Valcore, and
Klinger study (2010), male SWAT team members acknowledged they behave in a way that
would be viewed as professionally unacceptable to the outside world: “We would have to
watch our tongues”(p. 228). Female officers
are aware of the hyper-masculine culture
that exists in SWAT teams and are wary of
attempting access. One female police officer
in the Dodge et al. study reported, “I wouldn’t
fit in. I lack a penis” (p. 227).
Ultimately, the question is whether this culture of masculinity is necessary for the proper
functioning of SWAT teams. Research would
suggest that it is not as the integration of
women into military units has not significantly altered the group’s level of solidarity or
effectiveness (Harrell & Miller, 1997; Moskos,
1985). Little academic research, however, has
studied female police officers’ participation
in SWAT teams. Even less research has scrutinized the processes by which members are
selected for SWAT teams.
What Do We Know About Female
Police Officers and SWAT?
Today, there are many SWAT teams that have
not accepted female officers as members. In
fact, it becomes headline news when a female
officer successfully makes entrance into this
group. For example, in January 2012, the first
female officer on one SWAT team had to pass
testing, which included a one-mile run in less
than 12 minutes; three pull-ups; five dips; and
dragging the heaviest member of the SWAT
team, a man of nearly 300 pounds, 15 yards.
Each of these tests was in full gear, which
included the SWAT uniform, combat boots,
weapons, and a gas mask. Television cameras
rolled as the deputy was required to negotiate
the obstacle course. Bystanders can be heard
screaming, “I’m so pretty” or using a bullhorn just feet from her head to yell, “Come
on Goldilocks” (Fields, 2012). Some parts of
the entrance process are important and necessary, while other parts (like running through
a pool of water while being sprayed with a
fire hose) seem to border on hazing rituals.
The Los Angeles Police Department SWAT
team existed for over 40 years before allowing
the first female into their SWAT training program in 2008. Chief William Bratton drove the
change for the organization as he suggested
it was necessary to help “break all the glass
ceilings in the LAPD that kept women out of
many units” (Del Barco, 2008). Previous legal
challenges for the LAPD in 1994 had not been
enough to accomplish this as a $2 million
award for discrimination in the SWAT selection process did not result in any changes.
There is scant research on how many female
police officers are members of SWAT units.
Only recently have researchers studied female
police officers’ participation in SWAT or what
Dodge et al. (2010) described as “the last vestige of male dominance in law enforcement”
(p. 218). The study conducted by Dodge
and her colleagues examined the gendered
aspects of SWAT by interviewing both male
and female police officers. All of the male officers were SWAT team members, while 87% of
the female officers had no SWAT experience.
They found female officers felt SWAT was
a male-dominated subculture that tends to
exclude women. Also, both male and female
officers felt that female applicants had to prove
themselves in a fashion that creates barriers to
participating in SWAT. Both male and female
officers agreed that officer ability, not gender,
is the most important factor for being a SWAT
officer. Only 5% of male officers reported that
women have no place on SWAT, while 58%
said that women would be accepted on the
SWAT team. Comments from female officers
suggested that a “boys club” mentality existed
on SWAT and that women are not welcome.
Statements from male officers supported the
concern that women lacked the aggressiveness for SWAT and did not have the “search
and destroy” attitude that some feel is needed
to be an effective SWAT member.
Dodge, Valcore, and Gomez (2011) found
that male SWAT officers are becoming less
resistant to the presence of women in SWAT,
but they still question whether women have
the strength and skills necessary for the job.
They surveyed 117 male SWAT officers and
85 female officers of which only two had
SWAT experience. Male and female officers disagreed on a number of issues. Men
were more concerned than women about
female upper body strength and being able
to endure lengthy stakeouts. Male officers
were also more likely to believe that women
should not be on SWAT and were more likely
to feel women are not interested in being on
SWAT teams. Not only were male officers less
likely to believe female officers possessed the
unique skills which would make them valuable to SWAT, they also felt women were not
as capable as male officers in handling combative suspects and would be less likely to
use force. The study shows that the beliefs
held by male SWAT officers are reminiscent
of the beliefs women dealt with when they
first joined uniform patrol positions in the late
1960s. The reluctance of male SWAT members
to be accepting and encouraging of female
applicants suggests the inclusion of women
in SWAT will continue to be slow.
Overall, there has been limited research conducted on female officers and SWAT teams.
The research presented herein begins to fill
this gap in the literature by answering the following research questions:
19
1. How many female police officers are
members of SWAT teams in the 50 largest
police agencies in the U.S.?
2. What are the current requirements for
choosing SWAT team members in the 50
largest police agencies in the U.S.?
3. Have SWAT team requirements in the 50
largest police agencies changed in the last
four decades? If so, how have the requirements changed?
Methodology
Telephone interviews were conducted with
SWAT team representatives from 41 of the
50 largest local and state law enforcement
agencies in the U.S.1 The surveys were conducted from October to December 2012. The
list of the 50 largest agencies was drawn from
the Census of State and Local Law Enforcement
Agencies, 2008 (Bureau of Justice Statistics,
2011). All but one of the agencies had a SWAT
team or similar type of unit. Of the remaining
49 agencies, SWAT team representatives from
41 agencies participated in telephone interviews, resulting in a response rate of 84%. A
SWAT representative from one police agency
refused to provide information or participate
in the study, while SWAT representatives
from the remaining seven police agencies
either did not return phone calls or e-mail
requests after multiple attempts were made
to contact them.
The decision to use the largest police agencies
in the U.S. was based on recent statistics that
indicate that large police agencies employ a
higher percentage of female police officers
compared to smaller or medium-sized agencies. Specifically, 15% of sworn police officers
in large police departments and 13% in large
sheriff’s departments are women (Bureau of
Justice Statistics, 2010). In contrast, women
represent 6% of sworn officers in small local
police agencies and 4% in small sheriff’s
departments (Bureau of Justice Statistics,
2010). A second reason for using the 50 largest
20
police agencies is these agencies were more
likely to have had a SWAT team for a long
period of time, which allowed an examination of the changes in SWAT requirements
over time.
Telephone and/or e-mail contact was used to
initiate the interviews with SWAT team representatives. A request was made to speak with
either a SWAT supervisor or team member
who would have the most knowledge of team
selection procedures in each of the agencies.
A telephone interview was then conducted
that lasted from 15 to 60 minutes.
Interview questions inquired about the size
of the SWAT teams; the requirements necessary for a candidate to apply for SWAT; the
selection process used to find new SWAT
team members; the last time the selection process was changed and how; the number of
female officers currently on the team; the past
presence of women on the team; and the current number of sworn female officers within
the agency. The interview questions focused
solely on SWAT “operator” positions, meaning positions directly on the SWAT team and
not peripheral positions like canine officers or
hostage negotiators.
Each interview subject from the 41 participating police agencies provided current information on the staffing of the agency SWAT team.
Most team supervisors were unable to provide current data on the staffing and demographics of the agency overall. To get accurate
data for the size and total number of women
in each of the 41 agencies, several sources
were used, including agency websites, SWAT
team supervisors, agency personnel offices,
and agency recruitment/hiring offices.
Research Findings
The first research question inquired about the
number of women who are members of SWAT
teams. Table 1 provides a description of the
agency and the presence of women in both
the police agencies and the respective SWAT
teams. Among the law enforcement agencies
in this study, 14.6% of their sworn staff are
female officers, which is similar to the national
average of 15% among large municipal agencies and 13% among large sheriff’s departments (Bureau of Justice Statistics, 2010).
When examining the presence of women
on SWAT teams, a difference can be noted
between the SWAT team’s makeup and that
of the agency. While women represent 14.6%
of sworn patrol officers, women only represent 0.47% of SWAT team members.
Six of the 41 agencies had a female officer acting as an “operator” on their team
(see Table 2). Many of the SWAT teams had
female team members in the past, but 34.1%
of the teams had never had a female member
in their history. All eight of the current female
SWAT team members serve on teams in
municipal law enforcement agencies; none of
the county- or state-level SWAT teams have a
female team member. Four of the state teams
(40%) and four of the county teams (40%)
had female sworn team members in the past.
It was not possible to get an accurate total
count of female sworn staff ever serving on
these teams as few teams kept an accurate
count of females serving on their teams, especially if they had a long team history. Of the
27 teams who at one point or another had a
female team member, 13 had only one previous female member.
The second research question focuses on the
current requirements for choosing SWAT
team members in the 50 largest police agencies in the U.S. All of the police agencies in
this survey require some level of experience with the agency before they can apply
for the SWAT team. This varies from one to
five years of experience and, in most cases, it
would be rare for someone just meeting the
minimum level of experience to be able to join
the team. In one agency, joining the SWAT
team was strictly connected to seniority; and,
as a practical matter, about 20 years of experience was required to make it on the team.
Because of this close attachment to seniority,
no female officer has ever participated in the
testing process. The most senior female officer
in that agency was estimated to have about 16
years of experience and is still several years
away from reaching this standard for successful application.
One way of countering this requirement for
experience is to have past military involvement, especially if it was connected to elite
teams like Navy SEALS, Army Rangers, and
other similar specialized units. Several agencies mentioned altering their application
Table 1. Agency Staffing Information (41 Surveyed)
Agency Information
Municipal agencies sworn staff
County agencies sworn staff
State agencies sworn staff
Sworn staff summary
SWAT team staff
Agencies
21
10
10
41
41
Total Sworn
74,273
25,750
32,410
132,433
1,704
Female Sworn
13,571
3,561
2,178
19,310
8
% Female
18.3
13.8
6.7
14.6
0.47
Table 2. Participation of Female Officers on SWAT
Female SWAT Status
Ever had a female team member
Never had a female team member
Single female team member in team history
Teams with current female SWAT member
Number of Agencies
27
14
13
6
Percentage
65.9
34.1
31.7
14.6
21
protocols to allow candidates with this type
of experience to apply earlier in their career.
With the militaristic focus of these teams,
they are often led by team members with
past military experience. This leads to a belief
that candidates with past military experience
are a desirable addition. Some team leaders
mentioned how those with previous military
experience were able to make the transition to
these teams more easily than people with no
military experience.
process is the use of some type of obstacle
course. Although these courses commonly
require candidates to climb walls and fences,
get through a window, do a low crawl, or
drag a weighted dummy, they also differ from
agency to agency. Obstacle courses varied in
design from negotiating a four-foot high fence
to getting over a nine-foot fence or wall. The
dummy drag could vary from pulling a 120to 200-pound dummy anywhere from 40 feet
to 100 yards.
Agencies face the same challenges in developing selection processes for SWAT as they
do for their initial hire testing. Many agencies have tried to develop selection processes
connected to specific job criterion for hiring.
Unfortunately, the evidence for criterion-related validation is scant, and when it does
happen, it often leads to a questioning of the
legitimacy of the testing (Lonsway, 2003).
Although several SWAT commanders suggested they have tried to develop testing processes tied to job tasks, few could say the testing had undergone formal validation at any
time.
Fitness testing involved some standardized
components like push-ups, pull-ups, and situps (see Table 3). While the majority of agencies used these tests (78.0% use push-ups,
75.6% use sit-ups, and 68.3% use pull-ups), the
standards for these tests were anything but
standard. Minimum push-up requirements
went from a low of 18 to a high of 60, sit-ups
went from a low of 28 to a high of 80, and
pull-ups went from as low as one to as high as
12. For 85.4% of the agencies, these tests were
pass/fail; thus, candidates are eliminated if
they cannot meet the minimum score in any
of the elements of the process. Some agencies
had candidates test in gym gear, while others
required them to wear equipment like a ballistic vest, gun belt, police radio, or all of this
equipment. Pull-up testing was one area in
which candidates commonly had to wear a
weight or vest. One agency had the candidate
wear a 40-pound vest and complete one pull
up, while other agencies had the candidate
wear a 25-pound vest or weight for this test
and do as many as four pull-ups.
SWAT team operators are involved in many
activities that are different from normal patrol
officer activities. These activities may involve
physical exertion above normal conditions
and the wearing or carrying of heavier pieces
of equipment. SWAT teams use this reasoning
to justify having a higher selection standard
than for initial hire. Two of the 41 responding
agencies did not have a true physical fitness
testing process; one agency is implementing
their first physical fitness test just this year;
and one agency just implemented a physical
fitness standard only two years ago. For the
agencies that do not have fitness testing, they
rely more heavily on firearms testing or a few
specific job-related tasks such as demonstrating
the capacity to wear and function in a gas mask.
Many of the physical testing elements were
similar among agencies, but the overall combination of tasks and scoring were often
different. A common element in the testing
22
Other fitness testing was unstandardized in
its use. Distance running (one mile or more)
was used by 80.5% of the agencies, and the
bench press was used by 22.0% of the agencies. Every agency using the bench press used
a percentage of body weight as the standard,
but it varied from 72% of the candidate’s
body weight to 125% of the candidate’s body
weight. The most common distance for running was the 1.5-mile run with 61.0% of agencies using this test. Successful completion
times for the 1.5-mile run varied significantly
Table 3. Physical Agility Test Components (N = 41)
Fitness Tests
Number of Agencies
Percentage
Push-ups
32
78.0
Sit-ups
31
75.6
28
Pull-ups
68.3
1.5 mile run
25
61.0
Dummy drag
18
43.9
Obstacle course
17
41.5
Sprint
13
31.7
11
Hazmat suit/Phobia test
26.8
Bench press
9
22.0
6
Swimming
14.6
1.0 mile run
5
12.2
2.0+ mile run
3
7.3
Scoring of tests
Pass/fail†
35
85.4
Performance bonus††
20
48.8
†
Failure to complete an element of the fitness testing eliminates the candidate
††
Bonus points or consideration awarded to candidate for exceeding minimum requirements
from a low of 11:41 to a high of 17:00. The onemile run had a consistent time constraint with
most set at an 8-minute limit.
Another consideration in the fitness testing process was whether or not a candidate
received additional consideration for exceeding minimum fitness requirements or if fitness testing was strictly pass/fail with no
added benefit for exceeding minimum scores.
This is meaningful as it may place female candidates at a disadvantage if exceeding minimum strength testing increases the testing
score. Among these agencies, just under half
(48.8%) gave additional benefit for exceeding minimum standards. In several agencies,
it was formalized into a system of points for
completing each element with set standards
for more points at identified benchmarks.
The third research question asks whether
SWAT team requirements have changed in
the last four decades, and, if so, how have the
requirements changed over time. The interviews revealed that some agencies are making
changes to the SWAT application and testing
process. One commander mentioned that 20
years ago, “it was a good ol’ boy network”
where an applicant with the right connections just wrote a letter to get on the team.
For another county agency in the same state,
changes just occurred in 2012. Prior to the
testing being initiated in 2012, an applicant
only needed to put in a request for transfer,
and, if accepted, they became a SWAT team
member. Major metropolitan agencies are
facing similar challenges as one just implemented a fitness testing process in the last two
years, while two other agencies have little to
no physical screening process. In some cases,
this was a result of struggles with police
unions who resisted limitations being placed
on members that could remove them from
a position or consideration for one. Another
team commander specifically mentioned how
important it was that testing processes and
subsequent SWAT candidate training schools
not become a hazing ritual.
Table 4 shows that nearly half (46.3%) of the
agencies have revised their testing processes
in the past five years. The changes were wide
ranging and showed no real pattern. When
change occurred in these agencies, it did not
necessarily make the testing process easier.
In ten of the agencies, testing became more
23
Table 4. SWAT Team Testing History
Years Since Last Testing Change†
1-5 years
6-10 years
11-20 years
20 or more years
Unknown
Total
†
Figure based on estimate of respondent
Frequency
19
9
5
5
3
41
difficult as a result of changes to the process.
Some SWAT supervisors said it was to make
the testing fairer or more applicable to the job.
For eight of the agencies, the SWAT supervisor suggested the changes made the testing
easier. Although not in every case, the SWAT
supervisor often mentioned changes were
made to make the process more defensible
to legal challenge. Table 4 also shows that an
equal number of agencies (46.3%) have not
made changes to their selection process in six
or more years. For five agencies, the SWAT
supervisor reported it had been more than 20
years since the testing had changed for admission to SWAT.
Discussion
Until the 1970s, height and weight standards
were used by many law enforcement agencies
as part of their selection criteria. The inability
of police agencies to establish this as a bona
fide occupational qualification led to these
selection criteria being condemned by the
courts (Birzer & Craig, 1996). Law enforcement agencies are now facing similar challenges with fitness testing for SWAT team
applicants.
Female officers remain a token group in a traditionally male-dominated institution; thus,
their presence on a SWAT team would only
further highlight that status. The fact that
few female officers even apply to those teams
with little or no physical fitness testing would
seem to confirm these notions. Female applicants will face attitudes from some people
24
Percentage
46.3
22.0
12.2
12.2
7.3
100.0
that women do not belong on SWAT. An
example of the reception female candidates
sometimes receive is the comment heard by
the first female SWAT officer in one agency
during her first day in sniper training, “Are
you sure you’re in the right place?” (Russell,
2009). Most of the SWAT teams in the current study require a stringent, competitive
selection process, which they are resistant to
change. Team commanders were clear in their
position that female applicants are welcome
as long as they can meet the same standards
as male applicants.
Previous research (Dodge et al., 2010, 2011)
found that women police officers accept the
concept of heightened physical requirements
for membership on SWAT teams, but they
also noted that the atmosphere on SWAT
teams was not welcoming to women. This
same research found that male officers were
reluctant to lower physical standards in the
testing process, and they questioned the
ability of most female officers to be successful members on the team. While this emphasis on physical conditioning makes sense to
many team commanders, some recognize
other applicant characteristics might be more
important. As one team commander interviewed in the current study put it, “In today’s
climate, it’s really a thinking person’s game.”
He commented that many tests used in SWAT
team selection might be biased. This particular agency revised their testing process four
years earlier and in designing their new process, they strongly considered the question,
“How do you defend your process?” While
they use fitness testing to help determine who
gets into their SWAT candidate school, they
consider the SWAT school itself as the real
selection process.
physical requirements, women are at a disadvantage when compared to men. The product of these gendered procedures is a lack of
women in SWAT units.
The results of this study indicate that female
SWAT operators are rare, with only eight
women (0.47%) working in SWAT among
the 1,704 total SWAT operators in the 41
police agencies included in this study. Several
SWAT supervisors commented on how difficult it was to get women to apply for the
SWAT team or to retain women who had been
selected for SWAT. Part of the problem may
be linked to the selection processes, their scoring systems, and the testing process emphasis
on upper body strength. Nearly half (48.8%)
of agencies awarded a performance bonus for
exceeding the minimum requirements. Some
team commanders mentioned this was fair as
those who were more physically fit deserved
a higher score. The problem is that this could
minimize the importance of other skills critical for a good SWAT operator such as decision-making or the ability to effectively handle
stressful situations. This method of scoring
also reduces the likelihood female candidates
will be successful. Change is slow for many of
the agencies as almost half (46.3%) have not
changed their testing process in six or more
years. If and when change does occur, it does
not always increase the likelihood that female
candidates will be successful.
Acker’s (1992) second gendered process considers the images and ideologies institutions
use to legitimize their construction. Several
SWAT commanders mentioned the value of
military experience and the similarities they
share with elite military units. The image
that is constructed is one of elite fitness and
military tactics and appearance. As the militarization of SWAT units continues, the image
that is constructed is not one welcoming
to women. Women are rare in elite military
units like the Navy SEALS or Army Rangers.
As this image is used to model the construction of SWAT units, it is one that is notable for
the scarcity of women as members. This elite
military model provides legitimacy to SWAT
units and affords some justification for the
exclusion of women.
The results illustrate how SWAT is a gendered
institution within a gendered institution. As
Acker (1992) described in her first process of
gendered institutions, procedures are used
to control or exclude membership to an institution based on gender. This research shows
that women who have broken through the
initial barrier in a gendered institution face
another barrier when they consider applying
for SWAT. In addition to accessing a policing unit that is overwhelmingly masculine,
women face application and testing processes
that work to exclude them. By giving advantages to applicants with military backgrounds
or added benefit for exceeding minimum
Dodge et al. (2010, 2011) noted comments
from male officers about how they would
have to alter their behavior if women were
added to the SWAT team. This concern is consistent with Acker’s (1992) third and fourth
processes of gender in the context of SWAT.
How SWAT teams “do gender” and the methods they use for forming a masculine persona
have an effect on constructing this gendered
institution within a larger gendered institution. Results from this study confirm the presence of internal processes in SWAT, which
emphasize the value of physical strength
and perpetuate a masculine persona. Several
SWAT commanders mentioned their concern
with developing selection processes that were
strictly job related and avoiding an atmosphere conducive to hazing. They also mentioned the importance of good decision-making skills and other abilities that more traditional selection processes do not measure.
Should law enforcement agencies change the
focus of their selection processes in this direction, it will begin to alter the traditional male
25
SWAT persona and provide more access for
women.
The insightful examination of the processes
SWAT teams use for candidate selection is
the first step for agencies that want to erode
the hyper-masculine perception of this gendered institution. The results of this study
reinforce this but also highlight the importance of making the operational culture of
SWAT teams more welcoming to women.
Unnecessary standards, ritualistic training,
and hazing in the name of team building pervade the selection processes used in some law
enforcement agencies. Add to this a militaristic atmosphere of hyper-masculinity prevalent in some SWAT teams, and agencies
are unlikely to see a dramatic increase in the
number of female candidates. Agency leaders
must take an active role in the management of
SWAT teams to counter this environment and
to ensure that they operate in a legally defensible manner that is more inclusive to women.
Endnote
1
This study is limited to the 50 largest police
agencies in the United States. Future studies
should expand the scope of this research to
include small and mid-size police agencies.
Smaller agencies’ SWAT teams may differ in
selection processes and operational culture
as they cannot be as selective as larger agencies. The necessity of being less selective may
result in the participation of more women,
and, thus, those agencies may not have the
hyper-masculine ethos of the larger agencies
in this study. A larger study could also allow
for a quantitative analysis of the testing used
by SWAT teams to determine if different
requirements affect the rate of women applying for SWAT teams and succeeding in joining these teams. Combined with the broader
range of interviews, this may help to determine which factors contribute to the participation of women on SWAT teams.
Acknowledgment
The author would like to thank Dr. Carol
Archbold for her mentorship and for her
invaluable encouragement with this project.
References
Acker, J. (1992). From sex roles to gendered institutions. Contemporary Sociology, 21, 565-569.
Archbold, C. A., & Schulz, D. M. (2012).
Research on women and policing: A look at
the past, present, and future. Sociology Compass, 6, 694-706.
Birzer, M. L., & Craig, D. E. (1996). Gender
differences in police physical ability test
performance. American Journal of Police, 15,
93-109.
Boldry, J., Wood, W., & Kashy, D. A. (2001).
Gender stereotypes and the evaluation of
men and women in military training. Journal of Social Issues, 57, 689-705.
Bureau of Justice Statistics. (1991). Police departments in large cities, 1987. Washington, DC:
Office of Justice Programs, U.S. Department
of Justice.
Bureau of Justice Statistics. (2002). Police
departments in large cities, 1990-2000.
Washington, DC: Office of Justice Programs,
U.S. Department of Justice.
Bureau of Justice Statistics. (2010). Crime and
data brief. Women in law enforcement, 19872008. Washington, DC: Office of Justice Programs, U.S. Department of Justice.
Bureau of Justice Statistics. (2011). Census of
state and local law enforcement agencies, 2008.
Washington, DC: Office of Justice Programs,
U.S. Department of Justice.
Cordner, G., & Cordner, A. (2011). Stuck on a
plateau? Obstacles to recruitment, selection,
26
and retention of women police. Police Quarterly, 14, 207-226.
Del Barco, M. (2008, April 29). LA SWAT unit
on verge of accepting first woman. National
Public Radio. Retrieved from www.npr.org/
templates/story/story.php?storyId=90015810.
Dodge, M., Valcore, L., & Gomez, F. (2011).
Women on SWAT teams: Separate but
equal? Policing: An International Journal of
Police Strategies & Management, 34, 699-712.
Dodge, M., Valcore, L., & Klinger, D. A. (2010).
Maintaining separate spheres in policing:
Women on SWAT teams. Women & Criminal
Justice, 20, 218-238.
Fields, T. (2012, March 28). First female
SWAT member almost didn’t get her shot.
WSTP-TV. Retrieved from www.wtsp.com/
video/1533218156001/1/First-female-SWATmember-almost-didn't-get-her-shot.
Garcia, V. (2003). Difference in the police
department: Women, policing, and doing
gender. Journal of Contemporary Criminal Justice, 19, 330-344.
Harrell, M. C., & Miller, L. L. (1997). New
opportunities for military women: Effects upon
readiness, cohesion, and morale (Report No.
DASW01-95-C-0059). Santa Monica, CA:
RAND.
Kraska, P. B. (1999). SWAT in the commonwealth: Trends and issues in paramilitary
policing. Kentucky Justice and Safety Research
Bulletin.
Kraska, P. B. (2001). Militarizing the American criminal justice system: The changing
roles of the armed forces and the police. Boston:
Northeastern University Press.
Lonsway, K. (2003). Tearing down the wall:
Problems with consistency, validity, and
adverse impact of physical agility testing in
police selection. Police Quarterly, 6, 237-277.
Los Angeles Police Department (LAPD).
(2012). S.W.A.T.: Special weapons and tactics.
Retrieved from www.lapdonline.org/inside_
the_lapd/content_basic_view/848.
Martin, S. E. (1996). Doing gender, doing police
work: An examination of the barriers to the
integration of women officers. Presented at
the Australian Institute of Criminology
Conference.
Milton, C. H. (1978). The future of women in
policing. In A. W. Cohen (Ed.), The future
of women in policing (pp. 183-204). Beverly
Hills, CA: Sage.
Moskos, C. C. (1985). Female GIs in the field.
Society, 22, 28-33.
Prokos, A., & Padavic, I. (2002). “There
oughtta be a law against bitches”: Masculinity lessons in police academy training.
Gender, Work, and Organization, 9, 439-459.
Prussel, D. (2001). Women where? Law and
Order, 49, 86-90.
Russell, L. (2009, March 2). Female Va. SWAT
sniper tackles crime and stereotypes. The
Virginian-Pilot.
Schulz, D. M. (1993). From policewoman to
police officer: An unfinished revolution. The
International Review of Police Development, 16,
90-98.
U.S. Department of Defense. (2010). Demographics 2010: Profile of the military community. Washington, DC: Office of the Deputy
Under Secretary of Defense. Retrieved from
www.militaryonesource.mil/12038/MOS/
Reports/2010_Demographics_Report.pdf.
Weber, D. C. (1999). Warrior cops: The ominous growth of para-militarism in American
police departments (Briefing Paper No. 50).
Retrieved from http://object.cato.org/sites/
cato.org/files/pubs/pdf/bp50.pdf.
27
West, C., & Zimmerman, D. H. (1987). Doing
gender. Gender & Society, 1, 125-151.
Thorvald O. Dahle is a doctoral student, instructor, and teaching assistant
in Criminal Justice at North Dakota State
University in Fargo. His research interests
include policing and issues regarding ethics, race, and gender. He has published in
Police Quarterly, Race and Justice, and the
Journal of Interpersonal Violence. He earned
a Master’s degree in Public and Human
Service Administration from Minnesota
State University Moorhead. He spent 24
years in law enforcement, including serving as a chief of police.
Contact Information
Thorvald O. Dahle
North Dakota State University
Department of Criminal Justice and
Political Science
NDSU Department 2315
1616 12th Avenue North
PO Box 6050
Fargo, ND 58108-6050
(218) 329-0386
[email protected]
28
Keeping Kids Out of Corrections: Lowering
Recidivism by Strengthening Teamwork and
Collaboration Between Law Enforcement
Officers and Transition Coordinators
in Juvenile Correctional Facilities
Theresa A. Ochoa, PhD, Associate Professor of Special Education,
Indiana University
Lawrence J. Levy, PsyD, Florida Private Practice
Kelly M. Spegel, Doctoral Student, School Psychology, Indiana University
Yanua F. Ovares, Clinical Lecturer, Special Education, Universidad de Costa Rica
Abstract
Transition to community support is critical to efforts to reduce recidivism among juveniles involved with
the justice system. Police officers have contact at the time of arrest. Parole officers usually are involved
primarily when the adolescent is returned to the community. Transition coordinators are involved during
incarceration and after discharge. But, too often, these professionals work in isolation of each other to the
detriment of maintaining any gains made during an adolescent’s time in custody. In this article, we propose that working in isolation limits successful reintegration back into the community and that strengthening communication and teamwork collaboration between law enforcement personnel and the juvenile
correctional facility transition coordinator will directly benefit adolescents and reduce recidivism.
Juvenile delinquency and crime are significant national concerns in the United States.
Although crimes committed by juveniles
have decreased from 2006 to 2011, the Federal
Bureau of Investigation (FBI) reported a total
of 1,470,000 crimes committed by minors
ages 10 to 17 during the 2011 fiscal year compared to 2,213,500 crimes committed in 2006
(Puzzanchera & Kang, 2014). Offenses committed by juveniles include violent crimes
such as murder, rape, aggravated assault, and
robbery; property crimes such as burglary,
auto theft, and arson; and non-indexed crimes
such as drunkenness, drug abuse, prostitution, and vagrancy. The same FBI report
estimated that 4,396 minors were arrested in
2011. Statistics provided by Sickmund (2010)
show that in 2008, 81,000 juveniles committed crimes that led to them eventually being
remanded to state residential correctional
facilities. Nellis and Wayman (2009) report
that 100,000 youngsters are discharged from
the juvenile justice system each year after
confinement in a residential treatment center,
training school, boot camp, group home, private placement facility, or state juvenile correctional facility.
Juveniles of color (Puzzanchera, 2013) and
juveniles with disabilities (Archwamety &
Katsiyannis, 2000; Cavindish, 2013; Leone,
Krezmien, Mason, & Meisel, 2005) are at
higher risk of involvement with the juvenile justice system and make up the highest
proportion of incarcerated youth. Statistics
from the U.S. Department of Justice’s Office
of Juvenile Justice show that 68% of individuals detained and confined were children of
29
color—mainly African American and Hispanic
(Puzzanchera, 2013). Cavindish (2013) reports
that anywhere from 20 to 90% of the population of incarcerated juveniles is composed
of juveniles with disabilities. Adolescents
with learning disabilities and emotional and
behavioral disorders make up the largest
portion of juveniles in correctional facilities
(Mears & Aaron, 2003). The data show that
youth of color and youth with disabilities are
at a higher risk for delinquency and involvement with law enforcement.
National juvenile justice data show that
incarceration costs $66,000 per adolescent
per year (Sickmund, 2010). A more recent
report published by the Southern Education
Foundation (SEF) (Suitts, Dunn, & Sabree,
2014) shows that in the state of Georgia, the
cost of incarceration for any single juvenile in a residential facility is in the range of
$88,000 to $91,000 per year. In the states of
Tennessee and Virginia, the cost of incarceration is $92,000 and $101,000, respectively.
In contrast, a report by the New America
Foundation (2012) shows the annual cost per
student in community schools ranges from a
low of $6,612 in the state of Utah to a high of
$19,698 in the District of Colombia.
Financial considerations aside, it is not reasonable to expect that confinement alone will
result in a change of behavior for adolescents.
According to Snyder and Sickmund (2006),
some states report recidivism rates as high as
55%. Nellis and Wayman (2009) indicate that
50 to 70% of previously incarcerated juveniles return to a correctional facility within a
year of release. If the goal of incarceration is
to reform or rehabilitate juveniles who have
committed crimes, it is imperative to understand what supports are necessary to maximize the possibility that adolescents will not
come into contact with law enforcement for
new crimes once released from custody.
In this article, we summarize the laws which
mandate the provision of transition services
for juveniles leaving custody and describe the
30
traditional roles assumed by transition coordinators, juvenile parole officers, and school
resource officers during rehabilitation and
at the time of transition from juvenile correctional facilities. We provide a summary
of best practice guidelines for transition and
offer state-of-the-art recommendations for
strengthening collaboration and communication between law enforcement and juvenile
justice personnel to support youth when they
return to their communities after confinement.
Laws Requiring Transition Support
to Delinquent Juveniles
Two federal-level laws mandate transition
services for juveniles exiting correctional
facilities. The No Child Left Behind Act (NCLB),
the federal law that governs the treatment
and education of all students in the nation,
aims to ensure that all children, including
delinquent youth, have access to a high-quality education. Under the NCLB, juvenile correctional facilities in the U.S. are under obligation to provide not only educational services
to all youth in custody, but also transition
services (Sheldon-Sherman, 2010). The NCLB
explicitly requires that correctional facilities
hire a transition coordinator with the aim of
ensuring each adolescent is successfully reintegrated into his or her community.
The Individuals with Disabilities Education Act
(IDEA) of 2004, the federal law governing
the education and treatment of students with
disabilities, also mandates transition services.
Transition support is described within the
IDEA as
a coordinated set of activities for a child with
a disability that: (a) is designed to be within
a results-oriented process, that is focused
on improving the academic and functional
achievement of the child with a disability to
facilitate the child’s movement from school
to post-school activities, including post-secondary education, vocational education,
integrated employment (including supported
employment), continuing and adult education,
adult services, independent living, or community participation; (b) is based on the individual child’s needs, taking into account the
child’s strengths, preferences, and interests;
and (c) includes instruction, related services,
community experiences, the development of
employment and other post-school adult living
objectives, and, when appropriate, acquisition
of daily living skills and functional vocational
evaluation.
For adolescents with disabilities, transition
services encompass providing support when
an adolescent transitions from community
to correctional facility, from one facility to
another, and from the correctional facility
back to the community (Clark, Mathur, &
Helding, 2011; Osher, Amos, & Gonsoulin,
2012). The NCLB and the IDEA are not only
federal mandates to provide transition services when an adolescent leaves confinement;
they also provide opportunities for educators
and law enforcement personnel to collaborate
as a team to increase the chance that the youth
will avoid committing some crime again.
Best Practice Recommendations
for Transition Services and
Supports
Providing transition services when youth
leave confinement is essential to maintain
progress made during incarceration and in
reducing recidivism once back in the community. A comprehensive examination of
published literature on transition support,
conducted by two of the authors of this article
(Ochoa & Spegel, 2015), yielded four broad
phases (or stages) that require transition support: (1) intake, (2) during confinement, (3) the
period just prior to release (imminent release),
and (4) post release. From this model of focusing on transition as a means for reducing
recidivism emerge several guiding principles:
(1) transition should drive educational programming while juvenile is in the correctional
facility; (2) transition should be approached as
a multidisciplinary team effort; (3) transition
goals and progress should be monitored
regularly; and (4) transition services should
seek to reengage students immediately upon
release. Table 1 summarizes guiding principles and ideal activities that should be carried
out at each stage of incarceration.
Transition Support at the Point of Intake
The phrase “Think exit at entry” coined by
Risler and O’Rourke (2009) captures the
importance of approaching transition support as a process that begins when the individual enters the facility and which guides
the planning of services while the youth is in
custody. In essence, considerations and planning for transition back to the community
should drive programming while youth are
housed in correctional facilities (Baltodano,
Mathur, & Rutherford, 2005a; Stephens &
Arnette, 2000). Risler and O’Rourke (2009)
advocate for beginning this with an intensive intake procedure that includes a review
of educational records, completion of psychological and educational achievement
assessments, and the creation of a transition
portfolio that will contain important information for assisting the youth upon release
(e.g., items like academic records, job skills,
a résumé, and reference letters). Published
best practice guidelines show that developing
a comprehensive treatment and educational
plan that focuses on the social, emotional, and
adaptive living skills is critical to preparing
youth to return to the community and assisting them in not returning to the attention of
the legal system (Nellis & Wayman, 2009;
Risler & O’Rourke, 2009; Sheldon-Sherman,
2010; Stephens & Arnette, 2000). Best practice
guidelines also show that connecting youth
with a mentor or advocate in their communities to provide support and help the youth
advocate on their own behalf is advisable
(Osher et al., 2012). Considering that the
majority of incarcerated youth are below their
expected grade level in academic achievement
and since youth with disabilities are over-represented within correctional facilities, it is
imperative that educational records are sent
31
Table 1. Recommended Transition Support at Each Stage of Incarceration
Stages of Incarceration
Principle and Recommended Actions
Transition should drive educational programming while the juvenile is in a
Intake
correctional facility:
• Conduct a review of records, complete assessments, and begin developing
portfolio for transition
• Plan for access to mental health and substance abuse services
• Seek acquisition of records from youth’s school
• Design and implement a skills training individualized program
• Coordinate information from facility with parole officers and community
schools
• Include a mentoring support component in transition plan
Transition should be approached as a multidisciplinary team effort:
• Communicate and coordinate with all service providers in the first week
• Indicate each service provider’s responsibilities and create a system of
accountability for transition goals
• Include relevant nonprofessionals
• Should be coordinated by a community-based service provider
• May be guided by the adolescent’s family and the adolescent to the extent
appropriate
• Include formal collaboration among juvenile justice and community
agencies
Transition goals and progress should be monitored regularly:
Confinement
• Meet weekly with transition team to monitor progress and adjust plans
and services as needed
•
Conduct pre-release meeting 60 days prior to release to review portfolio,
Imminent Release
discuss transition, and finalize plans for return to community
• Develop educational plan from facility to community school two weeks
prior to transition
• Determine the most appropriate educational placement
• Visit the community school
• Assess family and living environment to which the student is returning
• Conduct formal exit interview 10 days prior to release to assess progress
• Provide finalized portfolio
Transition services should seek to reengage students immediately upon
Post Release
release:
• Assess strengths in family, community, and student, as well as risk factors
• Provide services that include social, educational, occupational, health, and
community supports
• Connect youth with mentor
• Create a support system for developing positive peer connections
Transition services should facilitate transition to school and employment:
• Send records from facility to educational placement
• Enroll youth in transitional educational placements where available
• Provide school-based probation officers for transitioning youth when
possible
Transition goals and progress should be monitored regularly:
• Continue to meet with transition team after the adolescent is released
32
to the correctional facility prior to arrival in
order to create an effective educational plan
(Risler & O’Rourke, 2009; Sheldon-Sherman,
2010; Stephens & Arnette, 2000). Furthermore,
the educational curriculum at the correctional
facility should be coordinated between the facility and the school in the community to which
the youth will return in order to better facilitate the transfer of credits (Stephens & Arnette,
2000). Sheldon-Sherman (2010) recommends
that these services be coordinated and communicated within the first week of incarceration.
One of the most important components of
transition support at the point of entry is the
formation of a transition team composed of
professionals from several disciplines. Ideally,
a service provider from the community to
which the youth will be returning, such as a
probation officer, should coordinate transition services instead of an individual based
out of the juvenile justice system (Barton,
2006; Stephens & Arnette, 2000). The transition team should be formed under the guidance of this community-based coordinator
and should include the youth and his or her
family (JustChildren, 2006; Risler & O’Rourke,
2009; Sheldon-Sherman, 2010); important
nonprofessionals in the youth’s life (i.e., youth
pastor, mentor) (Barton, 2006); and key stakeholders from within the facility, probation/
parole department, community agencies,
and the school to which the youth is returning. This multidisciplinary team should be a
formal collaboration between juvenile justice
and these community agencies (Barton, 2006;
JustChildren, 2006). Finally, in order to oversee that each key stakeholder is fulfilling his
or her assignment, transition plans should
explicitly and specifically describe each service provider’s responsibilities and create a
system of accountability for transition goals
(JustChildren, 2006; Sheldon-Sherman, 2010).
Transition Support During Confinement
It is important that educational and therapeutic services provided are monitored
throughout incarceration and that changes to
the transition services plan are made when
goals are not being met as expected (Nellis
& Wayman, 2009; Osher et al., 2012; Risler &
O’Rourke, 2009). Ongoing monitoring involves regularly scheduled meetings, led
and organized by the transition coordinator,
involving teachers, therapists, psychologists,
and correctional staff, during which each
youth’s progress is discussed and the educational and therapeutic treatment plan is
modified and refined based upon new assessment data and professional observations.
Ongoing monitoring means that treatment is
constantly revised based upon feedback from
each member of the multidisciplinary team,
including the adolescent and his or her family,
as managed by the transition coordinator.
Transition Support at the Point of
Imminent Release
The short period of time just prior to release
from the facility is also of vital importance
when considering the best way to support
youth being released into their communities. The development of educational support
from the previous stage is particularly important in the brief period just before release. It
should not be assumed that youth will return
to the same school from which they left. It
is up to the transition team to determine the
best educational placement for each youth
upon release (JustChildren, 2006). The transition team, including the youth, should also
visit the community school to set up a support system with the purpose of easing the
youth back to the school (Sheldon-Sherman,
2010; Stephens & Arnette, 2000). Risler and
O’Rourke (2009) propose a very explicit timeline of imminent release support. Program
counselors, facility administrators, and the
youth’s guardian should meet 60 days prior
to the youth’s release to review the youth’s
file, discuss transition, and finalize plans
for the youth’s return to the community. A
formal exit interview with facility personnel,
the youth’s guardian, and the parole officer
should take place 10 days before release to
33
assess progress and provide a finalized transition portfolio (Risler & O’Rourke, 2009).
Transition Support Post Release
The transition team, including the service
providers in the youth’s community, should
seek to reengage students in pro-social activities and connect them with support services
in the community immediately upon release
(Anthony et al., 2010; JustChildren, 2006; RoyStevens, 2004). In order accomplish this, transition services should include an assessment
of the family and living environment to which
the youth is returning (Sheldon-Sherman,
2010) and an assessment of the community
support systems available (Barton, 2006). This
assessment will guide coordinators in determining what services may need to be put in
place to best help families cope with the return
of the youth and the changes which happen
as a result and to reinforce skills their child
learned while incarcerated. In the same way
that services in the facility should be determined based on each youth’s specific needs,
the services provided after youth are released
should also be individualized based on the
youth’s needs (Anthony et al., 2010; Baltodano
et al., 2005a; Nellis & Wayman, 2009). It is
vital that the types of services (e.g., counseling, academic tutoring) provided by the facility be continued after release. This continuation and overlap of services from correctional
facility to community is critical for the success
of the youth. Services should include social,
health, and community supports (Anthony
et al., 2010); expanded mentoring support,
(Baltodano et al., 2005a; Osher et al., 2012;
Sheldon-Sherman, 2010; Stephens & Arnette,
2000); and support systems for developing
positive peer connections (Baltodano et al.,
2005a).
Youth returning to their communities from
confinement often find it difficult to find
employment and face barriers when attempting to return to school. Lack of vocational skills
and unsupportive school environments contribute to the high proportion of unemployed
34
and uneducated formerly incarcerated youth
(Baltodano et al., 2005a). To overcome these
barriers, transition services must be in place to
help facilitate the transition from incarceration
to school and employment (Baltodano et al.,
2005a). At the most basic level, it is recommended that facility charts, files, and records
for each youth be transferred back to community schools in a timely manner to help in
transition planning (Risler & O’Rourke, 2009;
Roy-Stevens, 2004; Stephens & Arnette, 2000).
Stephens and Arnette (2000) recommend placing youth in alternative or transitional schools
for a period of time to help youth readjust
to educational demands outside of the correctional facility prior to enrolling them in
mainstream public schools. The authors also
recommend placing probation officers within
community schools to further establish partnerships between juvenile justice and education as well as provide support for both
schools and returning youth. Ongoing monitoring and feedback continue to be a crucial part of transition support for youth even
after they have left the facility. The transition
team should continue meeting with regularity throughout the post-release transition
(Stephens & Arnette, 2000), and youth progress should be continually monitored, with
plans adjusted based on needs and problems
that arise (Barton, 2006; Risler & O’Rourke,
2009).
Role of Law Enforcement Personnel
and Transition Coordinators
Upon Release from a Juvenile
Correctional Facility
For an adolescent involved in the juvenile
justice system, returning to the community
after confinement can be more difficult than
the confinement itself (Clark & Unruh, 2010).
This section underscores the importance of
approaching transition support as a coordinated set of services between multiple settings and multiple professional disciplines
and agencies. It is important to first describe
the role of each professional involved with
Figure 1. Multidisciplinary Team Approach to the Transition Process
This figure depicts the unique role of each professional and the recommended overlap between all professionals
involved in the adolescent’s reintegration to the community.
the adolescent at the point of release from the
correctional facility. As previously stated, the
NCLB and the IDEA require correctional facilities to provide transition services to adolescents under their care. Accordingly, Figure 1
shows the transition coordinator as central in
the transition process.
Role of Transition Coordinator
After an adolescent is released from custody,
the role of the transition coordinator is central to successful community reintegration
and, therefore, central to reducing recidivism
(Unruh, Gau, & Waintrup, 2009; Waintrup
& Unruh, 2008). The transition coordinator
is hired by the correctional facility and is the
professional best trained to coordinate and
monitor transition support services between
different community service providers. The
role of a transition coordinator is multifaceted, requiring the professional to function as
guidance counselor in the facility and advocate for each adolescent outside of the correctional facility. As an advocate, the key role of
a transition coordinator at the point of release
is to ensure that communication with all service providers in the community are initiated
when the adolescent leaves confinement. The
role of transition coordinator is analogous to
that of a project manager in industry—that is,
someone who oversees the activities of professionals from various disciplines who have
the same goal, namely, keeping the youth
from recidivating. It is the responsibility of the
35
transition coordinator to ensure that records
from the facility reach the appropriate service
providers. If the adolescent is returning to
school, then the transition coordinator must
contact educators well in advance to give
notice that the student is returning to school on
a given date. The transition coordinator must
also ensure that documents from the facility
relevant to reentry are available to teachers
and school administrators. For example, the
school needs records of academic accomplishments the adolescent completed while in
custody. Some of the information educators
have said they need for students returning to
school include progress notes, assessments
conducted while incarcerated, and a description of the student’s level of motivation and
effort to complete work (Macomber et al.,
2010).
Two research projects provide practical
descriptions of the role of transition coordinators. In Project SUPPORT (Service Utilization
to Promote the Positive Reintegration and
Community Transition of Incarcerated Youth
with Disabilities), the transition coordinator
worked directly with the youth and other relevant professionals in the correctional facility to develop a transition plan (Unruh et al.,
2009; Waintrup & Unruh, 2008). Additionally,
the transition coordinator collaborated with
the parole officer to develop a complementary parole plan. The transition coordinator in Project SUPPORT was responsible for
coordinating the process of sharing information between different settings and agencies
(e.g., educational personnel in the facility and
the community, law enforcement personnel,
medical professionals, and employers). Karcz
(1996) describes a transition coordinator using
a different term, Youth Reentry Specialist
(YRS), but provides a similar description of
the role the YRS accomplishes in providing
transition support. The two-year research
project, funded by the U.S. Department
of Education and Rehabilitation Services,
made use of a specially trained professional
to provide transition services to youth with
disabilities leaving a correctional facility
36
in Wisconsin. The primary responsibilities
of the YRS (transition coordinator) were to
(1) determine re-entry procedures from the
correctional school to the special education
unit in the community; (2) obtain vocational
resources available in the community, including information about vocational program
requirements and state-level funding opportunities; (3) obtain permission from the juvenile parole officer to enroll the student in the
special education program and provide support to the juvenile parole officer and the adolescent’s parents afterwards; and (4) provide
technical assistance and information about
funding opportunities to the adolescent, his
or her family, and all related service providers. The YRS was housed and employed by
a school instead of a law enforcement unit.
To summarize, a transition coordinator is the
person responsible for developing and coordinating a re-entry plan for adolescents upon
discharge from the juvenile correctional facility and for monitoring the execution of that
plan post-discharge.
Role of Juvenile Parole/Probation Officer
The traditional and primary role of any law
enforcement officer has been to monitor and
respond when laws are broken. In the case
of juvenile delinquency, the relationship
between juveniles who commit crimes and
juvenile justice officers, whose responsibility
it is to enforce the law, has been, by and large,
antagonistic (Ochoa & Rome, 2009). Parole
or probation officers work as a part of either
their local juvenile court system or under the
administration of departments managed at
the state level (Torbet, 1997). The title and
responsibilities of parole officers and probation officers varies greatly from state to state
and jurisdiction to jurisdiction. Parole officers
are often thought of as the professionals who
provide court-ordered supervision following
a period of incarceration or long-term outof-home placement, while probation officers
are the professionals who supervise youth
as an alternative to out-of-home placement.
However, the titles of probation officer and
parole officer are often interchangeable for
adolescents caught up in the juvenile justice
system. In some states, parole and probation
officers are housed in the same department.
For example, in Louisiana, the official title
of professionals who provide court-ordered
supervision to juveniles is Probation and
Parole Officer or PPO (State of Louisiana, n.d).
It is estimated that there are 18,000 probation
officers providing services to juveniles in the
U.S. (Torbet, 1997). Probation officers have
the challenging and somewhat conflicting
task of balancing setting limits and monitoring adherence to rules and laws while also
providing social service type functions. The
various roles of the probation officer fall into
three broad categories: (1) law enforcement,
(2) social service, and (3) resource broker
(Rudes, Viglione, & Taxman, 2011). In the
category of law enforcement, responsibilities
include enforcing the terms of probation and
punishing noncompliance, with the overarching goal of protecting the community
(Rudes et al., 2011). As a social service agent,
the probation officer’s focus is on rehabilitation through case management. Their job is to
determine what services need to be in place
to help youth successfully reintegrate into
their communities. As a resource broker, it is
the probation officer’s job to connect youth to
those resources and services within their communities that are needed for them to transition
successfully out of confinement (Rudes et al.,
2011). In practice, these services may range
from electronic monitoring, surveillance, and
drug screenings to completing risk and needs
assessments, developing reintegration plans,
contacting local community mental health
centers to arrange for therapeutic services
upon return, and coordinating aftercare meetings between family, probation, and community-based service providers. Probation
and/or parole officers should be an integral
part of any transition planning and should
work closely with the transition coordinator to help ensure that youth are provided
with the services and resources they need.
Role of School Resource Officer
The school resource officer (SRO) represents
efforts on the part of police departments
to work proactively with schools to deter
crime and support juveniles to reach their
full potential (National Association of School
Resource Officers [NASRO], 2012). The SRO
is employed by the community’s police
department and is a law enforcement officer
charged to collaborate with schools and community-based organizations. The main goal of
the SRO is to keep order in schools (Omnibus
Crime Prevention Control and Safe Streets Act,
1968). The presence of SROs in public schools
is not without criticism (NASRO, 2012), with
critics arguing that schools should implement
discipline not law enforcement. Nonetheless,
data provided by NASRO indicates that the
presence of an SRO is effective at reducing
violence in schools (NASRO, 2010, 2012).
According to NASRO (2012), the role of the
SRO is not merely to serve as a police officer
who is based in a school. Through the Triad
Model of training, SROs are part educators,
part informal counselors, and part law enforcers. Duties of the SRO include, but are not
limited to, delivering information gathered
from the home or community to school principals at the start of the school day, meeting
with social workers to provide direct support
to an adolescent who is exhibiting disruptive
behavior, providing information to students
about violence and prevention programs, and
developing and implementing interventions
aimed at developing adaptive skills in adolescents (NASRO, 2012). Ochoa, Otero, Levy,
and Deskalo (2013) have suggested that the
current function of SROs can be strengthened
by redefining their primary function away
from punitive enforcers of the law to that of
a gentler adult whose purpose in schools is
to help adolescents learn to abide by the law.
With regard to recently released adolescents,
the SRO stands to provide reinforcement of
pro-social behavior as the adolescent is reintegrated into the school and community.
37
Role of Adolescent
Research shows that incarceration during
adolescence often leads to inadequate preparation for young adulthood (Abrams, 2006).
Research also shows that the period of transition, defined as the month before release to
six months after release, from a juvenile correctional facility to the community is important (Baltodano et al., 2005b). Yet, the role of
adolescents in custody at the point of transition is typically passive. That is, they often
do not participate in the planning efforts of
their own discharge. Clinkinbeard and Zohra
(2011) point out that plans to incentivize adolescents to be more involved in their transition are not yet a universal priority in juvenile corrections. Our own research in juvenile
correctional facilities has shown that many
adolescents in custody are enthusiastic about
leaving confinement, but few of them are
meaningfully involved in their own transition plans. Transition is approached in very
much the same way as their custody proceeded: they are confined and expected to
follow rules imposed on them by the juvenile
justice system, the correctional facility, and
the educators or service agents who work
with them. Little is known about adolescents’
thoughts, fears, and perceptions of the transition process. We agree with Clinkinbeard and
Zohra’s (2011) suggestion that increasing the
adolescent’s participation in their transition
from custody to the community is a necessary
component of successful transition planning
because it encourages adolescents to accept
the plan that has been developed for them.
Recommendations to Improve
Transition Outcomes
State-level recidivism rates ranging from
55% (Snyder & Sickmund, 2006) to 50 to 70%
(Nellis & Wayman, 2009) indicate that the
majority, if not all, of the behavioral goals
met by many adolescents and the academic
progress made by some adolescents while
in custody are lost once they leave custody.
38
This section provides recommendations to
improve transition outcomes for previously
incarcerated youth.
Increase Adolescent Involvement
The IDEA outlines regulations in regard to
transition which call for the involvement of
adolescents in the decision-making process
for the development of their own educational
program. We recommend that the transition
coordinator and the parole officer include
the adolescent in developing the transition
and parole plans. The two studies that follow
provide examples of positive outcomes when
researchers and service providers sought
involvement from adolescents. Including the
adolescent in the development of his or her discharge plan serves to elicit greater buy-in and
compliance. Clinkinbeard and Zohra (2011)
surveyed 543 incarcerated adolescents (384
males; 159 females), ranging from 12 to 22
years of age. Their goal was to understand
the adolescents’ future goals and determine
their levels of motivation, a factor linked to
successful reentry into a community. The
researchers posed four related questions
about the person a youth wanted to become:
(1) Next year, I expect to be . . .; (2) Am I doing
something to reach my goal of what I want
to be? (3) Next year, I fear I will be . . .; and
(4) Am I doing something to avoid becoming
the person I fear? The researchers concluded
that while plans do not automatically convert
into outcomes, youth with plans found more
success in community reentry than did adolescents without plans.
Silesky (2014) conducted a study to involve
adolescents in Hogar San Augustin, a residential placement in Costa Rica for children
who are wards of the state and who are at
risk for incarceration, in their own programming. A self-administered survey, including a combination of Likert scale items and
open-ended questions, was used to collect
data on what the 16 youth under their care
liked about the academic and vocational services, recreational activities, and areas that
are of concern to them or that they believe
need improvement. The residential facility
plans to continue to analyze results from the
survey to make programmatic adjustments as
deemed necessary. The reason to include the
results along with the research conducted by
Clinkinbeard and Zohra (2011) in this article
is to highlight that professionals who work
with delinquent youth, or youth at risk for
delinquency, are beginning to see the important role adolescents play in planning for their
future as young adults.
Increase Overlap Between Juvenile
Justice Personnel and Law Enforcement
Officers
This article describes the role each professional plays in the transition process. Figure
1 underscores the importance of increasing
the overlap between transition coordinator, parole officer, and SRO—all professionals employed by a law enforcement agency.
Stephens and Arnette (2000) and Barton (2006)
advise that transition support should be
delivered under the guidance of a community-based service provider rather than an individual based out of the juvenile justice system
like a probation officer. We recommend the
expansion of communication, cooperation,
and collaboration between the professional in
each agency who works with the individual
youth. Communication on the status of the
adolescent will prevent a silo effect, in which
one professional is not included and is not
contributing insights and suggestions which
could help that adolescent. We appreciate the
value of Barton’s (2006) recommendation to
move away from management of the transition process by a law enforcement professional to having management supervised by a
transition coordinator employed by a non-law
enforcement agency. However, we recognize
that this recommendation involves a cultural
shift in thinking and may not be possible at
the present time. As indicated previously, the
roles of parole officers include law enforcement, social service, and resource brokerage
(Rudes et al., 2011). Similarly, SROs are part
educators, part informal counselors, and part
law enforcers (NASRO, 2012). Instead, we
encourage law enforcement agencies to continue to expand the training of juvenile parole
officers and SROs to emphasize the social
service role and to appreciate the way emotional and learning disabilities affect adolescent behavior and their tendency to commit
criminal acts.
Further, we believe it is appropriate for the
transition coordinator to guide and be responsible for the creation of the transition plan
while the adolescent is incarcerated. We also
recognize the value in shifting the management and supervision of the transition plan
to the juvenile parole officer once the juvenile
exits the correctional facility. In an ideal situation, this hand-off process would be possible and efficient because there would be a
formal communication and collaboration policy
between the transition coordinator and the
juvenile parole officer during the period just
prior to release from the correctional facility.
The goal is to allow the multidisciplinary team
of professionals to continue to work with the
adolescent within the scope of each of their
professions and to have a smooth transition
of both the adolescent and of responsibility
for the adolescent and his or her treatment via
the transition plan.
Formalize Communication Between All
Transition Support Service Providers on
the Transition Team
As indicated previously, best practice guidelines recommend that transition support be
approached as a team effort instead of relegating all of the responsibility to a single individual. Barton (2006) promotes a formalized
approach to communication between all agencies and professionals involved in the transition process. For example, the use of formalized transition manuals created by the juvenile
correctional facility transition coordinator can
include a checklist indicating the actions that
need to be taken, specifically naming each
professional and the activity for which they
39
are responsible. The checklist should include
the date the action was accomplished and any
relevant information that needs to be communicated to the entire team. Formalizing procedures is important to minimize the likelihood
that an important component of the transition
process might fall through the cracks because
someone assumed that someone else was
responsible for a required action.
Promote a More Proactive and Positive
Approach to the Professional Preparation
of Juvenile Law Enforcement Personnel
Charged to Work with Juveniles
We acknowledge that juvenile delinquents
engage in antisocial behavior, and ignoring
their maladaptive behavior is not a solution for deterring them from further delinquency. We urge shifting away from reactive
and punitive responses to acts committed by
delinquent youth, which have been proven
ineffective in decreasing delinquency (Ochoa,
& Rome, 2009). Instead, we promote the adoption of non-punitive, positive approaches to
rehabilitation which, as described by Ochoa
and Rome (2009), support the development
of adaptive social skills. Law enforcement
agencies have already started to move toward
non-punitive measures. However, we also
acknowledge that there is a large gap between
intended role and actual practice. It is a fact
that many juveniles are exposed to additional
violence during incarceration. Some of the
violence is initiated by the youth. Yet, a good
portion of violence, even if it is initiated by the
adolescent, may be the result of an unhealthy
culture which condones heavy-handed punitive measures in juvenile correctional facilities that is allowed by the adults who interact
with juveniles in custody.
In conclusion, juvenile delinquency and the
high rate of recidivism is an important societal problem requiring involvement of individuals from many professional disciplines
and in various settings. However, once an
adolescent is involved with the criminal justice system, and particularly after they have
40
served a sentence for a crime, law enforcement professionals play a pivotal role in transitioning juveniles from custody to the community. Community-based law enforcement
agencies already have juvenile parole officers
and SROs whose job it is to deter youngsters
from crime and delinquency. Similarly, juvenile correctional facilities employ transition
coordinators whose primary job responsibility is to develop a plan of action when adolescents exit custody. Approaching the process
of transition from custody to the community
from a multidisciplinary, collaborative perspective and forging strong communication
ties between the professionals involved in the
transition process will lead to better reintegration of the juvenile into his or her community
and reduce the likelihood of juvenile recidivism for violations of parole or new crimes.
Acknowledgment
We thank Sarah Swank for providing organizational assistance in the process of writing the manuscript. The project was funded,
in part, by the Faculty Research Support
Program at Indiana University.
References
Abrams, L. (2006). From corrections to community: Youth offender’s perceptions of the
challenges of transition. Journal of Offender
Rehabilitation, 44(2/3), 31-53.
Anthony, E. K., Samples, M. D., de Kervor,
D. N., Ituarte, S., Lee, C., & Austin, M. J.
(2010). Coming back home: The reintegration of formerly incarcerated youth with
service implications. Children and Youth Services Review, 32(10), 1271-1277.
Archwamety, T., & Katsiyannis, A. (2000).
Academic remediation, parole violations,
and recidivism rates among delinquent
youth. Remedial and Special Education, 21(3),
161-170.
Baltodano, H. M., Mathur, S. R., & Rutherford,
R. B. (2005a). Transition of incarcerated
youth with disabilities across systems and
into adulthood. Exceptionality: A Special Education Journal, 13(2), 103-124.
Karcz, S. A. (1996). An effectiveness study of
the Youth Reentry Specialist (YRS) program
for released incarcerated youth with handicapping conditions. The Journal of Correctional Education, 47(1), 43-46.
Baltodano, H. M., Platt, D., & Roberts, C. W.
(2005b). Transition from secure care to the
community: Significant issues for youth in
detention. The Journal of Correctional Education, 56(4), 372-388.
Leone, P. E., Krezmien, M., Mason, L., &
Meisel, S. (2005). Organizing and delivering empirically based literacy instruction to
incarcerated youth. Exceptionality: A Special
Education Journal, 13(2), 89-102.
Barton, W. H. (2006). Incorporating the
strengths perspective into intensive juvenile
aftercare. Western Criminology Review, 7(2),
48-61.
Macomber, D., Skiba, T., Blackmon, J., Esposito,
E., Hart, L., Mambrino, E., . . . Grigorenko,
E. L. (2010). Education in juvenile detention facilities in the state of Connecticut: A
glance at the system. The Journal of Correctional Education, 61(3), 223-261.
Cavindish, W. (2013). Academic attainment
during commitment and postrelease education-related outcomes of juvenile justice-involved youth with and without disabilities. Journal of Emotional and Behavioral
Disorders, 20(2), 1-12. http://dx.doi.org/
10.1177/1063426612470516
Clark, H. G., Mathur, S. R., & Helding, B.
(2011). Transition services for juvenile
detainees with disabilities: Findings on
recidivism. Education and Treatment of Children, 34(4), 511-529.
Clark, H. G., & Unruh, D. (2010). Transition
practices for adjudicated youth with E/BDs
and related disorders. Behavioral Disorders,
36(1), 43-51.
Clinkinbeard, S. S., & Zohra, T. (2011). Expectations, fears and strategies: Juvenile offender
thoughts on a future outside of incarceration. Youth and Society, 44(2), 236-257.
Individuals with Disabilities Education Act, 20
USC § 1400 (2004).
JustChildren. (2006). A summary of best practices in school reentry for incarcerated youth
returning home. Submission to the Commonwealth of Virginia Board of Education,
Charlottesville.
Mears, D., & Aron, L. (2003). Addressing the
needs of youth with disabilities in the juvenile
justice system: The current state of knowledge.
Washington, DC: Urban Institute.
National Association of School Resource
Officers (NASRO). (2010). To protect and
educate: The school resource officer and the
prevention of violence in schools. Retrieved
from https://nasro.org/cms/wp-content/
uploads/2013/11/NASRO-To-Protect-andEducate-nosecurity.pdf.
NASRO. (2012). Class training. Retrieved from
www.nasro.org/class-training.
Nellis, A., & Wayman, R. H. (2009). Back
on track: Theory, research, and promising
approaches for youth reentry. Washington,
DC: Youth Reentry Task force of the Juvenile Justice and Delinquency Prevention
Coalition.
New America Foundation. (2012, March 28).
Federal education budget project. Retrieved from
http://febp.newamerica.net/k12/rankings/
ppexpend.
Ochoa, T. A., Otero, T. L., Levy, L. J., &
Deskalo, A. Y. (2013). Integration of the
41
school resource officer as a liaison between
law enforcement and school administration
in the discipline of students. Law Enforcement Executive Forum, 13(2), 129-144.
Ochoa, T. A., & Rome, J. (2009). Considerations for arrests and interrogations of suspects with hearing, cognitive, and behavioral disorders. Law Enforcement Executive
Forum, 9(5), 125-135.
Ochoa, T. A., & Spegel, K. M. (2015). Transition
from juvenile confinement: A literature review.
Manuscript in preparation.
Omnibus Crime Prevention Control and Safe
Streets Act, PL No. 90-351, §1709, 82 Stat. 197
(1968).
Osher, D., Amos, L. B., & Gonsoulin, S. (2012).
Successfully transitioning youth who are delinquent between institutions and alternative and
community schools. Retrieved from www.
neglected-delinquent.org/sites/default/
files/docs/successfully_transitioning_youth.
pdf.
Puzzanchera, C. (2013). Juvenile arrests 2011.
Juvenile Offenders and Victims: National
Report Series Bulletin. Retrieved from www.
ojjdp.gov/pubs/244476.pdf.
Puzzanchera, C., & Kang, W. (2014). Easy access
to FBI arrest statistics: 1994-2012. Retrieved
from http://ojjdp.gov/ojstatbb/ezaucr/asp/
ucr_display.asp.
Risler, E., & O’Rourke, T. (2009). Thinking exit
at entry: Exploring outcomes of Georgia’s
juvenile justice educational programs.
The Journal of Correctional Education, 60(3),
225-239.
Roy-Stevens, C. (2004). Overcoming barriers to
school reentry. Washington, DC: Office of
Juvenile Justice and Delinquency Prevention, Office of Justice Programs, U.S. Department of Justice.
42
Rudes, D. S., Viglione, J., & Taxman, F. S.
(2011). Juvenile probation officers: How the
perception of roles affects training experiences for evidence-based practice implementation. Federal Probation, 75, 3-10.
Sheldon-Sherman, J. A. L. (2010). No incarcerated youth left behind: Promoting successful school reentry through best practices
and reform. Children’s Legal Rights Journal,
30(2), 22-37.
Sheldon-Sherman, J. A. L. (2013). The IDEA of
an adequate education for all: Ensuring success for incarcerated youth with disabilities.
Journal of Law & Education, 42(2), 227-274.
Sickmund, M. (2010). Juveniles in residential
placement, 1997-2008. Washington, DC: Office
of Juvenile Justice and Delinquency Prevention, Office of Justice Programs, U.S.
Department of Justice.
Silesky, O. (2014). Percepciones de los jóvenes
del Hogar San Agustín en torno a los servicios
ofrecidos por la institución durante su tiempo de
permanencia en el Hogar [Youths’ perceptions
regarding services offered during their time
of residency at Hogar San Agustín]. San
Jose, Costa Rica. (Unpublished study)
Snyder, H., & Sickmund, M. (2006). Juvenile
offenders and victims: 2006 national report.
Washington, DC: National Center for Juvenile Justice.
State of Louisiana, Youth Services, Office of
Juvenile Justice. (n.d.). Probation & parole:
Duties and responsibilities of the probation and
parole officer (PPO). Retrieved from http://
ojj.la.gov/index.php?page=sub&id=32.
Stephens, R. D., & Arnette, J. L. (2000). From
the courthouse to the schoolhouse: Making successful transitions. Washington, DC: Office
of Juvenile Justice and Delinquency Prevention, Office of Justice Programs, U.S.
Department of Justice.
Suitts, S., Dunn, K., & Sabree, N. (2014). Just
learning: The imperative to transform juvenile justice systems into effective educational
systems. Atlanta, GA: Southern Education Foundation. Retrieved from www.
southerneducation.org/getattachment/
cf39e156-5992-4050-bd03-fb34cc5bf7e3/
Just-Learning.aspx.
Torbet, P. M. (1997). Juvenile probation: The
workhorse of the juvenile justice system.
Washington, DC: Office of Juvenile Justice
and Delinquency Prevention, Office of Justice Programs, U.S. Department of Justice.
Unruh, D. K., Gau, J. M., & Waintrup, M.
G. (2009). An exploration of factors reducing recidivism rates of formerly incarcerated youth with disabilities participating
in a re-entry intervention. Journal of Child
Family Studies, 18, 284-293. http://dx.doi.
org/10.1007/s10826-008-9228-8
Contact Information
Theresa A. Ochoa
School of Education
Indiana University
(812) 856-8135
[email protected]
Lawrence J. Levy
Florida Private Practice
[email protected]
Kelly M. Spegel
School of Education
Indiana University
[email protected]
Yanua F. Ovares
University of Costa Rica
[email protected]
Waintrup, M. G., & Unruh, D. K. (2008).
Career development programming strategies for transitioning incarcerated adolescents to the world of work. The Journal of
Correctional Education, 59(2), 127-144.
Theresa A. Ochoa is an associate professor
of Special Education at Indiana University.
Her research includes the preparation of
teachers of students with a range of cognitive and behavioral disorders and the
laws that govern the education and treatment of students with disabilities. The discipline of students with disabilities within
school settings and their treatment within
law enforcement is a specific area of her
research for which she intends to build
crossdisciplinary collaboration.
Lawrence J. Levy, PsyD, is a licensed psychologist in private practice in Boca Raton,
Florida. His research and clinical interests
focus on anxiety, impulsivity, and behavioral disorders in adolescents and adults.
43
Factoring Fatigue into Police Deadly Force
Encounters: Decision-Making and
Reaction Times
David M. Blake, MS, Blake Consulting and Training Group
Edward Cumella, PhD, Professor of Graduate Psychology, Kaplan University
Abstract
Significant evidence exists demonstrating the negative impact of fatigue on human cognitive performance
in such areas as decision making, reaction times, and memory. Law enforcement studies have shown that
officers suffer from high levels of fatigue from lack of sleep, unusual shift schedules, and exorbitant hours
awake; however, little empirical evidence exists directly relating the effects of fatigue to individual officer
performance in police specific tasks, particularly performance in deadly force situations. The current study
(N = 53) examined effects of fatigue, including total time awake (TTA), shift work, hours slept, and subjective sleep quality, on officers’ decision-making and reaction times when presented with simulated shoot/
don’t shoot and ambiguous target paradigms. The authors of this study hypothesized that fatigue would
negatively impact officers’ decision-making and reaction time accuracy. The hypothesis was confirmed in
that many of the fatigue measures correlated significantly with decreases in decision making in the deadly
force simulations and with increased reaction time. Specifically, poor sleep quality, greater TTA, more days
worked, and working night or swing shifts all decreased the accuracy of officers’ decision making, especially
when officers were presented with no-shoot and ambiguous scenarios. Greater TTA, more days worked, and
working swing shifts also increased officers’ reaction times during these deadly force simulations. Finally,
the effects of fatigue also increased throughout each work day, with officers’ reaction times increasing
consistently from their pre-shift assessment to their post-shift assessment. These findings have significant
implications for police performance in deadly force encounters, training, and scheduling.
A simple click of the mouse button while
searching through hundreds of YouTube
videos can provide single dimensional eyewitness views of United States law enforcement
officers in a multitude of unusual and deadly
situations. The most intriguing ones revolve
around officers’ decision making during rapidly evolving, dynamic, and highly stressful
incidents. Unfortunately, some of these incidents, about 0.2%, result in an officer’s most
powerful and devastating decision, the decision to use deadly force (Adams et al., 1999).
Also unfortunate is how society, through the
media, can sometimes misjudge these encounters based on limited information (Adams et al.,
1999; Johnson, 2007; Sharp & Hess, 2008).
44
Understanding police use of force in today’s
culture of unrelenting media access and personal video devices may require a paradigm
shift in how society looks at those who protect
and serve. As police are often held to scientifically deduced ideal human performance
standards that may be unattainable in real-life
encounters, they may be perceived as racist,
overly aggressive, or even murderous when
they fail to meet these standards (Johnson,
2007; Lewinski & Honig, 2008; Sharp & Hess,
2008). The idea of fallibility through human
performance error is rarely considered or
accepted when an officer uses lethal force.
It could be argued that social judgment
should and legal judgment must be derived
from solid empirical evidence. This evidence
would do well to provide a full accounting of
all the human factors involved and the context within which a specific situation exists,
along with close attention to the applicable
written legal codes.
To begin this accounting, a basic understanding of the aspects involved in a use-of-force
incident and how they pertain to the perspective of the officer must be compiled. This journey may begin in the legal realm. Miller (n.d.),
Branch Chief within the Legal Division of the
Federal Law Enforcement Training Center
(FLETC), states that assessment of a police
officer’s use of force is generally based upon
the 1989 Supreme Court decision, Graham v.
Connor (1989). In brief, the decision provides
that an officer’s use of force in any Fourth
Amendment seizure (i.e., an arrest or detention) should be judged based on an objective
reasonableness standard. The Court went on
to define the objective reasonableness standard as follows: “Would another similarly
experienced officer in a similar situation, utilize a similar amount of force while under
split second timing restrictions and operating
in “tense, uncertain, and rapidly evolving circumstances?” (p. 1).
The Supreme Court in Graham provided
the legal backdrop for a fair and balanced
approach by which police use of force can
be judged. However, to truly understand
the event as it was experienced by the officer involved, human factors relating to subjective perspective must be introduced. Yet,
significant uncertainty exists concerning the
human factors that officers may have experienced, are able to testify about, or even know
to exist within the situation. The aspects of
rapidly evolving violent encounters from the
perspective of an involved police officer are
unique and have not been previously studied
with the detail necessary to fully understand
them. Cognitive psychology and the study of
human factors in use-of-force encounters are
just beginning to close the gap of understanding (Honig & Lewinski, 2009).
Recent research provided insight into how
officers on the same scene might provide
differing accounts of an incident, how officers vary in threat perception, and even how
they might justifiably shoot a suspect in the
back (Blair et al., 2011; Lewinski & Honig,
2008; Lewinski & Hudson, 2003; Lewinski &
Redmann, 2009). Take for instance the case of
Randall Carr, who was shot and killed after
a deadly altercation with police officers. The
location of the fatal wound was questioned
because its placement in the buttocks meant
Carr was no longer a threat when shot by
officers. Ultimately, Dr. William Lewinski of
the Force Science Institute was able to adequately explain the human dynamics behind
the incident, and the officers were exonerated
by a jury (Force Science Institute, 2005). His
breakthrough research into the dynamics of
police-involved shootings demonstrated the
fluidity of a gunfight and provided an officer’s “stop shooting” reaction times, which
could result in a suspect justifiably being
shot in the back (Lewinski, 2000). Without
Lewinski’s research into the human factors of
use-of-force encounters, the officers may have
been held accountable for what appeared to
be an unjustifiable shooting when, in fact,
they were simply operating within the confines of human performance.
One human factor that has been ignored for
too long in policing is fatigue. Shift work,
court appearances, special assignments,
and the long hours officers usually work
have been suggested as contributing factors
to human error (Vila, Kenney, Morrison,
& Reuland, 2000). The concept of fatigue
is closely linked to sleep deprivation, and
these labels are often used synonymously
(Samkoff & Jacques, 1991; Sundelin et al.,
2013). A depth of research exists in the area
of sleep deprivation (SD) as it applies to several occupations, including policing, but its
application to specific police tasks has been
rather limited. Many of the other occupations
impact public safety and have been mandated
to have rest periods (Senjo, 2011). Policing,
which appears to have some of the poorest
45
working conditions in regard to sleep (Senjo,
2011; Vila et al., 2000), does not benefit from
such mandates. In fact, little regard is given to
the many sleep disturbances officers experience. SD is often deemed just “part of the job”
(Vila et al., 2000).
Bonnet and Arand (1995) discussed in great
detail what constitutes adequate sleep in their
literature review. The first prominent point
they make pertains to today’s society being
chronically sleep-deprived (sleep loss > 1 hour
nightly). They support an 8.5 hour standard
as being optimal and demonstrate that nightly
sleep lengths of 7.2 to 7.4 hours are deficient.
The authors stated that chronic sleep deprivation of less than 6.5 hours is potentially disastrous in regard to human performance.
A closer look should be taken at some nuances
of sleep deprivation because it may have different meanings depending on the type of
study involved. Following the definition supported by lead researchers (Dinges, Rogers, &
Baynard, n.d.; Durmer & Dinges, 2005; Lim &
Dinges, 2010), SD is simply a restriction of a
subject’s sleep to less than their usual amount
within any 24-hour period. SD can be as minor
as restricting a subject to 7 hours of sleep
nightly, which is the starting point for deficiencies in cognitive performance (Dinges et al.,
n.d.).
The U.S. Department of Health and Human
Services (DHHS) (2012) discusses the need
to combat SD through maintaining regularly
scheduled sleep habits of 7 to 8 hours daily
for most adults. Studies show that SD leads to
problems in many areas of human functioning, the most notable being deficits in decision
making, problem solving, attention, reaction
time (RT), and emotional control (Durmer
& Dinges, 2005; Rajaratnam et al., 2011; Vila
et al., 2000). Sadly, proof of the DHHS’s
pronouncements has been provided by the
National Highway Transportation Safety
Administration’s (NHTSA) (1996) report of
a yearly average of 56,000 traffic collisions
resulting in 1,550 fatalities occurring due to
46
driver fatigue. The NHTSA provides the main
characteristics of fatigued drivers as having
increased reaction times, attention deficits,
and a decreased ability to process information.
Durmer and Dinges (2005) performed an
extensive meta-analysis of the consequences
of SD. Their review began by discussing the
many vehicle-related accidents which occur
as a result of fatigue. The research suggests an
analogy between fatigued driver performance
and that of alcohol impaired drivers. Studies
have shown that drivers who are awake for 17
to 19 hours operate a motor vehicle with similar psychomotor skills to those with blood
alcohol content (BAC) between 0.05 and 0.1%,
with 0.08% being the typical legal definition
of driving while intoxicated (NHTSA, 2006).
Bryan Vila, Director of the Simulated
Hazardous Operational Task Laboratory at
the Washington State University Sleep and
Performance Research Center, and his colleagues (2000) have conducted studies on
police SD, which showed that 53% of U.S.
police officers receive less than the mean
amount of sleep needed per night. Study
results showed that 18% of officers experienced fatigue and a lack of motivation, and
another 16% stated they had trouble simply
staying awake on the job. Performance issues
related to this study were noted in the areas of
reduced patience, diminished decision-making capacity, decreased alertness, and slower
response times.
Neylan et al. (2002) conducted a study comparing subjective sleep quality in police officers, examining the effects of trauma exposure (critical incidents) and non-police routine
organizational stressors. The findings showed
that although officers suffer from trauma-related nightmares, the most significant
aspect affecting sleep quality was based in
the routine stressors experienced within the
non-trauma-related work environment.
Senjo (2011) researched 15 Western state law
enforcement agencies in the U.S. The study
provided self-reported sleep needs of 7 to 9
hours a night by 70% of the responding officers. However, two-thirds of the 70% reported
actual completed sleep ranging between 3 and
6 hours. Issues such as shift work, overtime,
secondary employment, and others were
listed as reasons for officers having insufficient sleep.
Rajaratnam et al. (2011) conducted a critical
study of 4,957 police officers from across the
U.S. and Canada. The research involved both
online surveys and onsite interviews. Results
showed that 40% of those tested suffered from
at least one sleep disorder. Of those suffering
from a sleep disorder, 6.5% suffered from
moderate to severe insomnia, and 5.4% tested
positive for shift-work disorder. Results also
showed that those who suffered from sleep
disorders more often reported having made
administrative errors, falling asleep while
driving, and committing safety violations due
to fatigue.
Beyond the previously discussed fatigue
issues is the concept of chronic partial sleep
loss, often called sleep debt, which has also
been shown to affect alertness and performance (Barger, Lockley, Rajaratnam, &
Landrigan, 2009). Sleep debt is often discussed in terms of its cumulative effect. Using
a simple example, cumulative sleep debt is
the total amount of time, typically measured
in hours, over a specified period in which the
required sleep was not achieved. Van Dongen,
Maislin, Mullington, and Dinges (2003) conducted a study restricting the sleep of 48 participants to either 6 hours or 4 hours over 14
days. Tests such as the Psychomotor Vigilance
Test (PVT; Dinges & Basner, 2011) and the
Stanford Sleepiness Scale (SSS; Hoddes,
Zarcone, Smythe, Phillips, & Dement, 1973)
were administered. The PVT measures alertness through required sustained attention
while requiring quick reactions to random
stimuli; it has been deemed highly reliable
with test results comparable to real-world
behaviors (Dorrian, Rogers, & Dinges, n.d.).
The SSS is a subjective test with demonstrated
validity that can reliably determine levels of
sleepiness in an individual. The results of
the SSS have been found to correspond significantly with performance on tasks related
to SD (Hoddes et al., 1973). Within this sleep
study, significant differences existed in both
the 6- and 4-hour sleep groups in comparison
to the 8-hour group, indicating deficits for
the 4- and 6-hour groups. Of importance for
this study was the result showing that sleep
restriction to 4 hours over 14 days resulted in
working memory and alertness levels equivalent to those in persons who had not slept for
two days. Those in the 6-hour group showed
deficits comparable to one day without sleep.
Thus, the empirical evidence from this study,
combined with the BAC comparisons, suggest that persons who suffer from chronic
cumulative sleep debt could be functionally
equivalent to highly intoxicated individuals.
Another study concerning cumulative sleep
debt shows that even minor restrictions, such
as one hour per night, can cause performance
deficiency (Belenky et al., 2003). Belenky et al.
(2003) conducted a study in which 66 participants were sleep deprived at levels from 3
to 7 hours over 7 days. The study utilized the
PVT and SSS to measure sleepiness and performance four times per day. Although the
7-hour group did not report having increased
sleepiness (SSS), they did show significant
decreases in RT performance within PVT
results.
Couyoumdjian et al. (2009) discussed realworld decision making by considering the
circumstantial uniqueness in which decisions
often occur. Some circumstances, such as
those of policing, require innovative thinking, distraction avoidance, ignoring irrelevant stimuli, and following unfolding events,
all of which are negatively impacted by SD.
These executive-level functions were assessed
through a task switching stimulus test. The
results indicated that one night of total SD
negatively impacted the participants’ ability
to shift between two different cognitive tasks.
This information is significant in that officers
47
are often required to switch between tasks,
especially in use-of-force situations.
From the available literature, there is little doubt
that police officers work within a SD occupation and are ultimately exposed to SD at levels
which have adverse effects on human performance (Alhola & Polo-Kantola, 2007; Antal,
1975; Barger et al., 2009; Couyoumdjian et al.,
2009; Durmer & Dinges, 2005; Edwards &
Waterhouse, 2009; Lewinski & Honig, 2008;
NHTSA, 1996; O’Brien et al., 2012; Rajaratnam
et al., 2011; Senjo, 2011; Vila et al., 2000).
Nevertheless, the specificity of research into
fatigue and police performance with the use
of firearms (i.e., deadly force) is lacking. A
few studies are suggestive, however.
Edwards and Waterhouse (2009) conducted
an experiment showing the effects of SD on
the ability to throw darts. This study provided
intriguing results because of its simplicity and
the demonstrated effects of relatively little
SD. Sixty participants were deprived of four
hours sleep on just one night and then asked to
throw darts at a dart board. Deficiencies were
noted in accuracy and reliability. These deficiencies increased as the subjects were tested
over a span of several hours after awakening.
Antal (1975) conducted a study of circadian
rhythm disruption and its effects upon competitive shooters. Although his data are rudimentary and provide no specific hours of SD
or levels of fatigue, he shows a correlation
between the interruption of natural sleep
cycles and accuracy with firearms. His study
reported that shooters with SD suffered from
an inability to concentrate, complaints of
fatigue, and a lack of vitality. Another study
involving SD of 22 hours on a range of shooting skills was completed with a group of 20
military subjects. The study revolved around
the effects of caffeine and performance, but
it provided solid SD data in several areas.
This military study concluded that SD causes
deficits in RT to engagement and accuracy of
shot placement (Tikuisis, Keefe, McLellan, &
Kamimori, 2004).
48
The effect of SD on human beings operating in
various environments appears clear. A stack
of empirical evidence shows lack of sleep
causes poor attention, errors in judgment and
decision making, and a slowing of reaction
times. There also exists a rather convincing
amount of evidence, indicating that SD exists
at significant levels within law enforcement.
The last dot to connect is between the stated
effects of SD and police use-of-force incidents.
Bill Lewinski of the Force Science Institute
has collected scientific studies and discussion papers suggesting such a link. Based on
reviews, Lewinski and Honig (2008) summarize many of the human cognitive dynamics
of police use-of-force encounters and point to
attention, perception, decision making, pattern recognition, and action/reaction time as
having much to do with officers’ successfully
overcoming violent encounters and making
correct decisions. Importantly, these very
cognitive functions are negatively affected by
SD (Alhola & Polo-Kantola, 2007; Barger et al.,
2009; Couyoumdjian et al., 2009; Van Dongen
et al., 2003).
Thus, a confluence of research in the areas of
SD, SD in policing, and use of force provides a
glimpse of the potential deadly consequences
created by a combination of SD-driven factors. SD is prevalent in policing, yet the very
cognitive functions that are so necessary for
attending to and ultimately making the correct decision in use-of-force environments are
decreased by SD. The need to more carefully
examine the association between use-of-force
decision making and SD is therefore surely
necessary. Similar research specific to other
professions has revealed serious deficiencies,
prompting laws and regulations governing
these fields (NHTSA, 1996).
The literature review provided empirical evidence to support the following hypothesis,
which is addressed in the current research
investigation. To begin, police officers’ job
demands likely create an environment of SD
through several means, most prominently the
disruption of the circadian rhythm resulting
from shift work, a continuously accumulating
sleep deficit, and excessive total hours awake.
Overall cognitive abilities in the areas of information processing, decision making, reaction
time, and attention are negatively affected
by SD. Thus, it is hypothesized that these
effects of SD will have a negative impact on
officers’ decision-making capabilities during
shoot/don’t shoot scenarios. Specifically,
reaction time has been found to be negatively
impacted by SD and is expected to be affected
in this study through slower reaction times to
shoot/don’t shoot scenarios among officers
experiencing SD, with officers’ ability to react
quickly to perceived threats and to correctly
identify a shoot target being decreased by SD.
Methods
Participants
Participants were police officers from several
national police departments (N = 53): 50 men
and 3 women, ages 25 to 54, M = 40 (SD = 7.8)
years. Due to the specialized nature of this
study, participants were all experienced police
officers having completed basic police and
recurring inservice training concerning police
use of deadly force. Participating officers
were sampled from all shifts: Day Shift, n
= 17; Swing Shift, n = 21; and Night Shift,
n = 15. Following other studies on SD, participants were aware of the purpose of the study
as it has been determined such studies are
valid under these conditions (e.g., Edwards
& Waterhouse, 2009; Tikuisis et al., 2004;
Williamson & Feyer, 2000).
An additional and separate sample of police
officers from across the country (N = 277) completed a 10-question online Fatigue Survey.
The participants were gathered from electronic posts in police-specific online groups
such as the California Association of Force
Instructors, Law Enforcement Professionals,
and the Officer Involved Use of Force Group,
which are all hosted by LinkedIn (www.
linkedin.com), a professional networking
website. To protect the anonymity of these
officers and encourage their honesty about
a potentially difficult subject—the effects of
fatigue on their own police work—absolutely
no descriptive information about the participants was collected.
Procedure
An online electronic platform was created
based on previous studies and validated
measures. The online platform was named
the Thesis Computer Program (TCP) and has
built-in parameters to compensate for the lack
of an in-laboratory testing environment. These
parameters ensured participants logged in as
required, completed all prescribed tests, and
completed those tests within specified limits.
The TCP’s design favored validity over user
friendliness to provide the best outside of a
laboratory results.
Prior to engaging in the study, participants
were provided with an overview of the purpose of the study and introduced to its methodology. An online introduction to the TCP
followed in which in-depth instructions and
screen shots were provided. All information
provided prior to testing remained available
to participants throughout the duration of the
study. Additionally, participants had continued access to the facilitator to answer questions or resolve issues.
Participants were provided a URL which
allowed them access to the TCP online. Upon
entering the site, participants were required
to create an account using a typical password
and username security combination. The registration process included a request for certain non-identifying personal information
such as age, gender, shifts worked, and days
in the work week. Although all areas listed
were self-explanatory, the wide range of shift
definitions across law enforcement required
independent definitions of day, swing, and
night shifts. Day Shift was defined as most
duty hours during daylight. Swing Shift was
defined as half duty hours in daylight and
49
half during darkness. Night Shift was defined
as most duty hours during darkness.
During the registration process, the consent
form was displayed on the page, and participants were unable to register without selecting
a “consent/register” button, allowing them
to move forward. Participants received a validation e-mail to the address they provided
and were required to activate their accounts
through a link provided to that e-mail account.
Activating the account allowed participants
to have access to the task sections of the TCP.
Once registered, participants were asked to
log into the TCP at the beginning of their duty
week. They were required to log-in as close
to the beginning and end of each individual
work shift as possible. Strict adherence to the
research design was required, and participants were aware of parameters invalidating
any improper actions or inputs by the user.
Participants understood that a failure to complete all tests on all log-ins would result in
nullifying that day’s data.
The TCP itself adhered to a very strict set of
guidelines which allowed users little leeway
to operate outside its design. Participants
were guided step by step through the online
testing platform by displayed instructions
as well as automated movement to the next
task after the former task was completed.
The TCP did not allow for log-ins outside
of certain parameters, such as a mandated 8
hours between shifts or a requirement to log
in for post-shift tasks within 24 hours of the
pre-shift log-in. Participants were unable to
log-in for post-shift task completion without
having first signed in for pre-shift completion.
Sleep Diary
The first task required for each log-in to the
TCP was the completion of a sleep diary.
Participants completed a sleep diary for the
three days prior and all four days of the testing cycle. The sleep diary required the participants to enter the time they awoke each day,
50
the total hours of sleep prior to waking that
day, and their opinion about the quality of
their sleep.
Due to the subjective nature of asking participants whether or not they had a good or
bad night’s sleep, the TCP defined each category. A good sleep cycle was defined as an
“uninterrupted sleep cycle while awakening
well rested.” A bad sleep cycle was defined as
an “interrupted sleep cycle while awakening
poorly rested.” These definitions appeared on
each log-in to ensure consistency and validity. The TCP sleep diary task provided dropdown menus or restricted data entry points
(e.g., HH:MM) for each required response,
ensuring only the correct type of answer was
provided.
The sleep diary information was requested for
several reasons. The first is its ability to provide a static picture of changes in sleep patterns between duty days and non-duty days.
Additionally, time awake, hours slept, and
sleep quality are all key points of correlation
to the performance tasks within the study,
and they allow a determination of whether or
not these factors have any effect on reaction
times or decision making (Dinges et al., 1997;
Lim & Dinges, 2010).
Epworth Sleepiness Scale (ESS)
Participants were asked to complete the ESS
daily during both pre- and post-shift log-ins
to the TCP. Johns (2000) studied various sleepiness scales and demonstrated the ESS as the
most valid and reliable test available for measuring the appropriate amount of sleep. This
self-administered questionnaire (ESS) provides empirical evidence of whether or not
test subjects are fatigued. The instructions for
the ESS are very specific, yet simple, requiring
participants to subjectively rate their potential for “dozing” under a series of eight conditions. These standardized instructions were
provided in two places within the TCP as well
as reprinted on the TCP’s ESS data input page.
Additionally, input selections were limited by
drop-down menus to the standardized ESS
responses. The drop-down menus were an
additional method of ensuring validity in the
responses provided.
Baumann & DeSteno, 2010; Correll et al.,
2007; Dinges & Basner, 2011; Gartenberg &
Parasuraman, 2010; Lewinski & Hudson,
2003).
Psychomotor Vigilance Task (PVT)
Shoot/Don’t Shoot Situations (SDS)
The PVT has been mentioned often as a prevalent and simple testing measure to determine the effects of sleep loss upon reaction
speed and lapses (Alhola & Polo-Kantola,
2007; Dinges & Basner, 2012). Gartenberg
and Parasuraman (2010) conducted a study
testing the validity of a shortened “reaction test” using the iPhone/iPad platform,
with the application titled Mind Metrics. The
study provided evidence of validity in using
this 3-minute form of the PVT. Additionally,
other shorter duration PVT tests (i.e., 3 to 5
minutes) have demonstrated validity (Dinges
& Basner, 2012). The TCP included a version
of the 3-minute PVT due to the amount of
required testing sessions and the total time
required per log-in session.
Participants were required to complete the
PVT daily during both the pre- and post-shift
log-ins. Specific instructions were provided
to participants to ensure strict adherence to
the PVT methods. These instructions were
provided within the computer program and
had to be viewed before each test began. To
ensure validity, participants were required
to use their dominant hand middle finger
to perform the test. The hand was required
to be static, positioned just below the keyboard, helping to standardize the distance
the middle finger would be from the data
input device, which was the space bar. This
method enhanced both within-subject and
between-subject validity. In addition to the
physical restrictions asked of the participants,
the PVT had restrictions on acceptable reaction time results. The PVT does not accept RT
scores faster than 100 ms or slower than 1,500
ms to further ensure validity of captured
data. The reaction time parameters are similar
to other studies measuring RT under similar
circumstances (Adam, Bays, & Husain, 2011;
Correll et al. (2007) tested police use of force
decision making on several occasions through
the use of SDS computer analysis. The studies used a computer-based simulation displaying photographs of armed or unarmed
subjects in various settings. The photographs
remained on the screen for a short period
of time, between 500 and 850 ms, and were
intended to elicit SDS decisions from the subjects. Points were added and subtracted based
upon the decisions made by the subjects.
Using similar methodology, a SDS process
was incorporated into the TCP. The process,
similar to Correll et al. (2007), records data
from SDS displays and participant inputs
regarding reaction time and decision making
in response to the stimulus photo. Due to the
different nature of measurements within this
project, a minor change in the SDS platform
from Correll et al. (2007) was required as follows. Police officers spend a sizable portion
of their day involved in low stress tasks, but
when necessary, they are required to switch to
aroused status in reaction to threatening stimuli (e.g., from report writing to a radio call of
an in-progress crime). Likewise, police shooting situations are often unexpected and occur
in confluence with any number of other low
to highly arousing daily duties. To replicate
a realistic switch between arousal states, or at
least provide for a realistic cognitive distraction, the TCP displayed a simple math equation between SDS stimuli. The math questions required the participants to respond by
striking the spacebar for correct answers. The
math problem remained on the screen for 2 to
10 seconds prior to the display of each new
SDS stimulus. This intervening math event
was not present in Correll et al. (2007).
51
Per log-in, participants in the current study
viewed 12 of 62 randomized and encoded
SDS photographs upon a computer screen:
six shoot, three no-shoot, and three ambiguous
scenarios. SDS decisions were made through
standard keyboard input used for gaming:
A = shoot, L = don’t shoot. Prior to inclusion,
the photographs in this study were reviewed
by a panel of tenured police use-of-force
instructors. Photographs which received anything other than full agreement by the panel
were deemed ambiguous. Thus, all SDS photographs used have 100% inter-rater reliability
by tenured police use-of-force instructors, representing definitive SDS situations or ambiguous
situations.
Participants were required to complete the
TCP daily during both pre- and post-shift
log ins. To avoid practice effects that could
degrade the validity of the testing process,
two procedural actions were put in place. The
first was a randomization of the SDS photographs as to where each appeared during
each session. The randomization of the SDS
scenarios should ensure a lack of familiarity
with each stimulus. The second method of
avoiding practice effects is the sheer number
of scenario photographs, which were greater
than 60. A random viewing of 12 of 62 photographs over just four days should ensure a
lack of familiarity with each photograph as
no photograph was likely to appear several
times for each officer across the testing days.
Specific instructions were provided to participants to ensure strict adherence to the TCP’s
parameters. These instructions were provided
within the computer program and had to be
viewed before each test began. Participants
were required to place both hands below the
keyboard in a specific manner while having
their “point” fingers hovering above the A
and L keys. In addition to the physical restrictions asked of the participants, the TCP contained restrictions on acceptable reaction time
results. The reaction time parameters selected
were similar to those used in other studies
measuring RT under similar circumstances
52
(Adam et al., 2011; Baumann & DeSteno, 2010;
Correll et al., 2007; Dinges & Basner, 2011;
Gartenberg & Parasuraman, 2010; Lewinski &
Hudson, 2003).
Fatigue Survey
A survey created on the SurveyMonkey website contained 10 questions. Table 1 lists the
questions, all based on self-report, related to
sleep, performance, and agency oversight.
The answers were limited to “Yes” or “No.”
The purpose was to assess a separate sample
of police officers, untainted by their experiences completing the TCP, and obtain their
personal experiences with and views of the
effects of fatigue on their police performance.
Data Handling and Statistical Treatment
Participants were directed to the TCP website to complete their assessments. The TCP
options were set properly to ensure none of
the participants’ names, police agency names,
and IP addresses were collected. All results
were presented in aggregate form to further
protect subjects’ identities and confidentiality of information. Data were only accessible
through the online TCP system using a strong
password known only to the researcher.
Once the data collection was completed, data
were downloaded into Microsoft Excel and
then SPSS, stored only on the researcher and
advisor’s laptop computers, and deleted from
the online survey system. The SPSS database
used for data analysis was accessible only by
using a strong password known only to the
researcher and thesis advisor. Neither dataset
contained any coded identifiers and, as such,
both are completely anonymous.
The thesis chair and the student researcher
had access to the downloaded SPSS data. The
data were stored on the two computers owned
by these individuals. The data resided in separate Windows folders on each computer,
segregated from unrelated files. The two computers were locked by strong Windows passwords known only to the computer owners.
Table 1. Fatigue Survey Results
Measure
1. I believe shift work interferes with my ability to achieve a reasonably good
night of rest.
2. I have different sleep habits when I am not working as opposed to during
my work cycle.
3. I sleep much better on my days off as opposed to during my work cycle.
4. I believe lack of sleep has been the cause of a mistake or error I have made
while working.
5. I believe I perform better with more sleep.
6. I require about 8 hours of sleep to perform my best.
7. I believe I can perform adequately when required regardless of how many
hours I am awake.
8. I believe police departments should formally explore the impact of sleep
deprivation on officer performance.
9. I don’t want to explore aspects of sleep deprivation in police work because I
am concerned about a change in schedule or limitations on overtime.
10. I believe the law enforcement career field (in general) does not adequately
concern itself with safety issues concerning sleep deprivation.
The data were retained on these computer systems for the duration of the research, and, following completion of the research, they were
retained on the researcher’s computer for a
minimum of five years along with related files
in case questions arise about the analyses. The
dataset and related files will be transferred to
any future computer owned by the researcher
until the five years have expired. Throughout
the study and subsequent five years, the
researcher will implement a weekly backup
plan wherein the dataset and related files are
backed up using a secure online data backup
system. After the five years, the researcher
will destroy the SPSS data file using then-current Department of Defense data destruction
standards. An affordable technique, such as
encryption, will likely be chosen.
The various measures were scored according
to published norms. Then, the several independent variables, which were measures of
fatigue, were correlated with the outcomes of
the SDS scenarios—scenario by scenario and
in the aggregate. Patterns of correlations were
detected by extracting significant correlations
from the correlation matrix and presenting
% Yes
73.5
82.2
67.9
68.5
92.7
55.7
41.4
94.5
12.0
91.6
such in tabular format. Because the direction of each correlation was predicted by the
hypotheses in the study, alpha levels were
one-tailed, set at p < 0.10 for significance.
Results
In response to requests for participants printed
in the Police One Magazine and the Force Science
Institute Newsletter, 53 subjects completed the
study. It is not possible to know how many
subjects actually saw the research announcement in these two venues; as such, a response
rate cannot be calculated for this study. To
protect subjects’ anonymity, minimal sociodemographic data were collected; it appears
in Table 2. The mean subject age was 40; most
subjects were men. Table 2 also reveals that
subjects were fairly evenly distributed among
the three typical shifts worked in policing:
Days, n = 17; Swings, n = 21; Nights, n = 15.
As expected, participants slept more hours
on off-duty days than on-duty days: M = 6.8
hours vs. 6.4 hours. Participants were awake
between 15 and 17 hours at the completion of
each duty day (see Table 2 for details).
53
Table 2. Background Characteristics of Participants (N = 53)
Measure
Shift
Response
17 Day shift
21 Swing shift
15 Night shift
Age
Mean
Gender
Male
Female
Mean sleep hours per night
Off-duty
On-duty
Mean total time awake at posttest
Day 1
Day 2
Day 3
Day 4
40 (7.83)
50
3
6.8 (1.76) hours
6.4 (1.49) hours
17 hours
16 hours
16 hours
15 hours
Instrument Validity
Correlations for Day 1 were computed to determine instrument validity. A sizable number
of significant correlations occurred in the
predicted direction (see Table 3). Those correlations were moderate for SQ and ESS,
moderate to strong for PVT and SDS RTs, and
strong for ESS and SDS RTs. Aggregate means
for the ESS and PVT over the course of the
study also showed movement in the predicted
direction. Table 3 demonstrates that subjective reporting of fatigue increased from preshift to post-shift: ESS pre-shift, M = 6 (4.83),
and post-shift, M = 11 (6.27). Likewise, RT
increased from pre-shift to post-shift: PVT
pre-shift, M = 414 (63) ms, and post-shift,
M = 461 (87) ms. Table 3 includes data showing daily increases in PVT RT on all but one
(Day 4 post-shift) for both pre- and post-shift.
These results suggest good predictive validity
for the TCP instrument.
Decision Making (DM)
Table 4 displays the coefficients of all significant DM (i.e., shoot, don’t shoot, or ambiguous) correlations and the number of significant correlations occurring in the predicted
54
direction for each independent variable and
the several DM outcome variables. Subjective
reports of sleep quality (SQ) yielded 20 significant correlations in the direction of prediction over the course of four days. Days 1
and 4 provided the strongest correlations
(> 0.41), with Days 2 and 3 providing moderate correlations (0.26 to 0.40). Total time
awake (TTA) yielded 14 significant correlations in the direction of prediction on Days 3
and 4, with Day 3 providing the strongest significant correlations (e.g., 0.727).
Table 5 displays the type of significant DM
correlations (i.e., shoot, don’t shoot, or ambiguous). TTA and SQ produced six significant
results in the shoot scenarios. TTA and SQ
produced 19 significant no shoot results. TTA
and SQ produced 12 significant results among
the ambiguous scenarios.
Reaction Times (RT)
Table 6 displays the coefficients of all significant
RT correlations and the number of significant
correlations occurring in the direction of
prediction. TTA in relation to RT yielded nine
significant correlations in the direction of
Table 3. Correlations Suggesting Instrument Validity
Measure
SQ & ESS
Pre-shift
Post-shift
SQ & Shoot Response (Aggregate day 1)
Pre1NoShoot
Post1NoShoot
PVT & SDS RT Times
PrePVT/PreRT6
PrePVT/PreRT12
PostPVT/PostRT2
PostPVT/PostRT3
ESS & SDS RT Times
PreESS/PreRT4
PreESS/PreRT8
PostESS/PostRT9
Mean Psychomotor Vigilance Task (PVT)
Pre-shift
Day 1
Day 2
Day 3
Day 4
Post-shift
Day 1
Day 2
Day 3
Day 4
Pre-shift (aggregate)
Post-shift (aggregate)
Mean Epworth Sleepiness Scale (ESS)
Pre-shift (aggregate)
Post-shift (aggregate)
Day 1
0.363
0.369
0.311
0.408
0.338
0.436
0.397
0.361
-0.318
-0.282
0.424
411 ms
410 ms
436 ms
434 ms
437 ms
484 ms
486 ms
467 ms
414 (63) ms
461 (87) ms
6 (4.83)
11 (6.27)
Table 4. Correlations of Participants’ Reaction Times with Independent Variables
Measure
Total time awake
# Significant correlations in predicted direction
Shift
Days worked
Day 1
-0.039
0
-0.169
0
0.063
Day 2
0.449
5
-0.010
1
Day 3
0.009
1
-0.034
1
--
Day 4
0.643
3
0.602
2
0.557
1
55
Table 5. Participants’ Significant Decision-Making Types
Measure
Significant Correlations
TTA & Shoot response
SQ & Shoot response
Total
TTA & No shoot response
SQ & No shoot response
Total
TTA & Ambiguous response
SQ & Ambiguous response
Total
3
3
6
6
13
19
8
4
12
Table 6. Correlations of Participants’ Decision Making with Independent Variables
Measure
Sleep quality/Mean correlation
Total time awake/Mean correlation
Day 1
0.440
5
-0.846
1
Day 2
0.383
3
-0.067
2
Day 3
0.276
4
0.727
7
Day 4
0.697
8
0.342
7
Table 7. Aggregate Mean SDS Reaction Times
Measure
Pre-shift SDS RT
SD
Post-shift SDS RT
SD
Total combined SDS RT
Pre-shift
SD
Post-shift
SD
Day 1
719 ms
29.04
767 ms
29.68
Day 2
726 ms
21.15
738 ms
32.79
prediction. Days 2 and 4 produced strong
correlations (> 0.41), but Day 3 produced weak
correlations (< 0.20). The work shift assigned
showed strong correlations on Day 4, with
three significant correlations occurring in the
direction of prediction. The total days worked
also had one significant correlation moving in
the predicted direction on Day 4. Both work
shift and days worked correlations were
strong (> 0.41).
Table 7 displays the mean RT for the SDS with
all RTs moving in the direction of prediction.
56
Mean
709 ms
19.93
745 ms
16.22
Day 3
705 ms
27.50
746 ms
37.48
Day 4
683 ms
30.47
729 ms
37.86
SDS RT increased from pre- to post-shift on
Day 1: for SDS pre-shift, M = 719 (29) ms,
and post-shift, M = 767 (30) ms. SDS RT
increased from pre- to post-shift on Day 2:
for SDS pre-shift, M = 726 (21) ms, and postshift, M = 738 (33) ms. SDS RT increased
from pre- to post-shift on Day 3: for SDS
pre-shift, M = 705 (28) ms, and post-shift,
M = 746 (37) ms. SDS RT increased from
pre- to post-shift on Day 4: for SDS pre-shift,
M = 683 (20) ms, and post-shift, M = 729 (38) ms.
Aggregate mean RT for the SDS from Day 1 to
Day 4 increased between pre- and post-shift:
for aggregate pre-shift, M = 709 (20) ms, and
post-shift, M = 745 (16) ms.
Fatigue Survey
Table 1 presents the results of the fatigue
survey. Most of the respondents (74%)
believed that shift work interferes with their
ability to achieve a good night’s rest. Most
said they had different sleep habits on and off
duty (82%), with 68% stating they slept better
on their days off. The vast majority said they
perform better with more sleep (93%), and 69%
of respondents pointed to lack of sleep as a
causal factor in one or more mistakes or errors
which they had made while working. About
half of the respondents said they require 8
hours of sleep for optimal performance, with
a minority, 41%, believing they can perform
adequately regardless of how many hours
they are awake. Almost all believed that the
law enforcement career field does not adequately concern itself with safety issues arising from sleep deprivation (92%). Likewise,
95% of respondents stated that police departments need to formally explore the impact
of sleep deprivation on officer performance.
A mere 12% of respondents did not want to
explore sleep deprivation research within law
enforcement due to concerns about changes
in scheduling or limitations on overtime.
Discussion
The authors of this study hypothesized that
SD would negatively impact officers’ accuracy of decision making during SDS scenarios as well as their reaction times in such scenarios. This hypothesis appears to be amply
confirmed by the results of the present study
because many of the measures of fatigue correlated strongly with decreases in decision
making in the deadly force simulations and
with increases in reaction time. Specifically,
poor sleep quality, greater TTA, more days
worked, and working night or swing shifts
all decreased the accuracy of officers’ decision making, especially when officers were
presented with no-shoot and ambiguous
scenarios. Greater TTA, more days worked,
and working night or swing shifts also
increased officers’ reaction times during these
deadly force simulations. Finally, the effects of
fatigue also increased throughout each work
day, with officers’ reaction times increasing
consistently from their pre-shift assessment
to their post-shift assessment.
The body of scientific literature regarding
standard sleep requirements, sleep deprivation, and cumulative sleep debt, along with
the effects of these factors on performance, is
large and continues to grow. Time and again,
the primary finding within the literature
was the statistically significant relationship
between sleep deprivation and performance
in that sleep deficiency leads to performance
deficiency. The law enforcement field is
aware of the deficits from sleep deprivation,
but never before to the knowledge of this
researcher has a sleep-related study been so
directed to law enforcement’s most crucial
element—the application of deadly force.
The starting point of the discussion revolves
around the amount of sleep deprivation
experienced by the participants in this study.
Participants were not requested to change
sleep patterns or restrict sleep as is often the
case in sleep-related studies. The current
study simply looked at the police officers’ real
life data and compared their reported experiences and assessment results to findings from
the existing literature. Therein lay empirical evidence for a general determination of
fatigue’s impact on, and even expectation of,
poorer performance during crisis situations,
which could potentially involve deadly force.
The total hours slept data included days offduty as well as days on-duty to determine what,
if any, change occurred. The results showed
that participants had a negative mean change
of 20 minutes between off-duty and on-duty
sleep. In this light, it is important to note that
the literature review provided scientific evidence that even minor sleep loss can cause
deficiencies in performance (Belenky et al.,
57
2003; Bonnet & Arand, 1995). That appears to
have been the case with the participants in the
present study.
Those same scientific articles point to both
sleep deprivation and cumulative sleep debt
as indicators of fatigue and performance deficiencies. The current participants not only
received less sleep than recommended, but
they also experienced a rather severe cumulative sleep debt over seven days: 1.5 hours
daily x 7 days = 10.5 hours cumulative sleep
debt. Here again, the literature review noted
the negative effects of cumulative sleep
debt on performance (Barger et al., 2009;
Couyoumdjian et al., 2009; Hoddes et al.,
1973; Van Dongen et al., 2003), which was evident in the present study in that fatigue measures correlated with poor performance and
increased reaction times more during the later
days of the officers’ work weeks.
The TTA of the participants must also be
looked at to determine whether or not fatigue
is likely to be present. Study participants
reported TTA on the cusp of the hours scientifically shown to provide performance deficiencies, equivalent to a 0.05% blood alcohol
content (NHTSA, 2006; Senjo & Heward,
2007), with TTA, M = 16 hours. Based on these
results, it is empirically evident that the current subjects did suffer some level of fatigue.
The present study used two additional scientific measures to assess fatigue and provide
additional validation of the evidence of fatigue
discussed thus far. The first method is the
well-validated ESS on which our participants
self-reported post-shift mean results indicating excessive daytime sleepiness: M = 11
(Rajaratnam et al., 2011). This reported level
of daytime sleepiness concurs with the participants’ reported TTA, hours slept, and sleep
debt. It should be noted that the most significant change in ESS scores occurred between
Day 1 pre-shift and Day 3 post-shift: Day 1
pre-shift ESS, M = 6.5, but Day 3 post-shift
ESS, M = 12.75.
58
The well-validated PVT was also utilized,
providing RT results for all days of the study
at the beginning and end of each duty day.
The results indicate that RT increased in the
direction of prediction over the duration of
the study: pre-shift PVT, M = 414 ms, but
post-shift PVT, M = 461 ms, an 11% increase
in RT. This is further empirical evidence of
increased fatigue and a coinciding performance decrease measured by RT. Coinciding
with the ESS and providing validation to the
present method, this PVT increase from Day 1
pre-shift to Day 3 post-shift was the most significant change in the direction of prediction.
Deficiencies remained on Day 4 for both validated tests but leveled off as expected, with
no increased sleep restriction, similar to what
others have documented (Banks & Dinges,
2007).
Deadly Force and Reaction Times
Based on the results from the PVT, it is clear
that RT was affected by increasing levels of
fatigue. However, a corollary question is
whether or not those RT deficiencies translated to the SDS task. Many significant correlations emerged in the predicted direction regarding TTA and SDS RT scores. The
shift and the number of days worked also
negatively affected SDS RT scores. The correlations between SDS RT and TTA, days
worked, and shift worked strongly suggest
that fatigue directly increases an officer’s
reaction time to deadly force decisions, at
least in the simulated environment of the
present study.
Deadly Force and Decision Making
The data clearly show that subjective SQ
and TTA had a great impact on the officers’
ability to decide correctly between the three
SDS possibilities. It should be noted that the
ambiguous responses were coded so that
only a total lack of action impacted the participants’ results negatively. Although not
directly related to the hypothesis of this
study, it is important to point out that decision making on the SDS in fatigue-related
performance changes were associated overwhelmingly with the no shoot and ambiguous targets. Likely, the reason for this is due
to the no shoot and ambiguous situations
requiring more cognitive processing power
than clear shoot situations, using a rule-based
decision-making model (Harrison & Horne,
2000; Maddox et al., 2009).
This is the first time a use-of-force decisionmaking sleep study has been conducted in
this manner. Specifically, participants were
not asked to sleep less or stay awake longer
as is often the practice in sleep-related studies. Rather, participants simply worked the
shifts and hours required by their respective
agencies. The data supported the hypothesis
by showing that fatigue does appear to affect
both deadly force reaction times and decision
making.
One different outcome of the present study
is that it did not suggest the extreme effects
of fatigue that have been reported in much
larger studies. For example, Rajaratnam et al.
(2011) conducted a large study of police
officers (N = 4,957) in which 40% screened
positive for a sleep disorder and about 18%
later reported making serious administrative
errors. Senjo and Heward (2007) found officers were working significantly longer hours
(66 to 75 hours weekly) and receiving much
less sleep (3 to 6 hours per night) than was
found in the present study. Vila et al. (2000)
conducted a very large study involving several law enforcement agencies and found that
59% of officers did not sleep an average of 7
or more hours per night, while 16% self-reported trouble staying awake while driving. In light of the effects of fatigue on the
deadly force decisions discussed in the present study, if these more extreme sleep deficits
are occurring in some police agencies, these
greater amounts would raise a serious concern about police decision making in deadly
force situations.
Fatigue Survey
The survey provided troubling results demonstrating that a very large portion of police officers believe that the law enforcement industry
needs to study the impact of sleep deprivation on officer performance. Respondents also
reported that the law enforcement industry
is not sufficiently concerned with the impact
of fatigue on police performance and errors.
These results support the literature review
about the negative effects of shift work,
changes in sleep patterns, and the relationship between fatigue, errors, and general job
performance, suggesting that most officers are
aware of these issues, contend with them regularly, and would like to see solutions to prevent deteriorated job performance and errors.
The survey also supports the TCP results in
areas such as changes in sleep patterns on and
off duty, shift work and fatigue, SDS errors
and slowed RT related to fatigue, and slowed
PVT RT based on fatigue, suggesting that
police in general have an awareness of these
factors, are doing their best to compensate for
them, and are requesting assistance in remedying the causes of fatigue.
Validity
Validity concerns in this study were always
within the researcher’s purview. The SDS,
although a new measurement device, was
based on a similar computer-based shoot/
don’t shoot program used in other studies
(Correll et al., 2007). The ESS/PVT results
complemented those of the SDS, providing
solid evidence of concurrent validity for the
SDS. In addition, both subjective reporting
of fatigue and reaction times increased from
pre- to post-shift testing, suggesting that the
SDS has predictive validity as well.
Limitations
This study entails several limitations. The
first limitation is the method of assessment
delivery. The TCP was administered online
and outside of tightly controlled laboratory
59
conditions, allowing for the possibility of
minor differences in how tasks were completed. To compensate, the TCP included photographs and clear instructions, but these do
not compete with the controls available within
a laboratory environment. Additionally, due
to technical issues with the TCP, we lost an
overwhelming amount of data. A total of 215
participants logged into the program, but
only a subset (n = 53) was actually able to
complete the assessment. Most of the technical issues revolved around compatibility and
could have been remedied within a laboratory environment. These issues have left this
investigation with a relatively small sample
size. To protect the anonymity of police officers who volunteered to be in this study, minimal sociodemographic data were collected.
Yet, there may be relationships between
some sociodemographic variables and performance. For instance, most participants within
this study were men. In addition, because a
portion of the subjects were unknown to the
researcher, although unlikely, it is possible
that not all subjects were in fact police officers; some subjects’ professional credentials
were not possible to verify.
Follow Up
This study applied the findings from previous investigations to the never before tested
area of officer fatigue and decision-making/
reaction time during deadly force encounters. The study found, not surprisingly, that
even minor amounts of sleep deprivation,
decreased sleep quality, and shift work all
have a negative effect upon officers’ speed
and ability to make appropriate decisions
in deadly force situations. What may stand
out to law enforcement administrators and
policymakers are the relatively low levels of
sleep deprivation among the subjects in this
study, which nevertheless were sufficient to
cause performance deficits. Several national
studies with larger sample sizes have suggested that officers are typically much more
sleep deprived than the present subjects. As
such, the probable impact of fatigue on the
60
outcomes of deadly force encounters may
be a serious concern in the law enforcement
community.
Clearly, much larger samples are needed
to provide a more detailed investigation of
officers’ sleep habits on and off duty over a
longer period of time and the effects of fatigue
on performance. Tracking for the nature of
the sleep disturbances (e.g., court appearances, overtime, and other work assignments)
should be included. A single day’s sleep disturbance could greatly increase total hours
awake and cumulative sleep debt, which
both evidenced powerful effects on decision
making and reaction times within the present SDS task. A between-subjects comparison comparing sleep deprived and non-sleep
deprived officers’ performance could also be
very productive.
Former Police Chief and current sleep
researcher Bryan Vila has been studying
police officer fatigue for decades. He has
established the necessity for changes in the
police community regarding sleep. Vila and
Kenney (2002) provided a list of what some of
those changes should be: (1) Police executives
should be concerned with the total number
of employee work hours; (2) Police executive
should provide employees a voice in their
shift and work hours; (3) Police executives
should assess levels of employee fatigue; and
(4) Police executives should provide employees with sleep- and fatigue-related training to
ensure good habits.
The results of the present study suggest that
law enforcement executives, risk managers,
and their legal representatives may need to
come to terms with the necessity for change
within law enforcement to reduce the adverse
effects of fatigue, particularly on the outcome of deadly force encounters. Ignoring the
risks of excessive overtime, randomized shift
schedules, and unforgiving court appearance
schedules would appear to be unwise in light
of the data. Empirical evidence published
prior to the present study has already shown
the negative effects of sleep deprivation on
performance and resulted in legislation and
policy changes for some industries involved
in ensuring public safety such as truck drivers, commercial airline pilots, medical residents, and air traffic controllers (Arora, 2010;
Halsey, 2012; Lockridge, 2014; Trejos, 2014). It
may be time for law enforcement to address
this long-standing issue.
The current study demonstrates agreement
with previous sleep deprivation studies in
regard to performance, and it builds on these
previous investigations by suggesting that
sleep deprivation adversely impacts law
enforcement officers’ most difficult decision
at the moment officers are faced with deadly
force encounters. Just as no law enforcement
executive would place an intoxicated officer
on the street, they may come to understand
the dangers of placing a fatigued officer in the
line of duty. Tired cops make inferior decisions and react more slowly, placing themselves and the public they serve at unnecessary risk.
References
Adams, K., Alpert, G. P., Dunham, R. G., Garner,
J. H., Greenfeld, L. A., Henriquez, M. A.,
. . . Smith, S. K. (1999). Use of force by police:
Overview of national and local data. Retrieved
from https://www.ncjrs.gov/pdffiles1/nij/
176330-1.pdf.
Adam, R., Bays, P. M., & Husain, M. (2012).
Rapid decision-making under risk. Cognitive Neuroscience, 3(1), 52-61.
Alhola, P., & Polo-Kantola, P. (2007). Sleep
deprivation: Impact on cognitive performance. Neuropsychiatric Disease and Treatment, 3(5), 553-567.
Antal, L. C. (1975). The effects of the changes of
the circadian body rhythm on the sport
shooter. Retrieved from www.ncbi.nlm.nih.
gov/pmc/articles/PMC1859287.
Arora, V. (2010). Patient handoffs limit residents
work hour cap gains. Retrieved from www.
kevinmd.com/blog/2010/03/patienthandoffs-limit-residents-work-hour-capgains.html.
Banks, S., & Dinges, D. F. (2007). Behavioral
and physiological consequences of sleep
restriction. Journal of Clinical Sleep Medicine,
3(5), 519-552.
Barger, L. K., Lockley, S. W., Rajaratnam, S. M.,
& Landrigan, C. P. (2009). Neurobehavioral,
health, and safety consequences associated
with shift work in safety-sensitive professions. Current Neurology and Neuroscience
Reports, 9(2), 155-164.
Baumann, J., & DeSteno, D. (2010). Emotion
guided threat detection: Expecting guns
where there are none. Journal of Personality
and Social Psychology, 99(4), 595-610.
Belenky, G., Wesensten, N. J., Thorne, D. R.,
Thomas, M. L., Sing, H. C., Redmond, D. P.,
. . . Balkin, T. J. (2003). Patterns of performance degradation and restoration during
sleep restriction and subsequent recovery:
A sleep dose-response study. Journal of Sleep
Research, 12(1), 1-12.
Blair, J. P., Pollock, J., Montague, D., Nichols,
T., Curnutt, J., & Burns, D. (2011). Reasonableness and reaction time. Police Quarterly,
14(4), 323-343.
Bonnet, M. H., & Arand, D. L. (1995). We are
chronically sleep deprived. Retrieved from
www.soundersleep.com/uploads/BonnetArand.pdf.
Correll, J., Wittenbrink, B., Park, B., Judd, C. M.,
Sadler, M., & Keesee, T. (2007). Across the
thin blue line: Police officers and racial bias
in the decision to shoot. Journal of Personality
and Social Psychology, 92(6), 1006-1023.
Couyoumdjian, A., Sdoia, S., Tempesta, D.,
Curcio, G., Rastenllini, E., De Gennaro, L.,
61
& Ferrara, M. (2009). The effects of sleep and
sleep deprivation on task-switching performance. Journal of Sleep Research, 19(1), 64-70.
Dinges, D. F., & Basner, M. (2011). Maximizing sensitivity of the Psychomotor Vigilance Test (PVT) to sleep loss. SLEEP, 34(5),
581-591.
Dinges, D. F., & Basner, M. (2012). An adaptive-duration version of the PVT accurately
tracks changes in psychomotor vigilance
induced by sleep restriction. SLEEP, 35(2),
193-202.
Dinges, D. F., Pack, F., Williams, K., Gillen, K.
A., Powell, J. W., Ott, G. E., . . . Pack, A. I.
(1997). Cumulative sleepiness, mood disturbance, and psychomotor vigilance performance decrements during a week of sleep
restricted to 4-5 hours per night. American
Sleep Disorders Association and Sleep Research
Society, 20(4), 267-277.
Dinges, D. F., Rogers, N. L., & Baynard, M. D.
(n.d.). Chronic sleep deprivation. Retrieved from
www.med.upenn.edu/uep/user_documents/
dfd22.pdf.
Dorrian, J., Rogers, N. L., & Dinges, D. F. (n.d.).
Psychomotor vigilance performance: Neurocognitive assay sensitive to sleep loss. Retrieved from
www.med.upenn.edu/uep/user_documents/
Dorrianetal.PVTchapterinKushida2005.pdf.
Durmer, J. S., & Dinges, D. F. (2005). Neurocognitive consequences of sleep deprivation. Seminars in Neurology, 25(1), 117-129.
Edinger, J. D., & Fins, A. I. (1995). The distribution and clinical significance of sleep time
misperceptions among insomniacs. SLEEP,
18(4), 232-239.
Edwards, B. J., & Waterhouse, J. (2009). Effects
of one night of partial sleep deprivation
upon diurnal rhythms of accuracy and consistency in throwing darts. Chronobiology
62
International, 26(4), 756-768. http://dx.doi.
org/10.1080/07420520902929037.
Force Science Institute. (2005, January). Force
Science News #10: Was suspect’s shooting a
police execution? Detailed findings from Force
Science Research Center help federal jury
decide. Retrieved from www.forcescience.org/
fsnews/10.html.
Gartenberg, D., & Parasuraman, R. (2010).
Understanding brain arousal and sleep quality
using a neuroergonomic smart phone application.
Retrieved from www.crcnetbase.com/doi/
abs/10.1201/EBK1439835012-c20.
Graham v. Connor, 490 U.S. 386 (1989).
Retrieved from http://supreme.justia.com/
cases/federal/us/490/386/case.html#399.
Halsey, A. (2012, June 15). Air traffic controllers
violating no-doze schedule. The Washington
Post. Retrieved from www.washingtonpost.
com/local/trafficandcommuting/air-trafficcontrollers-arent-keeping-to-no-doze-schedule/
2012/06/15/gJQAJD5FfV_story.html.
Harrison, Y., & Horne, J. A. (2000). The impact
of sleep deprivation on decision making: A
review. Journal of Experimental Psychology:
Applied, 6(3), 236-249. Retrieved from http://
postcog.ucd.ie/files/Hrrison%20and%20
horne.pdf.
Hoddes, E., Zarcone, V., Smythe, H., Phillips,
R., & Dement, R. (1973). Quantifications of
sleepiness: A new approach. Psychophysiology, 10(4), 431-436.
Honig, A., & Lewinski, W. J. (2008). A survey
of the research on human factors related
to lethal force encounters: Implications for
law enforcement training, tactics, and testimony. Law Enforcement Executive Forum, 8(4),
129-152. Retrieved from www.forcescience.
org/articles/Honig129-152.pdf.
Johns, M. W. (2000). Sensitivity and specificity
of the Multiple Sleep Latency Test (MSLT),
the maintenance of wakefulness test and
the Epworth Sleepiness Scale: Failure of the
MSLT as a gold standard. Journal of Sleep
Research, 9(1), 5-11.
Johnson, J. L. (2007). Use of force and the
Hollywood factor. Retrieved from www.aele.
org/law/2007-04MLJ501.pdf.
Lewinksi, W. J. (2000). Why is the suspect shot
in the back? Finally, hard data on how fast the
suspect can be in 11 different shooting scenarios. Retrieved from www.forcescience.org/
articles/shotback.pdf.
Lewinski, W. J., & Honig, A. (2008). A survey
of the research on human factors related to lethal
force encounters: Implications for law enforcement
training, tactics, and testimony. Retrieved from
www.forcescience.org/encounter.html.
Lewinski, W. J., & Hudson, B. (2003). Time to
start shooting? Time to stop shooting? The
Tempe Study. The Police Marksman, 28(5),
26-29. Retrieved from www.forcescience.org/
articles/tempestudy.pdf.
Lewinski, W. J., & Redmann, C. (2009). New
developments in understanding the behavioral science factors in the “stop shooting”
response. Law Enforcement Executive Forum,
9(4), 35-54.
Lim, J., & Dinges, D. F. (2010). A meta-analysis of the impact of short-term sleep deprivation on cognitive variables. Psychological
Bulletin, 136(3), 375-389.
Lockridge, D. (2014). Ferro discusses split sleep
study, CSA, hours of service, electronic logs
at MATS fleet forum. Retrieved from www.
truckinginfo.com/channel/drivers/news/
story/2014/04/ferro-discusses-split-sleepstudy-csa-hours-of-service-electronic-logsat-mats-fleet-forum.aspx.
Maddox, W. T., Glass, B. D., Wolosin, S. M.,
Bowen, C., Mathews, M. D., & Schnyer, D. M.
(2009). The effects of sleep deprivation on
information-integration categorization performance. SLEEP, 32, 1439-1448. Retrieved from
www.cti-home.com/wp-content/uploads/
2014/01/Effect-of-Sleep-Deprivation-OnPerformance.pdf.
Miller, T. (n.d.). Graham v. Connor. Retrieved
from https://www.fletc.gov/sites/default/
files/imported_files/training/programs/
legal-division/podcasts/fletc-legaldivision-use-of-force/podcast-transcripts/
PartIGrahamvConnor.pdf.
National Highway Transportation Safety Administration (NHTSA). (1996). Drowsy driving
and automobile crashes. Retrieved from www.
nhtsa.gov/people/injury/drowsy_driving1/
drowsy.html.
NHTSA. (2006). DWI detection & standardized field sobriety testing: Instructor manual.
Retrieved from www.aschemansmith.com/
Websites/aschemansmith/Images/NHTSA/
SupportDocs_DRE_Forms_Manuals_dwi_
Instructor%20Manual%20-%20February%
202006.pdf.
Neylan, T. C., Metzler, T. J., Best, S. R., Weiss,
D. S., Fagan, J. A., Liberman, A., . . . Marmar,
C. R. (2002). Critical incident exposure and
sleep quality in police officers. Psychosomatic
Medicine, 64(2), 345-352. Retrieved from www.
ncbi.nlm.nih.gov/pubmed/11914452.
O’Brien, M. J., O’Toole, R. V., Zadnik-Newell,
M., Lydecker, A. D., Nascone, J., Sciadini,
M., . . . Eglseder, W. A. (2012). Does sleep
deprivation impair orthopeadic surgeons’
cognitive and psychomotor performance?
The Journal of Bone and Joint Surgery, 94,
1975-1981.
Rajaratnam, S. M., Barger, L. K., Lockley,
S. W., Shea, S. A., Wang, W., Landrigan, C.
P., . . . Czeisler, C. A. (2011). Sleep disorders,
health, and safety in police officers. Journal
of the American Medical Association (JAMA),
306(23), 2567-2578.
63
Samkoff, J. S., & Jacques, C. H. M. (1991).
A review of studies concerning effects
of sleep deprivation and fatigue on residents’ performance. Academic Medicine, 66,
687-693. Retrieved from http://journals.
lww.com/academicmedicine/Abstract/1991/
11000/A_review_of_studies_concerning_
effects_of_sleep.13.aspx.
Senjo, S. R. (2011). Dangerous fatigue conditions: A study of police work and law
enforcement administration. Police Practice
and Research, 12(3), 235-252.
Senjo, S. R., & Heward, M. E. (2007). Sleep and job
performance in law enforcement: Measuring differences between highway patrol, sheriff, and municipal police officers. Retrieved from https://
kucampus.kaplan.edu/documentstore/
docs09/pdf/picj/vol2/issue2/Sleep_and_
Job_Performance_in_Law_Enforcement.pdf.
Sharp, M. J., & Hess, A. B. (2008). To shoot or
not to shoot: Response and interpretation of
response to armed assailants. Retrieved from
www.forcescience.org/articles/study%20
shoot-or-not.pdf.
Sundelin, T., Lekander, M., Kecklund, G.,
Van Someren, E. J. W., Olsson, A., & Axelsson,
J. (2013). Cues of fatigue: Effects of sleep deprivation on facial appearance. SLEEP, 36, 13551360. Retrieved from www.journalsleep.org/
ViewAbstract.aspx?pid=29095.
Tikuisis, P., Keefe, A. A., McLellan, T. M.,
& Kamimori, G. (2004). Caffeine restores
engagement speed but not shooting precision following 22 h of active wakefulness.
Aviation, Space, and Environmental Medicine,
75(9), 771-776.
Trejos, N. (2014, January 3). New pilot fatigue
rules go into effect this weekend. USA
Today. Retrieved from www.usatoday.com/
story/todayinthesky/2014/01/03/pilotfatigue-mandatory-rest-new-faa-rules/
4304417.
64
U.S. Department of Health and Human Services (DHHS) (2005). Your guide to healthy
sleep. Retrieved from www.nhlbi.nih.gov/
files/docs/public/sleep/healthy_sleep.pdf.
Van Dongen, H. P. A., Maislin, G., Mullington,
J. M., & Dinges, D. F. (2003). The cumulative
cost of additional wakefulness: Dose-response
effects on neurobehavioral functions and
sleep physiology from sleep restriction and
total sleep deprivation. SLEEP, 26(2), 117-126.
Retrieved from www.med.upenn.edu/uep/
user_documents/dfd16.pdf.
Vila, B., & Kenney, D. J. (2002). Tired cops:
The prevalence and potential consequences
of police fatigue. NIJ Journal, 248, 16-21.
Retrieved from https://www.ncjrs.gov/
pdffiles1/jr000248d.pdf.
Vila, B., Kenney, D. J., Morrison, G. B.,
& Reuland, M. (2000). Evaluating the effects
of fatigue on police patrol officers: Final report.
(Unpublished report). Retrieved from www.
ncjrs.gov/pdffiles1/nij/grants/184188.pdf.
Williamson, A. M., & Feyer, A. M. (2000). Moderate sleep deprivation produces impairments in cognitive and motor performance
equivalent to legally prescribed levels of alcohol intoxication. Occupational & Environmental
Medicine, 57, 649-655. Retrieved from http://
oem.bmj.com/content/57/10/649.full.
David M. Blake is a retired law enforcement
officer and avid student of law enforcement use of force. He is a Force Science
Certified Analyst with instructor certifications in Defensive Tactics, Firearms, Force
Options Simulator, and Reality Based
Training. He currently teaches Human
Factors and Force Encounters Analysis for
the California Training Institute. He is also
an adjunct professor of Criminal Justice, a
police academy instructor, and an inservice
use-of-force instructor. He owns the Blake
Consulting and Training Group.
Dr. Edward Cumella is a professor of
Graduate Psychology at Kaplan University. He received his Bachelor of Arts at
Harvard, and his Master of Arts/PhD in
Psychology at University of North Carolina Chapel Hill. He has worked in
mental health for 29 years. Previously,
he was Executive Director at America’s
largest eating disorder facility and in private practice. Dr. Cumella has published
50 peer-reviewed articles and has been
interviewed on TV, radio, and newspapers (e.g., ABC, FOX, New York Times).
He is the editor of a book on eating
disorders.
Contact Information
David M. Blake
[email protected]
Dr. Edward Cumella
[email protected]
65
Criminal Justice Practitioner Attitudes
Toward Risk Assessments in
Response to Domestic Violence
Lee E. Ross, PhD, Associate Professor, Department of Criminal Justice,
University of Central Florida
Abstract
The purpose of this study was to explore the attitudes of criminal justice practitioners toward risk assessments in domestic violence-related cases and to appreciate the context in which they are utilized. A total of
198 practitioners responded to an Internet-based survey consisting of 56 items. The data were examined
and presented in the form of univariate, bivariate, correlation, and regression analyses. A linear regression
model revealed three statistically significant factors that partially explain an agency’s use of risk assessments: (1) whether the agency has a domestic violence unit, (2) an awareness of other agencies that use risk
assessments, and (3) the ability to tell when someone is at risk for intimate partner homicide. In addition,
these results suggest that risk assessments are underutilized among law enforcement agencies. The implications of these findings are discussed in terms that promote an increased understanding and use of risk
assessments among criminal justice practitioners in hopes of preventing intimate partner homicides.
Introduction
Historically, risk assessments have been used
throughout the criminal justice system—from
managing the risk of police misconduct to
predicting recidivism among certain types
of offenders. Given the difficulty of predicting intimate partner homicides, recently, risk
assessment models have captured the interests of law enforcement when responding to
incidents of domestic violence. Thus far, however, very few researchers have been able to
document the extent to which risk assessments
are actually utilized by law enforcement when
responding to domestic violence (for a notable exception, see Campbell, Sharps, & Glass,
2001; Ross & Kane, 2014). Moreover, there is
very little research that explores the perceptions of law enforcement and other criminal
justice practitioners on the employment of
risk assessments when responding to domestic violence incidents. It would be extremely
useful to explore how these instruments and
66
practices are viewed—whether negatively,
positively, or indifferently—by law enforcement and other criminal justice practitioners.
For many actual and potential victims of
domestic violence, calling the police is one
of the most commonly employed strategies
used in abusive relationships. Beyond their
initial response—including arresting offenders, administering first aid, and making shelter referrals—police assume a vital role in
what happens to victims in the immediate
aftermath of a battering incident. Often times,
a successful intervention depends on the willingness and ability of police and victims to
appreciate and assess the degree of danger
a victim faces, especially the potential for
repeat victimizations that could prove lethal.
To facilitate these efforts, more and more
law enforcement agencies throughout the
United States, Canada, and even Australia
have begun to utilize an array of risk assessment instruments to gauge the victim’s risk
for lethality at the hands of their abuser
(Klein, 2012). Within some jurisdictions and
provinces, such as British Columbia, the use
of particular risk assessments1 is mandated
(Kropp, 2004). Among other jurisdictions,
partnerships have been developed between
researchers and law enforcement to standardize the use of risk assessment.
Risk Assessments
A thorough review of existing risk assessment instruments (variously referred to as
danger and lethality assessments) reveals a
surprising number, with most incorporating
similar measures of offender characteristics
and social dimensions of intimate relationships. Danger (i.e., risk) assessment, per se, is
a clinical and research tool that was designed
to assist battered women in assessing their
danger of being killed by their intimate partner (Campbell, 1986; Nicolaidis et al., 2003).
Conventionally, there are three methods (or
approaches) used when assessing an individual’s risk for violent victimization (i.e., danger):
(1) unstructured clinical assessments, (2) actuarial assessments, and (3) structured professional
assessments.
As the most commonly used method of
assessing spousal violence, unstructured clinical assessments are based upon the professional’s training, experience, and observations
of a specific client (see Campbell et al., 2001;
Dutton & Kropp, 2000). Moreover, decisions
are based on professional discretion and are
usually justified according to the qualifications of experienced practitioners. The actuarial method, in contrast, involves predicting
someone’s behavior based upon how others
have acted in similar situations (actuarial).
One attractive feature of the actuarial method
is that it was designed to predict specific
behaviors within a specific time-frame. It is
important to note that the main distinction
between an actuarial and the unstructured
clinical method is improved reliability and
validity. The third method, structured professional judgments, was developed as a middle
ground in response to limitations associated
with actuarial methods and unstructured
clinical assessments. With increasing appeal
to criminal justice professionals, this method
does not impose any restrictions for the inclusion, weighting, or combining of risk factors.
Rather, the primary goal of this approach
to risk assessment is to prevent violence by
any established means necessary (Douglas &
Kropp, 2002).
Among risk assessments in general—and
one of the earliest models used in response
to domestic violence—is the Lethality Assessment Program (LAP) that was developed by
the Maryland Network Against Domestic Violence (MNADV) in 2005. This model provided
an easy and effective method for law enforcement and other community professionals to
identify victims of domestic violence who
were at risk for being seriously injured or
killed by their intimate partners and immediately connected them to the domestic violence
service provider in their area (Klein, 2012). The
LAP consists of 11 items and is the counterpart
to the Danger Assessment (Campbell et al.,
2001) which consists of 20 items. Designed as
a field intervention, the LAP can assist practitioners who encounter a victim of intimate
partner violence during the course of their
work (Messing et al., 2014). When used by
police, the LAP typically involves a two-step
process. First, a police officer responding to
the scene of a domestic violence incident uses
it to identify victims in “high danger” or those
considered at high risk for homicide. Second,
all victims receiving a score of four (or higher)
are put in immediate telephone contact with a
collaborating social service provider who can
assist them with safety planning options and
encourage them to come in for services.
An Evaluation of Danger Assessments
Based on a study by Klein (2012), an important byproduct of the danger assessments—
beyond increasing victim safety—has been
improved partnerships and collaborations
among law enforcement personnel, domestic
67
violence programs, healthcare providers, the
faith community, and allied professionals.
For instance, although studies conducted in
Oklahoma concerning the effectiveness of
lethality assessments produced mixed results,
they were deemed quite successful in encouraging high-risk victims to (at least) talk with
a domestic violence advocate. Overall, these
sessions tended to increase a survivor’s use
of formal and informal protective strategies
while decreasing the frequency and/or severity of physical violence (Messing et al., 2011).2
Equally positive results were also reported in
Minnesota’s Anoka County, where between
2010 and 2012, the county’s judges, probation officers, and prosecuting attorneys began
using lethality scales to make determinations on bail and other pre-trial related issues
(Dahl, 2012). The implementation of lethality
assessments in Kansas City, Missouri, also
produced successful outcomes that rival those
of Minnesota and Oklahoma where over 900
officers were trained in their use. Here, one
of the more interesting findings is that while
rates of domestic violence had spiked in
cities around the country in the midst of an
economic recession, Kansas City’s domestic
violence homicides actually dropped by 25%,
and domestic violence aggravated assaults
fell by 7% in a one-year period (Forte, 2011).
Additional evidence regarding the effectiveness of lethality assessments as an effective
intervention is that shelter occupancy rates
increased as did calls to its domestic violence
hotline. Overall, the program proved so successful that, according to the police chief, the
number of victims requesting services far
exceeded the shelter’s capacity to accommodate all of them (Klein, 2012).
Problems and Concerns About Risk
Assessments
Clearly, the use of lethality assessments by
criminal justice practitioners has produced
some encouraging results, and additional evidence suggests that these can even be used
reliably and validly in forensic mental health
contexts (see Otto & Douglas, 2010). Still,
68
many jurisdictions throughout the U.S. do not
mandate nor encourage the use of formal risk
assessments when responding to domestic
violence-related incidents. Even among those
agencies that do, there are concerns that risk
assessment and violence reduction programs
are somewhat over-rated as some argue that
we should not be too surprised that lethality assessments are effective. Moreover, the
claim is that any program that limits firearms
in the hands of abusers should reduce domestic violence homicides as firearms are the
overwhelming weapon of choice for batterers
(Klein, 2012; Paulozzi, Saltzman, Thompson,
& Holmgreen, 2001). Whether a criminal justice practitioner or a seasoned researcher,
both tend to promote the argument that any
law enforcement program that expands the
number of domestic violence arrests will ultimately reduce the number of armed abusers
(Klein, 2012). Rounding out the chorus of
criticism is that the current research on risk
assessment focuses too exclusively on the prediction of future violence rather than on the
management of risk. This exclusive focus has
proven problematic for some researchers who
argue that the aim of practitioners who intervene in cases of intimate partner violence is to
prevent future abuse—not to necessarily predict it (Bennett Cattaneo & Goodman, 2007;
Kropp, 2004).
The above claims and concerns appear reasonable as the literature is relatively silent
on law enforcement’s appreciation (or lack
thereof) and perceptions of risk assessments.
For purposes of the present study, perhaps
the most immediate concern is that there is
relatively little research on the extent to which
danger assessments are utilized among law
enforcement and other criminal justice practitioners. Moreover, a thorough review of the
literature suggests that some states and many
police departments do not utilize any kind
of a structured risk assessment at all, but,
instead, encourage officers to rely on their
“gut feeling” concerning the danger a certain victim might face. In other words, many
of the officers engage in informal anecdotal
assessments of danger that tend to lack
structure and are commonly referred to as
“unstructured professional judgments” (analogous to unstructured clinical assessments).
To some degree, this might reflect a lack of
officer training in the use of risk assessment
instruments as some structured professional
judgments require both a level of familiarity with mental health concepts balanced
with considerable discretion (Storey, Gibas,
Reeves, & Hart, 2011). Nonetheless, there are
countless benefits when practitioners become
proficient in using risk assessments as these
might assist both victims and police in realizing the danger and gravity of certain situations. Furthermore, these would enable
practitioners to validate potential factors that
could reliably predict and hopefully prevent
lethal outcomes (see Ross & Kane, 2014).
Based on this literature review, we are left
with a set of lingering questions and concerns
regarding the acceptance of risk assessments
in the war against domestic violence. How
do law enforcement agencies feel about risk
assessments? Are risk assessments viewed as
potential obstacles and/or stumbling blocks
in law enforcement’s response to domestic
violence? Are agencies adequately equipped
with the requisite training, manpower, and
resources to conduct risk assessments? These
are all important questions that require
thoughtful answers if we are to become more
knowledgeable and better informed on criminal justice practitioner attitudes toward the
use of risk assessment in response to domestic violence. Therefore, in an effort to answer
these questions, this exploratory research
study seeks to (1) measure criminal justice
practitioner attitudes toward risk assessments,
(2) determine the degree to which risk assessments are used among law enforcement, and
(3) identify correlates and factors that explain
the use and non-use of risk assessments. As
part of a much larger study, these and related
questions were explored by surveying the
perceptions and attitudes of criminal justice
practitioners in Central Florida who work on
domestic violence-related cases.
Research Methodology3
Data Sources
The survey used in the present study was
constructed specifically for criminal justice
practitioners—including fatality review team
members who dealt with domestic violence
on a daily basis—to measure their attitudes
about domestic violence issues. Consisting of
56 items, the survey instrument included attitudinal questions regarding lethality assessment and perceived seriousness of domestic
violence calls. As part of a larger study of
intimate partner homicide, additional measures included attitudes regarding pro-arrest
versus mandatory arrest policies, restraining
orders, no-drop prosecution policies, and
proactive policing strategies.
Measures
The measures used in the present study parallel those of previous researchers who have
examined law enforcement attitudes toward
domestic violence (see Gover, Pudrzynska,
Dodge, & Dodge, 2011; Ross & Kane, 2014).
The survey contained a host of demographic
variables, including race, age, and gender. For
purposes of correlational analyses, most of
the multi-categorical variables were recoded
as dichotomous variables. For example, Does
your agency use a risk assessment?, originally
measured along three categories—yes, no,
and not sure—was recoded as yes (1), no (0).
The only other variable relevant to these analyses is gender, which was recoded as male (1),
female (0).
The survey also included a number of attitudinal items which were measured on a fivepoint Likert scale. The following six attitudinal items are most relevant to our analyses:
(1) I can tell when someone is at risk for intimate
partner homicide, (2) It is quite possible to predict
intimate partner homicide, (3) Risk assessment
tools are virtually useless in predicting intimate
partner homicides, (4) When a batterer is arrested,
the police should assess his risk of killing someone,
(5) Our agency does not have the resources
69
(or time) to conduct risk assessments, and (6)
Personally, I am not interested in conducting any
type of risk assessment. Among these cases and
for data analysis purposes, response categories for all of these attitudinal variables were
recoded (from a 5-point Likert Scale) to a trichotomous measure of “agree,” “disagree,”
and “undecided.”
Procedures
The initial survey targeted only those individuals comprising domestic violence fatality
review teams (DVFRT)4 throughout the state
of Florida. As this is a convenience sample,
potential respondents were initially identified
by telephone and e-mail contacts to coordinators of DVFRTs, both locally and statewide.
When initial survey results revealed a smaller
than expected number of law enforcement
officers working with DVFRTs, the survey
was expanded to include local law enforcement agencies. Administrators approved the
distribution of the questionnaire within police
agencies, and officers were made aware of the
survey through roll calls and were encouraged to participate.
Internet Surveys
The survey instrument was administered
over the Internet (via Qualtrics research software) during November and December of
2013. Internet surveys are quite appealing
in terms of convenience, lower cost, and in
assuring respondent anonymity. Of these,
assuring anonymity is perhaps its greatest
asset as there is no way to know who submits responses. There are disadvantages to
Internet surveys as well, including limited
sampling, respondent availability, and inadvertent deletions and/or omissions. Given
the difficulty of drawing probability samples
(based on e-mail addresses), Internet surveys
also increase the potential for sampling bias
(Withrow, 2014). Therefore, to obtain a larger
sample size, potential respondents were provided with advanced notice of an upcoming
survey (from agency administrators). Survey
participation was voluntary and anonymous
70
and, on average, took anywhere from 10 to 15
minutes to complete.
Respondent Demographics
Data collection efforts yielded a total of 198
surveys. Of these, 54 surveys were not fully
completed (but were still useful—except for
certain missing items). On average, 154 completed surveys comprise the basis of analysis, of which there were 52 responses from
females (32.2%) and 102 from males (68.8%).
Caucasians comprised the vast majority of
respondents (76%), while Latinos and African
Americans comprised 17% and 4%, respectively. Those identifying as bi-racial and
Native Americans comprised the remaining
3%. Most respondents were married (71%),
with an equal number being either divorced or
separated (at 14.5% each). Levels of education
were normally distributed, with nearly 42%
possessing either an Associate’s or Bachelor’s
degree. Nearly 18% had taken some college classes without earning a degree, while
13.5% had earned their master’s degree. In
terms of annual income, nearly 47% reported
earning between $40,000 and $70,000. On the
lower end, 15% earned between $10,000 and
$39,999; while on the upper end, 16% earned
at least $70,100.
Data Analysis
The findings produced in this study are
presented through four statistical analyses:
(1) univariate, (2) bivariate, (3) correlational,
and (4) linear regression. Univariate statistics were estimated on four agency profile
descriptors using response categories “yes,”
“no,” and “not sure.” In addition, six attitudinal items were examined separately, using the
recoded responses (“agree” and “disagree”).
To increase the validity and reliability of
these measures, “undecided” responses were
excluded from the analysis. In the second
analysis, six attitudinal items were analyzed
in relation to the respondent characteristics
to determine whether attitudes varied by
demographics. In the third analysis, differences among attitudinal items and recoded
(dummy) variables (i.e., race/ethnicity,
gender, and education level) were conducted
using t-tests. Finally, correlational analyses
were used to highlight the relationship among
all relevant independent variables. This was
followed by a linear regression model which
sought to determine contextual factors that
support the use (and non-use) or risk assessments among criminal justice agencies.
Univariate Analyses
Table 1 is presented in two parts, including (1) a
descriptive profile of agencies dealing with
domestic violence on a daily basis and (2) an
attitudinal profile among criminal justice
practitioners on various risk-related items.
As indicated in the first part of the table, a
sizeable majority of respondents (44%) were
not sure if their agency use[d] risk assessments
to measure someone’s risk for intimate partner
homicide. Of those who were sure, a greater
percentage indicated their agency did not
use risk assessments (33.3% versus 22.6%).
In terms of an awareness of other agencies
using risk assessments, most indicated that
they were not aware (44.4%). On the other
hand, respondents were slightly more likely
to indicate that their agency had a domestic violence unit that only handled domestic-violence
related calls (48.9% versus 43.2%). However, of
all respondents, relatively few (12%) claimed
membership in a domestic violence task force.
The second part of Table 1 provides an attitudinal profile of criminal justice practitioners regarding risk-related items. As illustrated, greater percentages of agreement than
Table 1: Agency and Attitudinal Profile of Criminal Justice Practitioners on Risk-Related
Items (n = 186)*
Agency Profile
Does your agency use risk assessment tools that measures
someone’s risk for intimate partner homicide?
Are you aware of any agencies that use risk assessments (in general) for cases of domestic violence?
Does your law enforcement agency have a domestic violence
unit that only handles domestic violence-related calls?
Are you a member of a domestic violence task force?
Yes
22.6%
(40)
28.1%
(50)
48.9%
(87)
12.3%
(40)
No
33.3%
(59)
44.4%
(79)
43.3%
(77)
87.7%
(21)
Not Sure
44.1%
(78)
27.5%
(49)
1.7%**
(3/11)
-(150)
Attitudinal Profile
I can tell when someone is at risk for intimate partner homicide.
Disagree
17.9%
(30)
21.4%
(36)
57.7%
(97)
13.8%
(23)
47.6%
(80)
73.2%
(120)
Undecided
50.6%
(85)
38.1%
(64)
38.7%
(65)
14.4%
(24)
37.7%
(55)
20.1%
(33)
Agree
31.6%
(53)
40.5%
(68)
3.6%
(6)
74.9%
(120)
19.7%
(33)
6.7%
(11)
It is quite possible to predict intimate partner homicide.
Risk assessment tools are virtually useless in predicting
intimate partner homicides.
When a batterer is arrested, the police should assess his risk
of killing someone.
Our agency does not have the resources (or time) to conduct
risk assessments.
Personally, I am not interested in conducting any type of risk
assessment.
* In cases where numbers do not equal (n = 186), it reflects incomplete surveys.
** In 11 cases, the respondent agency was non-law enforcement.
*** These items were originally measured on a 5-point Likert scale. “Agree” reflects combinations of “agree”
and “strongly agree” while “Disagree” reflects combinations of “disagree” and “strongly disagree.”
71
disagreement were found among the following items: (1) I can tell when someone is at risk
for intimate partner homicide (31% versus 17%),
(2) It is quite possible to predict intimate partner
homicide (40.5% versus 21.4%), and (3) When a
batterer is arrested, the police should assess his risk
for killing someone (74.9% versus 13.8%). On
the other hand, respondents were more likely
to disagree with the following: (1) Risk assessment tools are virtually useless in predicting intimate partner homicides (57.7%), (2) Our agency
does not have the time to conduct risk assessments
(47.6%), and (3) Personally, I am not interested
in conducting any type of risk assessment (73.2%).
Bivariate Analyses
Table 2 presents measures of agreement in
practitioner attitudes toward risk assessments based on gender.5 As evidenced, male
and female practitioner attitudes along these
items are remarkably similar in terms of their
agreement, disagreement, and indecisiveness. There were, however, a few instances
for which differences are noted. In terms of
agreement, females were more likely than
males to agree to I can tell when someone is at
risk for intimate partner homicide (39.2% versus
28.4%). Males, on the other hand, were more
likely to agree that When a batterer is arrested,
the police should assess his risk for killing someone
(69.6% versus 19.6%). In terms of disagreement, females were twice as likely to disagree
that It is quite possible to predict intimate partner homicide (34.6% versus 16.7%). While both
disagreed, females were far more likely than
males to disagree with the item Our agency
does not have the resources (or time) to conduct
risk assessments (59.1% versus 17.1%). In terms
of indecisiveness, males were twice as likely
to be undecided regarding the item Risk
assessment tools are virtually useless in predicting
intimate partner homicide (45.1% versus 23.1%).
Correlational Analyses
The next data analyses involve correlations
and are presented in Table 3. Beginning with
gender, a relatively weak, though significant,
correlation exists between gender and whether
an agency uses risk assessment as women are
more likely than men to respond positively
Table 2: Univariate Measures of Agreements* in Attitudes Toward Risk Assessments
Based on Gender (N = 154)
I can tell when someone is at risk for
intimate partner homicide.
Some homicides simply cannot be
prevented.
When a batterer is arrested, the police
should assess his risk of killing someone.
It is quite possible to predict intimate
partner homicide.
Risk assessment tools are virtually
useless in predicting intimate partner
homicides.
Our agency does not have the resources
(or time) to conduct risk assessments.
Personally, I am not interested in conducting any type of risk assessment.
Males (n = 102)
Females (n = 52)
Disagree Undecided Agree Disagree Undecided Agree
21.6
39.2
39.2
21.6
39.2
39.2
21.6
8.8
69.9
21.2
21.2
57.7
18.6
11.8
69.6
25.5
54.9
19.6
16.7
44.1
39.2
34.6
25.0
40.4
52.0
45.1
2.9
73.1
23.1
4.8
17.1
32.4
20.5
51.9
26.9
21.1
72.5
19.6
7.9
80.8
13.5
5.8
* These items were originally measured on a 5-point Likert scale. “Agree” reflects combinations of “agree”
and “strongly agree” while “Disagree” reflects combinations of “disagree” and “strongly disagree.”
72
(r [99] = 0.24, p < 0.05). Being female is also significantly correlated, albeit weak, with knowledge of other agency use of risk assessment
(r [129] = 0.29, p < 0.01). The last and more moderate size correlation suggests that women are
far more likely to belong to a domestic violence
task force than men (r [171] = 0.40, p < 0.01).
These findings also lend additional support to
the results presented in Table 2.
Turning to whether agencies use risk assessments, we find it is positively correlated with
an awareness of other agencies that use risk
assessments (r [129] = 0.69, p < 0.01). Agency
use of risk assessment is also positively correlated with the presence of domestic violence
units within agencies (r [164] = 0.42, p < 0.01)
and whether someone is part of a domestic
violence task force (r [171] = 0.24, p < 0.01). The
final positive correlation was found between
agency use of risk assessments and whether
police should conduct a risk assessment
on arrestees (r [113] = 0.24, p < 0.05). There
are negative correlations as well, including
agency use of risk assessments and views that
risk assessments were useless (r [103] = -0.25,
p < 0.05) and attitudes suggesting agencies did
not have the time to conduct risk assessments
(r [113] = -0.25, p < 0.05). On the other hand,
awareness of other agency use of risk assessments was positively correlated with being
a member of a domestic violence task force
(r [171] = 0.44, p < 0.01), yet negatively correlated with attitudes suggesting agencies did
(r [113] = -0.25, p < 0.05).
Awareness of whether one’s agency had a
domestic violence unit was negatively correlated with attitudes suggesting agencies did
(r [113] = 1.37, p < 0.05). There is a significant
correlation between discerning when someone
was at risk for intimate partner homicide and
the attitude that it is possible to predict intimate partner homicide (r [104] = 0.50, p < 0.05).
Being a member of a task force was positively
correlated with an attitude that one’s agency
did not have the resources (or time) to conduct
risk assessments (r [113] = 0.28, p < 0.05). Being
a member of a task force, however, is negatively correlated with whether police should
assess an offender’s risk of killing someone
(r [143] = -0.35, p < 0.01) and disinterest in
conducting risk assessments (r [131] = -0.25,
p < 0.05). Further negative correlations are
found in attitudes of whether police should
assess an offender’s risk of killing someone
and attitudes that one’s agency did not have
the resources (or time) to conduct risk assessments (r [113] = -0.31, p < 0.05) and disinterest
in conducting risk assessments (r [131] = -0.38,
p < 0.05). Finally, there is a positive correlation between the attitude that agencies did
and disinterest in conducting risk assessments
(r [131] = 0.29, p < 0.05).
Regression Analyses
The final analysis includes a linear regression
model wherein several risk-related independent variables were regressed on the dependent variable (an agency’s use of risk assessments).6 Since the ultimate goal of using risk
assessments is to prevent a worst case scenario (i.e., homicide), the regression model
is used to contextualize factors that partially
explain why agencies may (and may not) use
risk assessments. Using the “enter method,”
a significant regression equation resulted—
F (2, 168) = 10.82, p < 0.001—with an adjusted
R2 value of 0.35. Overall, this model was able
to explain 35% of the variance in whether an
agency uses risk assessments. As shown, an
agency’s use of risk assessments appears to
result from three statistically significant factors: (1) an awareness of other agencies that
use risk assessments, (2) whether the agency
has a domestic violence unit, and (3) an individual’s ability to tell when someone was
at risk for intimate partner homicide. These
results are presented in Table 4.
Discussion
This is the first known study of its kind to
measure the attitudes of criminal justice
73
74
X4
1.00
-0.043
-0.070
-0.082
-0.084
0.128
-0.374**
-0.083
X3
1.00
0.133
0.446**
0.087
0.067
-0.149
0.078
-0.305**
-0.082
1.00
0.057
0.141
0.007
0.163
-0.145
-0.042
X5
1.00
0.509**
-0.160
0.079
0.070
-0.022
X6
X8
1.00
-0.291** 1.00
0.390** -0.357**
-0.228*
0.287*
-0.188
0.258*
X7
1.00
-0.315**
-0.382**
X9
1.00
0.299*
X10
1.00
X11
M
0.33
0.40
0.38
0.53
0.12
0.63
0.65
0.05
0.83
0.29
0.08
X1 = Gender
X2 = Does your agency use risk assessment tools that measures someone’s risk for intimate partner homicide?
X3 = Are you aware of any agencies that use risk assessments (in general) for cases of domestic violence?
X4 = Does your law enforcement agency have a domestic violence unit that only handles domestic violence-related calls?
X5 = Are you a member of a domestic violence task force?
X6 = I can tell when someone is at risk for intimate partner homicide.
X7 = It is quite possible to predict intimate partner homicide.
X8 = Risk assessment tools are virtually useless in predicting intimate partner homicides.
X9 = When a batterer is arrested, the police should assess his risk of killing someone.
X10 = Our agency does not have the resources (or time) to conduct risk assessments.
X11 = Personally, I am not interested in conducting any type of risk assessment.
X1
X2
X1
1.00
X2
0.244*
1.00
X3
0.298**
0.690**
X4
0.133
0.424**
X5
0.406**
0.247**
X6
0.001
0.234
X7
-0.167
-0.054
X8
-0.008
-0.258*
X9
0.142
0.258*
X10
0.016
-0.252*
X11
-0.053
0.006
*p < 0.05, ** p < 0.01
Table 3. Pearson’s (R) Correlation Matrix of Attitudes Toward Risk Assessments
SD
0.47
0.49
0.48
0.50
0.32
0.48
0.47
0.23
0.36
0.45
0.27
N
154
99
129
164
171
83
104
103
143
113
131
Table 4. Regression Results for Agency Use or Risk Assessments (N = 178)
Constant
Are you aware of any agencies that use risk assessments (in
general) for cases of domestic violence?
Does your law enforcement agency have a domestic violence unit
that handles only domestic violence cases?
Are you a member of a domestic violence task force?
I can tell when someone is at risk for intimate partner homicide.
Risk assessment tools are virtually useless in predicting intimate
partner homicide.
When a batterer is arrested, the police should assess his risk for
killing someone.
Our agency does not have the time or resources to conduct risk
assessments.
Personally, I am not interested in conducting any type of risk
assessment.
Adjusted R2 value of 0.35
Standard errors are reported in parentheses.
*p < 0.05, **p < 0.01, ***p < 0.001
practitioners toward the use of risk assessment
in response to domestic violence. Initial ideas
for this study were formed when previous
research revealed that rarely—if ever—were
risk assessments mentioned in the narrative
of fatality reviews (see Ross & Kane, 2014).
To find out why that was the case, the present
study first sought to establish whether risk
assessments were actually used “in the field”
by criminal justice practitioners. As illustrated,
the findings presented in Tables 1 through
4 present a mixed picture of how criminal
justice practitioners utilize and regard risk
assessments. Overall, less than 25% of all practitioners report having used risk assessments,
and the vast majority were not aware of any
agencies that used them. Still, criminal justice
practitioners reported they were highly interested in risk assessments, did not view them
as useless, and denied not having the time
and resources to conduct them (see Table 1).
When risk assessments are used, they are, as
expected, most likely used among law enforcement agencies (comprised mostly of men). On
Unstandardized Coefficients (β)
-0.013
(0.096)
0.444***
(0.061)
0.176***
(0.049)
0.024
(0.082)
0.159
(0.071)*
-0.214
(0.131)
0.141
(0.075)
0.043
(0.068)
0.113
(0.099)
the other hand, while women in the survey were
more likely to constitute domestic violence task
forces, they were significantly more aware (than
men) of agencies that use risk assessments. This
probably reflects the positions held by many
female respondents—mainly those who are
domestic violence counselors, shelter workers,
and victim advocates who constantly administer risk assessments to new clients. Some of the
more straightforward and logical findings are
revealed from a joint examination of Tables 3
and 4. Moreover, the multiple regression model
produced three factors that partially explain
an agency’s use of risk assessment to measure
someone’s risk for intimate partner homicide.
Conceivably, having a domestic violence unit
within a police agency should increase awareness about other agencies that use risk assessments while improving an individual’s ability
(and intuition) to tell when someone is at risk
for intimate partner homicide.
75
Study Limitations
While the findings presented here are important for criminal justice practitioners to consider, there are several methodological limitations that severely restrict their generalizability to other settings. The first limitation
concerns the sample selection which was not
random but purposeful and is best described
as a convenience sample. Another limitation
involves the manner in which the surveys were
administered. As an Internet-based survey,
respondents were not monitored during the
administration process, and there is no way to
ensure that participants did not discuss their
responses with others who had not yet taken
the survey. Also, the method of survey administration may have contributed to the lower
than expected response rate as the surveys
were not distributed, filled out, and returned
at the same time. Another issue concerns the
missing items associated with a number of
respondents who began the survey but failed
to complete it (i.e., 32). A final drawback is
that recoding some of the Likert scale variables into dichotomous (“agree”/“disagree”)
variables resulted in a loss of even more cases
among those who responded as “neither agree
nor disagree.” This, in effect, further reduced
the sample size available for different types
of statistical analyses. Therefore, these results
require a cautious interpretation and may not
generalize to other jurisdictions.
merits of using lethality assessments, including improved community partnerships, as
they tend to encourage victims to seek assistance (Klein, 2012). These results from the
present study indicate that criminal justice
practitioners are not necessarily opposed to
using risk assessments, but, rather, they have
not been trained in their proper use.
Interestingly, it is encouraging to see that risk
assessments are used among law enforcement
agencies that have specific domestic violence
units that respond primarily to domestic
violence-related calls. Being able to reliably
measure an individual’s risk for intimate
partner homicide is not an easy task by any
means. Nonetheless, department chiefs and
other supervisory personnel must be willing
to use all of the resources at their disposal.
To that end, all criminal justice practitioners
who work with domestic violence victims
are encouraged to appreciate the value of a
risk assessment and its potential to serve as
another weapon in the toolbox to combat
domestic violence.
End Notes
1
For purposes of this study, the term risk assessment is used interchangeably with danger
assessment and lethality assessment.
2
In the same study, however, there was no evidence of decreased presence of intimate partner violence or severe violence, and there was
no effect on the utilization of some measured
protective strategies (Messing et al., 2014).
3
As part of a larger study, the “Methods” section used in the present study parallels the
“Methods” section from an earlier publication from the same survey data. See Ross and
Leslie (2014).
4
These teams review and analyze domestic
violence homicide cases to uncover basic
knowledge about causes and factors that
increase or decrease the risk for death and
injury, and specific ways to prevent further
injury and death. For further reference, see
Wilson and Websdale (2006).
Conclusions
Although these research limitations merit consideration, they do not diminish the importance of these findings. This research is the first
of its kind to measure the attitudes of criminal
justice practitioners toward risk assessments
in domestic violence cases. As more and more
criminal justice practitioners embrace the
use and potential of risk assessments when
responding to domestic violence, the more
likely we are to witness more protective measures that not only predict increased risk for
lethality, but, more importantly, prevent it as
well. Previous research has established the
76
5
Statistical significance was not found among
other demographic variables (i.e., race and
ethnicity) and attitudinal measures, and,
therefore, are not included here.
6
The question concerning agency use of risk
assessments was worded specifically as Does
your agency use risk assessment tools that measure someone’s risk for intimate partner homicide?
References
Bennett Cattaneo, L., & Goodman, L. (2007).
New directions in IPV risk assessment: An
empowerment approach to risk management. In K. Kendall-Tackett & S. Giacomoni
(Eds.), Intimate partner violence (pp. 1-17).
Kingston, NJ: Civic Research Institute.
Campbell, J. C. (1986). Nursing assessment
for risk of homicide with battered women.
Advances in Nursing, 8(4), 36-51.
Campbell, J. C., Sharps, P., & Glass, N. (2001).
Risk assessment for intimate partner homicide. In G. F. Pinard & L. Pagani (Eds.),
Clinical assessment of dangerousness: Empirical contributions (pp. 137-157). New York:
Cambridge University Press.
Dahl, J. (2012, April 20). A new model aims to
catch domestic violence that is likely to
turn fatal. CBS News. Retrieved from www.
cbsnews.com/8301-504083_162-57417868
-504083/a-new-model-aims-to-catchdomestic-violence-that-is-likely-to-turnfatal.
Douglas, K. S., & Kropp, P. R. (2002). A prevention-based paradigm for violence risk assessment: Clinical and research applications.
Criminal Justice and Behavior, 2, 617-658.
Dutton, D. G., & Kropp, P. R. (2000). A review
of domestic violence risk instruments.
Trauma, Violence and Abuse, 1, 171-181.
Forte, D. (2011, August 18). Partnership
to protect domestic violence victims so
successful shelter must expand. Kansas
City, Missouri, Police Department Chief’s Blog.
Retrieved from http://kcpdchief.blogspot.
com/search/label/domestic%20violence.
Gover, A. R., Pudrzynska, D., Dodge, P., &
Dodge, M. (2011, May 6). Law enforcement
officers’ attitudes about domestic violence.
Violence Against Women, 17(5), 616-636.
Klein, A. R. (2012). Lethality assessment and
the law enforcement response to domestic
violence. Journal of Police Science Negotiations,
12(2), 87-102. http://dx.doi.org/10.1080/
15332586.2012.720175
Kropp, P. (2004). Some questions regarding
spousal assault risk assessment. Violence
Against Women, 10, 676-697.
Messing, J. T., Campbell, J., Wilson, J. S.,
Brown, S., Patchell, B., & Shall, C. (2014).
Police departments use of the lethality assessment program: A quasi-experimental program.
Washington, DC: U.S. Department of Justice.
(Document No. 247456).
Messing, J. T., Cimino, A., Campbell, J. C.,
Brown, S., Patchell, B., & Wilson, J. S. (2011).
Collaborating with police: Recruitment in
the Oklahoma Lethality Assessment (OKLA) study. Violence Against Women, 17(2),
163-176.
Messing, J. T., & Thaller, J. (2014). Intimate
partner violence risk assessment: A primer
for social workers. British Journal of Social
Work, 1-17. http://dx.doi.org/10.1093/bjsw/
bcu012
Nicolaidis, C., Curry, M. A., Ulrich, Y., Sharps,
E., McFarlane, J., & Campbell, D. (2003).
Could we have known? A qualitative analysis of data from women who survived an
attempted homicide by an intimate partner. Journal of General Internal Medicine, 18,
788-794.
77
Otto, R. K., & Douglas, K. S. (2010). Handbook of violence risk assessment. New York:
Routledge.
Paulozzi, L. J., Saltzman, L. E., Thompson, M. P.,
& Holmgreen, P. (2001). Surveillance for
homicide among intimate partners—United
States, 1981-1998. Morbidity and Mortality
Weekly Surveillance Summaries, 50, 1-16.
Ross, L. E., & Kane, K. L. (2014). Exploring
the utility of actuarial assessment in cases of
intimate partner homicide. Law Enforcement
Executive Forum, 14(2), 44-58.
Ross, L. E., & Leslie, T. (2014). Criminal justice
practitioner attitudes toward domestic violence: Another day in paradise. Law Enforcement Executive Forum, 14(3), 18-32.
Storey, J. E., Gibas, A. L., Reeves, K. A., &
Hart, S. D. (2011). Evaluation of a violence
risk (threat) assessment training program for
police and other criminal justice professionals.
Criminal Justice and Behavior, 38(6), 554-564.
http://dx.doi.org/10.1177/0093854811403123
Wilson, J. S., & Websdale, N. (2006). Domestic
violence fatality review teams: An interprofessional model to reduce deaths. Journal of
Interprofessional Care, 20(5), 535-544.
Withrow, B. L. (2014). Research methods in
crime and justice. New York: Routledge.
Contact Information
Lee E. Ross, PhD
Associate Professor
Department of Criminal Justice
University of Central Florida
12805 Pegasus Drive
Orlando, FL 32718-1600
(407) 823-0757
[email protected]
78
Detection and Prevention of Racial Profiling
Practices: Case Study of a Medium-Sized
City in Texas
Won-Jae Lee, PhD, Department of Security Studies and Criminal Justice,
Angelo State University
Shawn S. Morrow, PhD Candidate, Department of Security Studies and
Criminal Justice, Angelo State University
Seungmug (Zech) Lee, PhD, School of Law Enforcement and Justice
Administration, Western Illinois University
Abstract
Empirical literature, beyond descriptive analyses, on mandatory racial profiling reports pursuant to the
Texas Law on Racial Profiling is scant. Using aggregate citation-based stop and consent search following
citation data from a medium-sized city in Texas and Municipal Court, and baseline population data derived
from the Fair Roads Standard, this study is designed to seek and provide a more accurate and sophisticated
analysis in determining racial profiling practices. In addition to a typical descriptive analysis, the two
more robust analytical techniques—racial distribution analysis and logistical regression—were utilized to
determine the existence of institutional racial profiling in citation-based stops and to determine whether
the race of residential drivers is the determinate of racial profiling that occurred during consent searches
following citations. The findings indicate that there is little evidence to substantiate that both White and
minority officers, as a whole, were systematically engaging in racial profiling practices, but inconsistent
with anecdotal findings, minority officers are more likely than White officers to perform consent searches of
minorities. Discussions and policy suggestions will be provided to help police administrators better recognize the importance of accurate and thorough racial profiling accountability to the public.
Introduction
The widespread media coverage of racial
conflict, such as the Rodney King incident in
California and the Michael Brown incident in
Missouri, continues to intensify tension and
promote a lack of trust between law enforcement and local residents. Similarly, the issue
of racial profiling has become one of the most
controversial and sensitive issues nationwide.
Racial profiling practices during a police–citizen encounter result in citizens questioning
police legitimacy coupled with the procedural fairness of police organizations and
thereby undermine the minority community’s attitudes about the police (Engel, 2005;
Lundman & Kaufman, 2003; Smith & Holmes,
2003). A number of racial profiling studies
have revealed minorities’ belief in the existence of racial profiling and their experience
of racial profiling practices in police stops,
searches, and arrests, eventually leading to
their negative attitudes toward the police
(Huebner, Schafer, & Bynum, 2004) and their
lower levels of trust in and cooperation with
the police than Whites (Brunson, 2007; Reitzel
& Piquero, 2004; Stewart, 2007; Weitzer &
Tuch, 1999).
In addition to research primarily focusing on
traffic stops, Iomio and his associates (2007), in
79
their survey of the police’s view of bias-based
policing that occurred in police departments
in Virginia, found that 21% of survey respondents believed that bias-based policing is
presently practiced by officers in their department. Over time, racial profiling practices on
the streets and bias-based policing practices
in police departments will lead to the overall
failure of both effective community policing
and crime reduction. This erosion of the relationships between the community and police
departments suggests a need for an aggressive response by police administrators.
In the year 2002, the Texas legislature, in an
attempt to address the issue of racial profiling
in policing, passed the Texas Law on Racial
Profiling (Senate Bill 1074), which mandates
that all law enforcement agencies in Texas
develop a policy prohibiting racial profiling,
compile stop data, and provide an annual
report to the governing body of the municipality or county. Passing legislation monitoring
discrimination that may occur could help officers participate in corrective acts while performing their official duties (Nier, Gaertner,
Nier, & Dovidio, 2011). A thorough analysis
of the gathered data for citation-based stops,
searches, and arrests offers the opportunity
for police departments to proactively recognize and address racial profiling issues, and,
in turn, build trust between the police and the
citizens.
However, most police self-reports on racial
profiling pursuant to Senate Bill 1074 appear
to be limited to only a descriptive analysis
which is simplistic and suggests disparity
between the racial proportions of those cited
and the baseline population. Unfortunately,
the report with the descriptive data analysis and interpretation is not a valid and conclusive accounting and will not provide the
public with a reliable and accurate view of
racial profiling nor of the racial disparity in
citation-based stops and consent searches.
Since findings from the descriptive analysis
cannot determine if racial disparity occurred
80
during citations and consent search following citations, data were produced primarily
from the police racial profiling practices
(Fridell, 2004; McMahon, Garner, Davis, &
Kraus, 2002) or by other significant factors
(Batton & Kadleck, 2004) such as legal variables (Klinger, 1994; Tillyer & Engel, 2013);
officer characteristics, including officer race
(Cochran & Warren, 2012; Withrow, 2004); situational characteristics (Alpert, Dunham, &
Smith, 2007; Falik & Novak, 2012; Novak, 2004;
Pickerill, Mosher, & Pratt, 2009; Withrow,
2004; Worden, McLean, & Wheeler, 2012);
race-sensitive police deployment (Withrow,
2004); social characteristics of minority neighborhoods (Carroll & Gonzalez, 2014; Smith
& Alpert, 2007; Tillyer & Engel, 2013); subtle
cognitive bias (Tomaskovic-Devey, Mason,
& Zingraff, 2004); and so on. In response,
this study is designed to seek and provide
a more accurate and sophisticated analysis in determining racial profiling practices.
Accordingly, in addition to a typical descriptive analysis, the two more robust analytical
techniques—racial distribution analysis and
logistical regression—were utilized to determine the existence of institutional racial profiling in citation-based stops and to ascertain
whether the race of residential drivers is the
determinate of racial profiling that takes place
during consent searches following citations.
Data Analyses and Findings
Baseline Population
One medium-sized city in Texas with a population of about 100,000 was used to collect
data for one year. Pursuant to Senate Bill
1074, the police department citation-based
stop data must be compared against a baseline population to assess the occurrence of
racial profiling. Without a valid baseline population, it is impossible to determine if any
race is being cited and consent-searched at a
disproportionate rate. However, little consensus exists regarding how to conduct a valid
and appropriate baseline population analysis.
In this study, the methodologies that provided racial mix data and, thus, were considered were (1) Texas licensed drivers, (2) The
U.S. Census Bureau, and (3) The Fair Roads
Standard. Of the three available population
data resources, only the Fair Roads Standard
based upon data collected in the American
Community Survey was utilized for this analysis as the most relevant data available for
the baseline population since it provides an
adjusted estimate that includes the number
of racial households with vehicle accessibility. It should be noted that the data from the
Texas licensed drivers do not allow for a separate data comparison for White or Hispanic
licensed drivers, while the U.S. Census Bureau
data are not adjusted for the eligible driving
population.
By using the Fair Roads Standard data for the
baseline population, the nonresident driving population cited (Total: 3,684) makes up
approximately 20% out of the total driving
population cited (Total: 18,166). In a comparison between the racial proportions for
each of the two different driving populations
(resident drivers and nonresident drivers)
that were cited, Figure 1 indicates that the
racial proportion of the cited nonresident
driving population is considerably closer to the
baseline population than the resident driving
population who were cited, suggesting that
if the nonresident driving population is combined with the resident driving population, a
possibility exists for a reduction of racial disparity between the cited resident drivers and
the baseline population. Thus, excluding all
of the nonresident driving population from
the analyses in this study is more reasonable.
Descriptive Analyses Through
Comparisons
Among the 14,482 citation-based stops
included within the resident population,
Table 1 presents the comparison between citation-based traffic stops and baseline population and indicates that White residents were
stopped and received a citation at a disproportionately lower rate, while Hispanic and
Black residents were stopped and cited at a
disproportionately higher rate.
The racial breakdown in Table 2 presents data
for consent searches after citation-based stops
by percentages of the 197 consent searches
after citation-based stops that occurred. It
was found that 197 of 14,482 resident drivers received some type of traffic citation that
Figure 1: Traffic Stop and Baseline Population Rates by Race
81
Table 1. Racial Breakdown of Traffic Stop and Baseline Population Rates
No. of stops
Percentage of total stops
Percentage of baseline population*
Difference
White
7,549
52.13
66.07
-13.94
Hispanic
6,079
41.98
29.06
12.92
Black
788
5.44
3.61
1.83
Asian
60
0.41
0.67
-0.26
Native
American
6
0.04
0.60
-0.56
Total
14,482
100.00
100.00
* Fair Roads Standard population obtained from the American Community Survey adjusted for households with vehicle access
Table 2. Racial Breakdown of Consent Search and Traffic Stop Rates
No. of consent searches
Percentage of total consent searches
Percentage of total stops
Difference
Resident Driving Population by Race/Ethnicity
Native
White Hispanic
Black
Asian
American
65
100
32
0
0
32.99
50.76
16.12
0
0
52.13
41.98
5.44
0.41
0.04
-19.14
8.78
10.68
-0.41
-0.04
involved a consent search. However, 50.8%
of all consent searches were performed on
Hispanic drivers, followed by White (33%)
and Black (8.1%). Sixty-seven percent of
consent searches involved Hispanic (50.8%)
and Black (16.1%) drivers. During the same
period, 47% of the total traffic stops that
received a citation were Hispanics (42%) and
Blacks (5.4%). The results indicate that both
Hispanics and Blacks were more likely to be
the subjects of a consent search than Whites.
Summary and Limitations of the
Descriptive Analysis
The purpose of the descriptive analysis is
to examine if minorities (Hispanic or Black)
are being disproportionately stopped and
consent-searched by the city police. The
results from each of these analyses suggest
an unequal usage of citations and consent
searches. Specifically, White residents were
underrepresented in the racial proportions
of residential drivers cited and cited then
consent-searched, while both Hispanic and
African-American residents were overrepresented compared to the racial breakdown of
the baseline population. However, although
82
Total
197
100.00
100.00
the findings from the descriptive analysis can
determine the existence of racial disparity, it
is impossible for them to determine if there is
evidence of racial profiling.
Racial Distribution Analysis in CitationBased Stops and Findings
Although racial profiling may be practiced
by some racially prejudiced police offers, it
may also be considered a form of institutional
racial profiling, which police departments
first create and which, because of this, police
officers have cognitive biases and act on a set
of racial characteristics (Tomaskovic-Devey
et al., 2004). Following their hypothetical
models, this analysis utilized two models—
(1) no racial bias and (2) racial bias mechanisms—to assess and determine the existence of institutional racial profiling in citation-based stops.
Based upon an odds ratio, by utilizing all
officer-level counts of the residential race
distribution of citation-based stops and the
baseline population (Fair Roads Standard
population indicative of race composition of
at-risk residential drivers), Figure 2 shows
Figure 2. Hypothetical Race Distribution of Citation-Based Stops Without Racial Bias
Figure 3. Race Distribution of Citation-Based Stops Indicative of Racial Profiling
and compares both hypothetical models (see
Tomaskovic-Devey et al., 2004, for each model’s computation formula). Specifically, the
hypothetical model in Figure 2 denotes the
absence of racial profiling in citation-based
stops. For both minority and White officers,
the distribution is centered on the odds ratio
of 1.0. Interpreted, the odds of a minority or
a White being cited, adjusting for the race
composition of at-risk drivers, are equal. The
distribution, in the absence of racial bias, generally centers on or around an even minority–
White odds of citation-based stops. Hence,
when there is no institutional racial profiling,
it is reasonable to expect all officers to stop
and cite residential minority and White drivers with nearly equal probabilities.
In contrast, the hypothetical model in Figure 3
presents the existence of institutional racial
profiling in citation-based stops. For both
minority and White officers, the distribution is
centered beyond the odds ratio of 1.0. This is,
by assumption, that racial profiling produces
organizational rules that are presumably followed by most or all individual officers and
that encourage all officers to stop and cite
83
minority drivers in racially biased ways as a
standard practice.
Figure 4 shows the race distribution of citation-based stops by White and minority officers in the police department. Each distribution of the White and minority officers has a
right skew, indicating that minority drivers
were more likely to be stopped and cited than
White drivers by both White and minority
officers. However, this skew is even with more
than a half odds ratio, which is centered on
and around the ratio of 1.0, indicating a positive disposition of the citation-based stops
toward minority drivers. Therefore, there is
little evidence of institutional racial profiling
in citation-based stops.
Two Sets of Logistic Regression Analyses
and Findings
The findings from the descriptive analysis indicate that both Hispanics and Blacks
were overrepresented in traffic citations and
consent search following the citation-based
stops, while Whites were underrepresented.
In response, there are two primary concerns:
(1) which factors, especially officer race,
predict each race of vehicle drivers in citation-based stops and (2) whether the race of
vehicle drivers is the determinate of racial
disparity in consent searches following the
citation-based stop. Implementing two sets
of logistical regressions, all variables were
drawn from anecdotal findings (Alpert et al.,
2007; Cochran & Warren, 2012; Falik & Novak,
2012; Pickerill et al., 2009; Worden et al., 2012)
and are expected to be meaningful to prove
the primary concern. Different from removing nonresidential drivers to secure the valid
racial disparity between the resident driving
population cited and the baseline population
in both the descriptive and racial distribution
analyses, in these logistical regression analyses, the residency of driver is included in order
to test anecdotal findings (Withrow, 2004) that
nonresident drivers, especially minorities, are
significantly more likely to be cited and consent-searched than resident drivers.
Table 3 presents three logistical regression
models for predicting a driver’s race. In
Model 1, the dichotomous dependent variable, White drivers stopped to receive a citation,
Figure 4. Race Distribution of Citation-Based Stops in the City
84
was regressed on a set of variables representing officer characteristics and situational
characteristics. Only one variable, officer race,
was not found to be a significant predictor
of White drivers stopped to receive a citation, indicating that the officer’s race (White
or minority) was an insignificant factor in
predicting the race of the White drivers. In
contrast, the other five factors in Model 1
were found to be significant in predicting
the race of the White drivers. Overall, given
each value of Exp [B], the first three significant predictors in the officer characteristics
(officer sex, age, and division) were negligibly associated with White drivers. Two situational characteristics provided a greater
explanatory value than did that of the officer
characteristics. White drivers were 1.5 times
more likely to be stopped to receive a citation during the daytime (B = 0.404, p < 0.001),
and nonresident White drivers were 2.1 times
(1/0.472; Exp [B] = 0.472) more likely to be
stopped and to receive a citation (B = -0.751,
p < 0.001).
Model 2 in Table 3 is a logistic regression
model of factors predictive of Hispanic drivers who were stopped and received a citation.
Out of the four officer characteristics, only the
officer’s age was found to be a significant predictor of Hispanic drivers who were stopped.
However, it did provide a negligible association with the dichotomous dependent variable of Hispanic drivers (Exp [B] = 0.992). In
contrast, the situational characteristics, time
of day, and residency of citizen indicated a
greater explanatory value than did the officer characteristics. Contrary to the findings
in Model 1, Hispanic drivers were 1.4 times
(1/0.731; Exp [B] = 0.731) more likely to be
stopped to receive a citation at nighttime
(B = -0.313, p < 0.001), and resident Hispanic
drivers were approximately 2 times more
likely to be stopped to receive a citation
(B = 0.699, p < 0.001).
The final model in Table 3 presents a logistical regression analysis to predict the dichotomous dependent variable of Black drivers
who were stopped and received a citation.
Neither the officer’s sex nor the officer’s division characteristics were correlated or associated to the Black drivers who were stopped
and received a citation. In contrast, both the
officer’s race and the officer’s age were found
to be significant predictors of Black drivers
who were stopped and received a citation.
The officer’s age was negligibly related to the
Black driver who was stopped and received a
citation, however. Instead, the officer’s race,
the variable of real interest for this analysis,
was significantly correlated to the Black drivers who were stopped and received a citation—Minority officers (Hispanic or Black) were
1.3 times more likely than White officers to stop
Black drivers and issue citations.
Consistent with the findings in Models 1 and 2,
two situational characteristics—time of day
and the residency of the citizen—provide an
interesting value other than the officer’s characteristics. Contrary to the findings in Model 1
but similar to the findings in Model 2 , Black
drivers were 1.6 times (1/0.615; Exp [B] = 0.615)
more likely to be stopped to receive a citation
at nighttime (B = -0.486, p < 0.001), and resident
Black drivers were approximately 1.8 times
more likely to be stopped to receive a citation
(B = 0.576, p < 0.001).
The three findings are worth mentioning.
First, only a small percentage of the variance in each race of vehicle drivers who were
stopped to receive a citation, in each model
(4.4% in Model 1, 3.3% in Model 2, and 1.8%
in Model 3) was taken into account. However,
in each mode, the situational characteristics
provided a greater explanatory value than
did that of the officer characteristics, suggesting that situational characteristics make
a more substantial contribution to predicting
each race of drivers stopped to receive a citation than officer characteristics.
Second, the focus of this analysis is the
officer’s race since racial profiling may be
less likely to be committed by a minority
(Hispanic or Black) officer than by a White
85
86
Officer Characteristics
Officer race (Hispanic or Black = 1)
Officer sex (male = 1)
Officer age (years)
Officer division (patrol division = 1)
Situational Characteristics
Time of day (daytime or high visibility = 1)
Residency of citizen (resident = 1)
X2
2LL
Nagelkerke R2
a
Whites = 1; other = 0
b
Hispanics = 1; other = 0
c
African Americans = 1; other = 0
* p < 0.05; ** p < 0.01; *** p < 0.001
Variables
0.050
0.072
0.002
0.045
0.927
0.848
1.010
1.097
0.036
1.498
0.040
0.472
594.240***
23,922.194
0.044
0.404***
-0.751***
-0.075
-0.165*
0.010***
0.092*
Model 1
Whitesa (n = 10,137)
B
S.E.
Exp (B)
0.051
0.073
0.002
0.045
-0.313*** 0.036
0.699*** 0.042
440.625***
23,398.460
0.033
0.005
0.123
-0.008***
-0.040
0.731
2.013
1.005
1.130
0.992
0.961
Model 2
Hispanicsb (n = 7,037)
B
S.E.
Exp (B)
Table 3. Logistic Regression Analysis for Correlates Predicting Driver's Race
0.108
0.156
0.005
0.098
1.263
1.048
0.987
0.829
-0.486*** 0.076
0.615
0.576*** 0.104 1.779
107.068***
6,911.440
0.018
0.233*
0.046
-0.013**
-0.187
Model 3
African Americansc (n = 907)
B
S.E.
Exp (B)
officer. However, the finding in Model 3—
Minority officers were 1.3 times more likely than
White officers to stop and cite Black drivers—provides an interesting contrast with the findings in Models 1 and 2, which indicate no
significant association between the race of the
officer and the other races of drivers (Whites
and Hispanics). These mixWed findings are
not consistent with anecdotal findings. Also
inconsistent with the anecdotal findings, this
analysis conclusively found that both White
and minority police officers were more likely
to stop and cite minorities at night and White
drivers during the day. The conclusion here
is that there appears to be no racial profiling
practices against Black drivers cited since it
is undoubtedly more difficult to identify the
race or ethnicity of drivers during the nighttime hours (Novak, 2004; Worden et al., 2012).
Lastly, it may be expected that officers are
more likely to stop and cite nonresidents,
especially minorities by racially motivated
officers, since there is less risk of being criticized for disparate enforcement or racial profiling practices. Contrary to this assumption
and based on anecdotal evidence, minorities
living in the city were significantly more
likely to be stopped and cited than nonresident minorities. This indicates that officers
did not differentially target minority nonresidents. Collectively, the finding provides
an interesting contrast with anecdotal findings. A conclusion that can be drawn is that
although there was a disparity between the
racial breakdown of citation-based stops and
the baseline population, the officer’s race
did not produce a disparity within all citation-based stops.
Note that the findings from the descriptive
analysis indicate that both Hispanics and
Blacks were overrepresented among persons
who were selected for consent searches, while
Whites were underrepresented. Relatedly, an
important question is whether the race of a
Table 4. Logistic Regression Analysis for Correlates Predicting Consent Search
Variables
Driver Variables
Race of driverb (Hispanic or Black = 1)
Driver sex (male = 1)
Driver age (years)
Officer Characteristics
Officer raceb (Hispanic or Black = 1)
Officer sex (male = 1)
Officer age (years)
Officer division (patrol division = 1)
Situational Characteristics
Time of day (day time or high visibility = 1)
Residency of citizen (resident = 1)
Disposition of Traffic Stop
Consent search (yes = 1)
Probable cause search (yes = 1)
X2
-2LL
Nagelkerke R2
a
yes = 1; no = 0
b
Hispanic or Black = 1; White = 0
* p < 0.05; ** p < 0.01; *** p < 0.001
B
(Total 224 Consent Searchesa)
S. E.
Exp (B)
0.637***
1.025***
-0.012*
0.150
0.180
0.006
1.891
2.787
0.988
1.015***
0.789*
-0.062***
0.162
0.194
0.399
0.010
0.252
2.758
2.201
0.940
1.176
-1.813***
0.499*
0.179
0.222
0.163
1.646
389.277***
1,945.715
0.176
87
vehicle driver, after being stopped and cited,
produces racial disparity in police consent
searches. As presented in Table 4, logistic
regression analysis was employed to identify
significant variables in predicting a consent
search following a citation-based stop. The
dichotomous dependent variable of consent
searches of minority drivers as a disposition
of the citation-based stop was regressed on a
set of the variables. The eight significant predictors included driver variables (e.g., race,
sex, and age), officer characteristics (e.g., race,
sex, and age), and situational characteristics
(e.g., time of day and residency of citizen). The
eight significant predictors were found to be
17.6% of the variance in the consent search of
minority drivers following the citation-based
stop (Nagelkerke R2 = 0.176).
Of the three significant driver variables, the
driver age (Exp [B] close to 1) was negligibly
associated with the consent search. On the
other hand, the minority drivers (B = 0.637,
p < 0.001) and male drivers (B = 1.025, p < 0.001)
were significantly more likely to be the subjects of the consent search than White, female
drivers (1.9 and 2.8 times, respectively).
Minorities, especially male, were being disproportionately targeted for consent searches
by the police department. The results suggest
that the race of the driver, in the best interest
for this analysis, produced the disparity in the
consent searches.
Among the three significant officer characteristics, the officer’s age was found to be a significant predictor but provided a negligible association with the consent search (Exp [B] = 0.940).
In contrast, the minority (B = 1.015, p < 0.001)
and male (B = 0.789, p < 0.05) officers were significantly more likely than White, female officers to perform consent searches (2.7 and 2.2
times, respectively). That is, minority officers,
especially male officers, conducted consent
searches at a disproportionately higher rate as
compared with the White officers, indicating
that the race and gender of an officer did produce the disparity when a consent search was
conducted following a citation-based stop.
88
Each of the variables in the situational characteristic was found to be a significant predictor
of a consent search. Specifically, the consent
searches that occurred during the nighttime
hours were approximately 6.13 times (1/0.163;
Exp [B] = 0.163) more likely to be conducted
than those occurred during the daytime
(B = -1.813, p < 0.001), and resident drivers
were 1.7 times more likely to be subjected to
a consent search than the nonresident drivers
(B = 0.499, p < 0.05).
It is important to note three important findings.
First, the proportion of variance explained
by the model in predicting consent search
(Nagelkerke R2 = 0.176) substantially exceeds
that explained by each of the three models in
predicting driver’s race (Nagelkerke R2 = 0.044
in Model 1; 0.033 in Model 2; and 0.018 in
Model 3). Compared to the interplay of officer
and situational characteristics in predicting
driver’s race, the model in predicting consent
search focused on the interplay of driver, officer, and situational characteristics. Adding
driver characteristics appears to explain the
difference in the portion of variance. Second,
in this analysis, minority drivers were 1.9
times more likely to be subjected to the consent search than White drivers. This result is
consistent with anecdotal findings (Cochran
& Warren, 2012; Falik & Novak, 2012; Gibson,
Walker, Jennings, & Miller, 2010; Pickerill
et al., 2009) that Black and Hispanic drivers
were consent-searched following a traffic stop
at a significantly higher rate than statistically
expected.
Lastly, it is reasonable to assume that minorities were significantly more likely to be the
subjects of a consent search conducted by
White officers (Cochran & Warren, 2012; Falik
& Novak, 2012; Gibson et al., 2010; Pickerill
et al., 2009). Inconsistent with anecdotal findings, however, a further descriptive analysis
found that 75% of all consent searches by
minority officers targeted minorities compared to 65.5% of all consent searches conducted by White officers with minorities. This
finding is evidence that an officer’s race is a
significant predictor of a consent search following a stop-based citation. Further analysis
found minority officers were 2.7 times more
likely than White officers to perform consent
searches of minorities. Taken together, from a
statistical point of view, these findings have
led to the conclusion that the interesting interplay of a driver’s race and officer’s race and
gender in consent searches directly following
citation-based stops that occurred during the
nighttime hours did produce a racial disparity in consent searches by minority male officers against minority drivers, suggestive of
racial profiling practices.
Discussion
This study is an attempt to establish an accurate but thorough empirical benchmark to
assess one medium-sized city in Texas and
address and prevent racial profiling in the
city. There are three different analytical statistics employed for this study: (1) descriptive,
(2) racial distribution, and (3) logistical regression analyses. The overall analyses from three
analytical sections lead to three important
findings. First, compared to the results from
the descriptive analysis suggesting disparate
treatment in citations, the results from the first
logistic regressions indicate no evidence that
the police officers, especially White officers,
differentially stopped minority resident drivers. Thus, it can be concluded that the race of
the driver is not a significant factor to racial
disparity in traffic stop-based citations.
Second, in the race distribution of citation-based stops by White and minority officers, the right skew around the odd ratio
of 1.0 in each distribution of the White and
minority officers may indicate that minority
drivers were more likely to be stopped and
cited than White drivers by both White and
minority officers. Cognitive stereotyping can
make the best plausible explanation about this
slightly negative finding from TomaskovicDevey and his associates (2004). The cognitive stereotyping, as a subtle unconscious
bias process, might lead some officers to be
racially prejudiced or operate with a cognitive
bias that may result in minority drivers being
stopped at a higher rate than White drivers.
Thus, it can be concluded that there is little
evidence for either White or minority officers,
or their department, to be accused of institutional racial profiling practices as a standard practice of stopping and citing minority
drivers.
Third, consistent with anecdotal findings,
67% of all consent searches following traffic
citation stops (47% of Hispanics and 5.4% of
Blacks who were stopped and received a citation during the same period) of the total traffic
citation stops that were given a citation were
Hispanics (42%) and Blacks (5.4%), suggesting racial disparity in the consent searches. In
contrast, the results from the logistic regression analysis in predicting consent search
following a citation-based stop does not support anecdotal findings that White officers are
more likely than minority officers to perform
consent searches of minorities.
It is very difficult to conclude whether there
were racial profiling practices by minority officers that occurred during the consent search
as a post-stop activity since 83% of variance
was unexplained by the logistic regression
model in predicting consent search. The lack
of statistical explanation hardly supports the
finding that the officer’s race is a determining
factor for racial disparity in consent searches.
Accordingly, the results should be interpreted
with some caution, and there are two plausible explanations. The first explanation is that a
de-policing phenomenon occurs. That is, officers tend to curtail their proactive practices
in the post-stop activity. In fact, compared
to minority officers, White officers are more
likely to fear being accused of racial profiling
practices, which leads to de-policing even after
stopping minority drivers. This may be a plausible explanation but only partially at best.
A better and more plausible explanation is
related to both police deployment and social
characteristics of minority neighborhoods.
89
Police departments deliberately tend to assign
minority officers to the same racial minority
neighborhoods based on the assumption
that minority officers are more capable and
can relate to minority citizens and, thus, will
not engage in racial discriminatory practices
(Cochran & Warren, 2012). Also, as noted by
Carroll and Gonzalez (2014), Smith and Alpert
(2007), Tillyer and Engel (2013), TomaskovicDevey et al. (2004), and Withrow (2004), effective deployment depends on crime rate, calls
for service, and population density, which,
in fact, tend to be higher in lower socioeconomic class communities largely populated
by minorities. Unfortunately, these minority
neighborhoods are more likely to be inadvertently subjected to more patrol resource-based
aggressive policing and, accordingly, increase
the likelihood of police–citizen encounters
such as disproportionately higher rates of
police stop and post-stop dispositions. The
race-sensitive police deployment coupled
with the social characteristics of minority
neighborhoods might make minority officers
assigned to minority neighborhoods appear
to be more likely to conduct consent searches
of minority drivers than White officers.
The present study has one limitation which
should be addressed in future studies. Despite
the important impact of social characteristics
on an officer’s decision to stop a vehicle and
the related post-stop activity, it is not statistically possible to reveal by analysis a correlation between U.S. census tract areas (conceptually, indicative of social characteristics of
the tract area) and police beat areas (conceptually, indicative of patrol deployment strategy and patterns in each police beat in the
city). These additional variables in a multivariate-level analysis will lead to a better statistical explanation for predicting racial profiling
practices (Carroll & Gonzalez, 2014; Smith &
Alpert, 2007; Tillyer & Engel, 2013).
Conclusion
Racial profiling is possibly the most significant threat to solid police–community
90
relationships. A prevalent theme among law
enforcement agencies in the U.S. is the necessity to provide the general public with accurate data and analysis of racial profiling practices among police officers. In response, the
overall analyses and findings lead to the conclusion that there is little evidence to substantiate that both White and minority officers,
as a whole, were systematically engaging in
racial profiling practices in both traffic citation-based stops and the subsequent consent
searches.
Despite the positive conclusion, this study
provides two policy implications for law
enforcement agencies. All of the publicly
available reports reveal disparities in the percentages of minority citizens who are stopped
and consent-searched in comparison to select
benchmarks. In response, as evidenced by
the findings of this study, law enforcement
agencies need to utilize more than descriptive analysis to avoid being easily accused of
racial profiling practices. Also, it is strongly
recommended that Tomaskovic-Devey and
associates’ (2004) racial distribution analyses be used as an internal benchmark (also
known as early warning system) against officers’ racial profiling practices. As presented
in this study, these analyses are useful for
monitoring and preventing any improper
racial profiling practices and patterns by both
White and minority officers, and for visually
comparing an individual officer’s data with
the department, division, unit, and beat data,
eventually reducing subtle cognitive stereotyping against minorities.
Inevitably, as noted by Fridell (2004) and
Fridell et al. (2004), law enforcement administrators should recognize the importance
of accurate and thorough racial profiling
accountability to the public and come up with
institutional policies with which officers must
fully comply. Otherwise, they essentially overlook any police misconduct in stop, search,
and arrest procedures primarily based on
race alone, rather than on any individualized
suspicion, and, thus, contribute to an eroding
public confidence in the police.
References
Alpert, G., Dunham, R., & Smith, M. (2007).
Investigating racial profiling by the MiamiDade police department: A multimethod
approach. Criminology and Public Policy, 6,
25-56.
Batton, C., & Kadleck, C. (2004). Theoretical
and methodological issues in racial profiling research. Police Quarterly, 7, 30-64.
Brunson, R. (2007). “Police don’t like black
people”: African American young men’s
accumulated police experiences. Criminology and Public Policy, 6, 71-102.
Carroll, L., & Gonzalez, M. L. (2014). Out of
place: Racial stereotypes and the ecology of
frisk and searches following traffic stops.
Journal of Research in Crime and Delinquency,
51, 559-584.
Cochran, J. C., & Warren, P. Y. (2012). Racial,
ethnic, and gender differences in perceptions of the police: The salience of officer
race within the context of racial profiling.
Journal of Contemporary Criminal Justice, 28,
206-227.
Engel, R. S. (2005). Citizen’s perceptions of
distributive and procedural injustice during
traffic stops with police. Journal of Research
in Crime and Delinquency, 42, 445-481.
Falik, S. W., & Novak, K. J. (2012). The decision to search: Is race or ethnicity important? Journal of Contemporary Criminal Justice,
28, 146-165.
Fridell, L. (2004). By the numbers: A guide
for analyzing race data from vehicle stops.
Washington, DC: Police Executive Research
Forum.
Fridell, L., Lunney, R., Diamond, D., &
Kubu, B., with Scott, M., & Laing, C. (2004).
Racially biased policing: A principled response.
Washington, DC: Police Executive Research
Forum.
Gibson, L. G., Walker, S., Jennings, W. G., &
Miller, J. M. (2010). The impact of traffic
stops on calling the police for help. Criminal
Justice Policy Review, 21, 139-159.
Huebner, B. M., Schafer, J. A., & Bynum, T. S.
(2004). African American and white perceptions of police services: Within- and
between-group variation. Journal of Criminal
Justice, 32, 123-135.
Iomio, R., Tears, R. S., Meadows, L. A., Becton,
J. B., & Charles, M. T. (2007). The police view
of bias-based policing. Police Quarterly, 10,
270-287.
Lundman, R. J., & Kaufman, R. L. (2003).
Driving while black: Effects of race, ethnicity, and gender on citizen self-reports of
traffic stops and police actions. Criminology,
41, 195-220.
McMahon, J., Garner, J., Davis, R., & Kraus,
A. (2002). How to correctly collect and analyze
racial profiling data: Your reputation depends
on it! Washington, DC: U.S. Government
Printing Office.
Nier, J. A., Gaertner, S. L., Nier, C. L., & Dovidio,
J. F. (2011). Can racial profiling be avoided
under Arizona immigration law? Lessons learned from subtle bias research and
anti-discrimination law. Analyses of Social
Issues and Public Policy, 12, 5-20.
Novak, K. (2004). Disparity and racial profiling in traffic enforcement. Police Quarterly,
7, 65-96.
Pickerill, M., Mosher, C., & Pratt, T. (2009).
Search and seizure, racial profiling and traffic stops: A disparate impact framework.
Law and Policy, 31, 1-30.
91
Reitzel, J. D., & Piquero, A. R. (2004). Does it
exist? Studying citizen’s attitudes of racial
profiling. Police Quarterly, 10, 1-23.
Smith, M. R., & Alpert, G. P. (2007). Explaining police bias: A theory of social conditioning and illusory correlation. Criminal Justice
and Behavior, 34, 1262-1283.
Smith, B. W., & Holmes, M. D. (2003). Community accountability, minority threat, and
police brutality: An examination of civil
rights criminal complaints. Criminology, 41,
1035-1063.
Stewart, E. A. (2007). Either they don’t know
or they don’t care: Black males and negative police experiences. Criminology & Public
Policy, 6, 123-130.
Tillyer, R., & Engel, R. S. (2013). The impact of
drivers’ race, gender, and age during traffic
stops: Assessing interaction terms and the
social conditioning model. Crime & Delinquency, 59, 369-395.
Tomaskovic-Devey, D., Mason, M., & Zingraff,
M. (2004). Looking for the driving while
black phenomena: Conceptualizing racial
bias processes and their associated distributions. Police Quarterly, 7, 3-29.
Contact Information
*Won-Jae Lee, PhD
Associate Professor of Criminal Justice
Department of Security Studies and
Criminal Justice
ASU Station #10896
San Angelo, TX 76909
(325) 486-6717
[email protected]
Shawn S. Morrow, PhD Candidate
Instructor of Criminal Justice
Department of Security Studies and
Criminal Justice
ASU Station #10896
San Angelo, TX 76909
(325) 486-6692
[email protected]
Seungmug (Zech) Lee, PhD
Assistant Professor of Criminal Justice
School of Law Enforcement and Justice
Administration
Western Illinois University
Macomb, IL 61455
(309) 298-2746
[email protected]
* Corresponding author
Weitzer, R., & Tuch, S. (1999). Race, class, and
perceptions of discrimination by the police.
Crime & Delinquency, 45, 494-507.
Withrow, B. L. (2004). Driving while different: A potential theoretical explanation for
race-based policing. Criminal Justice Policy
Review, 15, 344-364.
Worden, R. E., McLean, S. J., & Wheeler, A. P.
(2012). Testing for racial profiling with the
veil-of-darkness method. Police Quarterly,
15, 92-111.
92
Law Enforcement Executive Forum
Illinois Law Enforcement Training and Standards Board Executive Institute
Western Illinois University
510 N. Pearl Street, Suite 4000
Macomb, IL 61455
www.ILETSBEI.com
Phone: (309) 298-2646
Fax: (309) 298-2642
Email: [email protected]

Read the article - cohpa - University of Central Florida

Transcription

Similar documents

July 28 - Royal Floridian Suicide

nmgh boggart hole clough moston collyhurst village

OFFICER DOWN:

RoboCam

Iowa`s 2013 Strategic Highway Safety Plan

Abbreviated Bio for Sue Rahr 2015

Tualatin`s Police Officers Annual Physical Fitness Test: ORPAT

How to Prepare for Promotion - International Association of Chiefs of

English - System of Strategy

Elena Ford, Why haven`t you responded to my