Girl delinquency A study on sex differences in - VU-dare

Transcription

Girl delinquency
A study on sex differences in (risk factors for) delinquency
Photography and cover design: Viewtography || www.floorknaapen.com
Cover model: Lotte van der Vleuten
Author: Thessa Wong
ISBN: 978-90-8891-394-5
Printing: Proefschriftmaken.nl || Printyourthesis.com
Published by: Uitgeverij BOXPress, Oisterwijk
© 2012 Thessa Wong
VRIJE UNIVERSITEIT
Girl delinquency
A study on sex differences in (risk factors for) delinquency
ACADEMISCH PROEFSCHRIFT
ter verkrijging van de graad Doctor aan
de Vrije Universiteit Amsterdam,
op gezag van de rector magnificus
prof.dr. L.M. Bouter,
in het openbaar te verdedigen
ten overstaan van de promotiecommissie
van de faculteit der Rechtsgeleerdheid
op woensdag 18 april 2012 om 15.45 uur
in de aula van de universiteit,
De Boelelaan 1105
door
Thessa Margaretha Loisie Wong
geboren te Zeist
promotoren:
copromotor:
prof.dr.mr. C.C.J.H. Bijleveld
prof.dr. J.M. Koot
dr. A. Slotboom
Leescommissie:
prof.dr. M. Dekovic
prof. C. Kruttschnitt
prof.dr. P.H. van der Laan
prof.dr. N.W. Slot
dr. F.M. Weerman
Voor tante Karin
TABLE OF CONTENTS
Chapter 1
Introduction
9
Chapter 2
Prevalence of juvenile delinquency
25
Chapter 3
Risk factors for delinquency in adolescent and young adult
females: A European review
41
Chapter 4
The sex difference in delinquency explained? Differential
exposure and vulnerability in the parent and peer domain
63
Chapter 5
The sex difference in delinquency: Do girls need a bigger
push?
89
Chapter 6
Sex and age differences in the risk threshold for delinquency
115
Chapter 7
General discussion
139
References
157
Publication list
173
Samenvatting (summary)
175
Dankwoord (acknowledgment)
179
Curriculum Vitae
181
7
Chapter 1
Introduction
9
10
For a long time, researchers generally neglected criminal behaviour of girls,
most probably because committing crimes was seen as a male phenomenon
with males outnumbering females in official statistics. However, the last
decades the arrest rates for females have gone up in Western countries, like
the United States and The Netherlands (Steffensmeier & Schwartz, 2009;
Wong, Blom & Van der Laan, in press). This has led to more attention for
girls’ crime involvement by scholars. As the number of arrested and
adjudicated girls is rising, there is an increasing need for knowledge how to
effectively cope with these delinquent females. After all, only when
interventions target adequate criminogenic needs (i.e. important risk factors),
will they be successful in reducing recidivism (Dowden & Andrews, 1999).
Although researchers only recently became more interested in girl
delinquency, there has been some research in the past about sex differences
in delinquency (e.g. Bontekoe, 1984; Bruinsma, & Lissenberg, 1987;
Dijksterhuis & Nijboer, 1987; Mertens, Grapendaal, & Docter-Schamhardt,
1998; Rutenfrans, 1989). However, this previous research had many
limitations and therefore not much is known yet about the development and
correlates of delinquent behaviour of girls. Without a doubt, most of the
research on juvenile delinquent behaviour has focused on boys and much
more is known about the origin of their behaviour. Theories about juvenile
delinquency and interventions to diminish it are for that reason mostly based
on research about males’ delinquent behaviour. And since the risk and
promotive factors for girl delinquency are not decisively identified yet, it is
not clear whether these theories and interventions are suitable for female
delinquency.
The first aim of the current dissertation is therefore to examine which
factors are related to delinquency of girls and which of these are similar to
those for boys. Furthermore, since the sex difference in the prevalence of
delinquency is so prominent, but so little understood, a second aim of this
dissertation is to search for explanations why girls are less involved in
criminal behaviour than boys.
How to study girl delinquency?
Different approaches have been used to study girl delinquency. In the current
paragraph three approaches (i.e. the add-and-stir, the female-only-focused
and the cumulative risk approach) will be described and related to girl
delinquency research in general.
Add-and-stir approach
The first approach, the so-called add-and-stir approach, is embraced by most
researchers and studies the contribution of mainstream theories – mostly
developed in research about male delinquency – to account for delinquent
behaviour of girls (Hoyt & Scherer, 1998). In other words, this approach
assumes that girls become delinquent for the same reasons as boys.
11
According to the control theory of Hirschi (1969), for instance,
everybody is inclined to commit crimes, boys as well as girls. The reason
that most people do not, is that they have learned that they might jeopardize
the bonding they have with society when they become delinquent. This
bonding consists of four elements: attachment to parents, teachers and peers;
commitment to conventional institutions such as school or work; involvement
in conventional activities such as sports; and belief in the validity of moral
values and social norms. Although all four elements have a unique influence
on delinquent behaviour, they also appear to be related to each other
(Chapple, McQuillan, & Berdahl, 2005). Hirschi claims that these
mechanisms can be applied to both males and females, but that girls,
compared to boys, have higher quality relationships with friends and parents
and achieve better in school. Because of their stronger bonds, girls have a
lower tendency to commit offences.
Another example of a mainstream theory often used to explain girl
delinquency – also originally developed to explain delinquent behaviour of
boys – is Moffitt’s developmental taxonomy (1993). According to Moffitt, at
least two different types of offenders can be distinguished: life-coursepersistent delinquents and adolescence-limited delinquents. Life-coursepersisters, on the one hand, are thought to have behavioural problems
already early in life and commit offences from early ages through adulthood.
Their criminal behaviour is not only persistent, but also severe. The
explanation for their behaviour is that they suffer from neurological deficits
that are originated in pregnancy or were incurred during birth or after birth.
These deficits can be manifested in cognitive, physical or behavioural
problems. Furthermore, these juveniles are generally born in a high-risk
environment that interacts cumulatively with their neuropsychological
problems.
Adolescence-limited offenders, on the other hand, show less severe
delinquent behaviour and only commit offences during puberty when they
experience a gap between their social and biological maturity, the so-called
‘maturity gap’. During this maturity gap, these adolescents want to be
independent from their parents and show rebellious behaviour, such as
committing (minor) offences. Delinquent behaviour can thus be explained by
situational risks, not by background or personality risks. The group of
adolescent-limited offenders is therefore much bigger than the life-coursepersistent group. Moffitt claims that these developmental trajectories can be
found among males as well as among females. Females less often suffer
from neuropsychological problems and therefore relatively few females
follow the life-course-persistent trajectory (Moffitt, Caspi, Rutter, & Silva,
2001).
A limitation of the add-and-stir approach is that researchers assume that
girl delinquency can be explained in the same way as boy delinquency,
without being sensitive to the possible differences between boys and girls in
12
terms of pathways to delinquency and context of crimes (Steffensmeier &
Allan, 1996). The focus of theories in this approach is on gender neutrality
and as such they preclude sex-specific mechanisms.
Female-only-focused approach
The second approach in female delinquency research, the female-onlyfocused approach, states that existing mainstream theories are not adequate
to explain delinquent behaviour of females. Girls are thought to have
different risk factors and hence need different theories that can explain their
delinquent behaviour. Therefore, not mainstream, but female focused
theories are used (Chesney-Lind, 1989; Hoyt & Scherer, 1998; Miller &
Mullins, 2009). Such theories claim that patriarchal power relations are a
reason for the sex difference in delinquency. In addition, researchers taking
this approach, believe that it is the reality that males and females have
different experiences in life because of their gender (Daigle, Cullen, &
Wright, 2007). For instance, females are at larger risk of being victimized
then males (Holsinger, 2000) or of being depressed (Calhoun, 2001). Also,
they are supervised more closely than boys (Leaper, Anderson, & Sanders,
1998). For this reason, the pathways to crime of females are different from
those of males. Also, it is thought that females who commit crimes are
usually forced to do so through victimization by males, because of their
economic disadvantage or because of other manifestations of gender
inequality (Chesney-Lind, 1989; Gaarder & Belknap, 2002).
A shortcoming of the female-only-focused approach is that researchers
are only interested in unique female factors and assume that there are no
shared factors and therefore do not examine these. These researchers even
believe that studies should focus only on delinquent females and ignore
delinquent males. In their point of view females are different from males
anyway and therefore there is no need to compare findings with those for
males. However, in order to know whether females are indeed different from
males, males should be used as a comparison group. As a result, the femaleonly-focused view of female delinquency is, just like the add-and-stir
approach, selective: it precludes gender neutral theorizing.
Since the add-and-stir approach is not very informative regarding femalespecific pathways to crime and the female-only-focused approach neglects
resemblances between males and females, girl delinquency should be
examined through the integration of these two approaches. Especially while
delinquent behaviour of girls is still less understood, it is essential in girl
delinquency research to study as many possibly related factors, either shared
or sex-specific and not to preclude any explanations.
13
Cumulative risk approach
In line with this broad focus, a third approach is needed to supplement the
discussed approaches. This is the so-called cumulative risk approach which
asserts that it is a not a single risk factor that affects delinquency, but instead
a whole range of factors that influence whether adolescents become
delinquent or not. The more risk factors adolescents experience, the more
likely they are to be delinquent, for boys as well as girls (Johansson &
Kempf-Leonard, 2009; Loeber, Slot, & Stouthamer-Loeber, 2008; Van der
Laan & Van der Schans, 2010).
Because the cumulative risk approach is very flexible, it is an adequate
approach to study girl delinquency. As a consequence of its flexibility, it is
possible within this approach to integrate the add-and-stir and the femaleonly-focused approach. The cumulative risk approach thus allows to include
shared as well as sex-specific factors in the study on girl delinquency.
Furthermore, this approach captures the intrinsic covariation of risk factors
(Luthar, 1993).
The present dissertation investigates which factors are related to girl
delinquency by considering those factors that are shared with boys, as well
as factors that may be sex-specific. Factors are taken into account in the
individual, family, peer and school domain. This cumulative risk approach
provides the stepping stones for the development of theories on specific
domains and factors to analyze through what mechanisms these domains and
factors influence delinquent behaviour of girls.
Previous research
This dissertation is not the first one to study girl delinquency and sex
differences in risk factors of delinquency. A number of previous studies have
examined risk factors for delinquent behaviour of females. However, these
studies did not integrate the discussed approaches. Studies either examined
whether known risk factors for males could be applied to girls – and mainly
found similarities between boys and girls – or they focused only on sexspecific factors – and found differences between boys and girls. In any case,
no comprehensive picture is provided yet about delinquent girls.
The studies of Rowe, Vazsonyi, and Flannery (1995), Moffitt et al.
(2001), and Junger-Tas, Rbeaud, and Cruyff (2004) are examples of studies
that examined whether known risk factors for males were related to girls’
delinquency. In the study of Rowe et al. (1995), 499 10-to-16-year old boys
and girls were analyzed, but no sex differences were found in risk factors.
Delinquent behaviour of both boys and girls was related to sexual
experience, parental affection, parental encouragement, school achievement,
anger, emotionality, impulsivity, deceitfulness, rebelliousness, and
delinquent peers. Also Moffitt et al. (2001) concluded that the same risk
factors predicted delinquency in girls as in boys (i.e. a young mother, a low
14
intelligent mother, deviant mother-child interaction, harsh or inconsistent
discipline, living with a single parent, low socio-economic status, low school
involvement, low intelligence, difficult temperament, hyperactivity, being
rejected by peers, delinquent peers). Their conclusions were based on
longitudinal analyses of delinquent behaviour from the age of 13 to 18 of
adolescents in the Dunedin study. Some sex differences were found (that is,
a higher vulnerability of boys for family risk factors), but according to the
researchers these were not very consistent if different measurement
strategies or statistical procedures were used. Junger-Tas et al. (2004), who
carried out research among juveniles from eleven different countries, also
concluded that no large sex differences could be found in the relations
between delinquency and important risk factors such as family bonding,
parental knowledge, the bond with teachers, school commitment, truancy,
and number of friends. Hence, these studies found that boys and girls have
similar risk factors for delinquency. However, the studies might have missed
the unique risk factors for boys and girls, because they only focused on risk
factors previously identified for boy delinquency.
Not many studies are carried out yet that focused on sex-specific factors
of delinquency. The ones that did examine sex-specific factors, mostly
compared characteristics of delinquent girls with those of delinquent boys.
They found some differences in characteristics between boys and girls.
Delinquent girls appeared to have lower levels of self-esteem, lower quality
parent-child relationships, lower family climates, to be more often raised in a
household with a single parent or with domestic violence, and to be more
often physically or sexually abused than delinquent boys (Belknap &
Holsinger, 2006; Emeka & Sorensen, 2009; Gavazzi, Yarcheck, & ChesneyLind, 2006; Gover, 2004; Johansson & Kempf-Leonard, 2009). However,
non-delinquent comparison groups were mostly not included and the
samples were probably biased and it is therefore unclear whether these
characteristics are indeed risk factors for delinquency. A more detailed
discussion of these and other limitations of previous research will be
discussed in a later paragraph. In any case, previous research is limited and
therefore it is difficult to draw conclusions yet about risk factors of girl
delinquency.
The explanation of the sex difference in the prevalence of delinquency
Next to knowledge about girl delinquency, it is important to gain knowledge
about the nature of the sex difference in the prevalence of delinquency.
Being able to explain why girls are less often delinquent than boys can
facilitate our understanding of girl delinquency. Although the add-and-stir,
female-only-focused and the cumulative risk approach have their limitations,
they contain elements that can be used to generate hypotheses to explain the
sex difference in the prevalence of delinquency: the differential risk
exposure hypothesis, the differential risk vulnerability hypothesis and the
15
differential threshold hypothesis. All these hypotheses will be examined in
this dissertation.
The differential risk exposure hypothesis
The differential risk exposure hypothesis assumes that the origin of
delinquency is similar for both sexes, but that boys are exposed to higher
levels of these similar risk factors than girls and are therefore more likely to
become delinquent (Keenan, Loeber, & Green, 1999; Lahey et al., 2006;
Moffitt et al., 2001; Murray, 2002). For the sake of simplicity, the hypothesis
will be referred to as the exposure hypothesis in the rest of this dissertation.
Differential exposure regards the level of risk and not the presence of risk.
For instance, adolescents are at little risk to become delinquent if they have
only one delinquent friend, but if all of their friends are delinquent, their risk
for delinquency is much higher. The exposure hypothesis states that girls are
exposed to lower level of risks and are therefore less likely to become
delinquent. To illustrate this, let us go back to the developmental taxonomy
of Moffitt (1993). Girls were found to have less (i.e. were less exposed to)
neurological deficits than boys and, according to Moffitt, therefore there are
fewer girls who follow the life-course-persistent trajectory. Along the same
line of reasoning, girls are thought to be less often delinquent because they
have higher levels of so-called promotive factors than boys: Hirschi (1969)
claims that girls have (i.e. are exposed to) stronger attachment to their family
and stronger bonds to school than boys, and are therefore more likely to be
withheld from delinquency. Girls thus are exposed to more protection in
terms of bonds and are therefore less often delinquent. In this dissertation, it
is examined whether sex differences in exposure can explain the sex
difference in the prevalence of delinquency. The hypothesis is confirmed if
the association between sex and delinquency is not significant anymore
when differences in exposure are taken into account.
Previous research indeed found that differential exposure could partly
explain the sex difference in delinquency. Moffitt et al. (2001), for instance,
found that because boys were exposed to higher levels of harsh parental
discipline, neuro-cognitive risk, hyperactivity and peer influences than girls,
part of the sex difference in the prevalence of delinquency was explained.
Messer, Goodman, Rowe, Meltzer, and Maughan (2006) found as well that
higher exposure of boys to harsh punishment, neuro-cognitive risk and peer
problems accounted for the sex difference in delinquency, as well as lower
reading performance of boys and higher prosocial behaviour of girls. These
factors were all assessed through parents and teachers though, and therefore
may contain gender bias. Differences in parental supervision (i.e. girls were
controlled more by their parents than boys) could also partly explain the
difference between boys and girls in delinquency (Junger-Tas et al., 2004).
Sex differences in school problems and in the quality of relationships with
parents did not contribute to the explanation though. Differential exposure
16
regarding the quality of the parent-child relationship (i.e. girls higher quality
than boys) did seem to partly explain sex differences in serious delinquency
(Fagan, Van Horn, Hawkins, & Arthur, 2007). Findings regarding the
parent-child relationship appeared to be more consistent when mothers and
fathers were considered separately: boys were found to receive more
paternal affection than girls, whereas girls were found to receive more
maternal affection (Fagan et al., 2007; Rowe et al., 1995).
The differential risk vulnerability hypothesis
Another possible explanation for the sex difference in delinquency, the
differential risk vulnerability hypothesis, states that girls and boys are
differently vulnerable for risk and promotive factors (Lahey et al., 2006;
Moffitt et al., 2001; Murray, 2002). This indicates that different associations
exists between risk and promotive factors on the one hand and delinquency
on the other. Vulnerability can be different in two ways. First, factors can be
associated with delinquency for one sex, but not for the other. This would
imply that boys and girls have completely different risk and promotive
factors for delinquent behaviour. In addition, differential vulnerability can
signify differences in terms of the strength of the association between factors
and delinquent behaviour. This is when factors are related to delinquency of
both boys and girls, but have a stronger relationship to delinquency for one
of the sexes. In other words, even if delinquency of boys and girls is affected
by the same risk factors, these may still have a different effect on them. For
instance, it might be that delinquency of both boys and girls is affected by
having delinquent peers, but that boys are more sensitive to peer influences
than girls. This would result in differential vulnerability towards delinquent
peers. It should be noted that the two types of vulnerability are actually
similar to each other, because a non-significant relationship between a factor
and delinquency is similar to an extremely weak relationship. The
consequences are different, however: in case of the first type of
vulnerability, the factor is uniquely related to delinquency of one sex,
whereas in case of the second type of vulnerability, the factor is a risk for
both boys and girls, even though it is more relevant for one of them.
According to the vulnerability hypothesis, girls are less likely to be
delinquent than boys, because boys are more vulnerable for risk factors than
girls, or have more risk factors they are vulnerable for (Lahey et al., 2006;
Moffitt et al., 2001; Murray, 2002). The same goes for promotive factors:
girls might be more susceptible for promotive factors than boys or might be
sensitive to more promotive factors, and therefore are less inclined to
commit offences. In this dissertation an attempt is made to test whether
differential vulnerability could explain the sex difference in the prevalence
of delinquency. If differences in vulnerability are taken into account and sex
cannot predict delinquency anymore, the vulnerability hypothesis is
confirmed.
17
In the previous paragraph, prior findings about differences in risk
factors between boys and girls were already discussed. However, none of the
studies that examined whether boys and girls had different risk factors,
tested the vulnerability hypothesis: whether differences in the effect of risk
and promotive factors can explain the sex difference in delinquency.
In this dissertation, the exposure hypothesis and the vulnerability
hypothesis are also tested concurrently. That is, it is very well possible that
girls are less exposed to risk factors than boys and at the same time are less
vulnerable for the risk effects, or not vulnerable at all. In that case one of the
hypotheses may not by itself be able to account for the difference in
delinquency between boys and girls, but in unison the two hypotheses have
more explanatory power. The two explanations should therefore be
examined concurrently. As far as we are aware, no previous study has done
so.
The differential threshold hypothesis
The third hypothesis to be tested that aims to explain the sex difference in
delinquency is the differential threshold hypothesis. Please note that for the
sake of simplicity the hypothesis is referred to as the threshold hypothesis.
This hypothesis claims that the more risk factors someone has, the more
likely he or she is to become delinquent (Eme, 1992). This implies that
people will only commit offences if they have a certain number of risk
factors. This particular number has been identified as the ‘threshold’ for
delinquency. A threshold is a cut-off value from which point onwards the
probability to be delinquent is larger than the probability not to be
delinquent. The threshold hypothesis asserts that this threshold is higher for
girls, which explains why delinquency is less common among girls.
Not many studies have examined whether girls indeed have a higher
threshold than boys. Existing studies suggested that more risks are needed
for girls to become delinquent, but they have used select samples and did not
operationalize the threshold properly. Alemagno, Shaffer-King, and Hammel
(2006), for instance, considered the number of risk factors of 250 boys and
girls who were incarcerated and found that detained girls had more risk
factors than detained boys. The same was found by Van der Laan and Van
der Schans (2010) regarding family risks. They showed that arrested girls
appeared to be exposed to more risk factors in the family domain than
arrested boys. They also found that arrested boys had a higher number of
risky lifestyle factors compared with arrested girls though.
Prior studies did not investigate the threshold as such, but instead
compared the number of risk factors of delinquents with those of nondelinquents. In order to adequately examine (sex difference in) the threshold,
it is essential to assess the threshold itself. Put differently, the exact ‘tipping
point’ should be identified for boys and girls.
18
In addition, it is possible that the threshold differs by age since the
delinquency prevalence of boys and girls varies by age and criminal careers
also develop differently for boys and girls (Junger-Tas, Haen-Marshall, &
Ribeaud, 2003; Wong et al., in press). As Moffitt (1993) suggested, during
adolescence boys and girls often commit crimes, as an expression of the
maturity gap they experience. Therefore, during puberty, delinquent
behavior is almost normative behavior, also for girls. Besides, since girls’
delinquency peaks earlier than that of boys, i.e. at age 15 versus at age 16
(Junger-Tas et al., 2003; Slotboom, Wong, Swier, & Van der Broek, 2011),
the age at which delinquent behavior is relatively most common, differs
between the sexes.
Limitations of previous research about girl delinquency
Previous studies about girl delinquency and about explanations for the sex
difference in delinquency had a number of methodological limitations. First,
many girl delinquency studies – especially those that considered sex-specific
factors and those that considered the threshold hypothesis– were carried out
in juvenile justice institutions or with a sample of arrested or adjudicated
girls. Since it is very likely that girls who are arrested or incarcerated are a
select group, results from these studies probably suffer from selection bias.
Investigating the characteristics of these girls therefore only resulted in
knowledge about girls that are involved in the juvenile justice system (i.e.
risk factors for delinquency, getting arrested, and incarcerated), but not in
knowledge about delinquency per se. Whether arrested, adjudicated or
incarcerated girls are indeed a special group of girls is not verified yet, but
indications in this direction can be made if we consider the juvenile
‘prosecution gap’, i.e. the difference between the number of offenses
committed and the number of offenses prosecuted (Rouwette, Van Hooff,
Vennix, & Jongebreur, 2007). A longitudinal study from Farrington et al.
(2003) showed with data from the US that according to self-reports, 85.9%
of the juveniles had committed a crime at least once as a 11 to 17 year old,
whereas only 34.0% of them was referred to court for at least one offence.
For females the prosecution gap appears to be even bigger than for males
(for more details, see Chapter 2) and, compared to their male counterparts,
the arrested and prosecuted females seem to be a group even more select.
For instance, Sealock and Simpson (1998) found that juvenile females were
relatively less often arrested than juvenile males for crimes. Also, if they
were arrested, they were relatively less often prosecuted compared to males,
even if it was taken into account that females commit less serious crimes
than males (Jeffries, Fletcher, & Newbold, 2003; Spohn & Beichner, 2000;
Williams, 1999). Also in the Netherlands, the prosecution gap of girls
appeared to be bigger than of boys. Van der Geest and Bijleveld (2008)
examined characteristics of institutionalized boys and found that according
to self-report 78% of the boys had committed one or more offences before
19
admission to the institution, whereas 62% of the boys was convicted for an
offence before admission. Girls from the same institution were examined by
Megens and Day (2007) and showed that while 75 per cent had committed at
least one offence before admission, only 41% was prosecuted for an offence.
These results all suggest that the juvenile justice system is more ‘lenient’
towards girl offenders than towards boy offenders. However, there are also
indications that girls receive a harsher punishment for minor offences;
especially in the US where juveniles are adjudicated for status offences, such
as truancy and running away (Hoyt & Scherer, 1998; Rosenbaum &
Chesney-Lind, 1994). In any case, research that involves only arrested or
incarcerated girls is not suitable to investigate all delinquent behaviour of
girls. A better way of examining girl delinquency might be by using selfreported delinquent behaviour of a population-based sample.
Second, in many studies there was a lack of a non-delinquent
comparison group. Without such a comparison group it is difficult to
conclude whether certain factors are indeed risk factors, since it might be
that all girls (and not only delinquent girls) score high on these factors. This
can only be tested with a comparison group.
Third, previous research on girl delinquency and on explaining the sex
difference in delinquency was often based on small samples. With small
samples, it is difficult – if not impossible – to detect effects. However, since
delinquency among girls has low prevalence, studies have difficulties
finding a group of delinquent females that is large enough to research girl
delinquency with enough power. In order to be certain to have a sufficient
number of delinquent girls, researcher therefore often focused on samples of
juvenile justice females.
Fourth, many studies do not include both boys and girls. Many studies
that only contained girls in their samples, compared their results with
previous boys’ studies. However, with such a comparison it is not clear
whether differences found between girls and boys are actual differences or
are caused by method differences between the studies. To accurately
examine sex differences in risk and promotive factors, information about
boys and girls should be assessed in the same manner, preferably within the
same study.
Fifth, studies that include girls in their samples, usually have no or only
a short follow-up period. Without longitudinal data, however, no conclusions
can be drawn about causal relations of factors (Singer & Willett, 2003).
Moreover, it is known that the prevalence of offending varies greatly by age
and different domains and factors can be important for delinquency at
different age periods. Much research has been carried out with large selfreport studies that have followed males for a (long) period of time, with
studies such as the Pittsburgh Youth Study, the Cambridge Study, and the
Montreal Longitudinal Study. Although there also exist some large studies
that include females (i.e. the Pittsburgh Girls Study, the Rochester Youth
20
Study, The Denver Youth Study, RADAR, the Dunedin Study, Trails), not
much research has been done yet with longitudinal self-reported female
delinquency data.
Sixth, and this applies actually to delinquency research in general, much
attention has been paid to risk factors, but hardly any attention has been paid
to promotive factors or to the cumulative effect of risk and promotive
factors. However, although risk factors can increase the likelihood of
becoming delinquent, promotive factors can decrease this likelihood. Risk
and promotive factors can thus be of similar importance and should therefore
receive the same amount of attention in research. Most delinquency studies
define their factors in such a way that they all represent risk factors. For
instance, not supervision, but lack of supervision is examined, assuming a
risk effect of low supervision. However, whether delinquency is related to
high supervision, low supervision or to both (i.e. whether supervision has a
risk effect, a promotive effect, or both) is generally never tested.
Aims and overview of the thesis
The central objectives of this dissertation are twofold. The first objective is
to establish an overview of factors related to delinquent behaviour of
adolescent girls and to identify which of those factors are unique for girls
and which of them are shared with boys. To start, a summary is given of risk
factors for girl delinquency found in previous European research and a
comparison is made with risk factors for boy delinquency. Subsequently, it is
empirically tested whether adolescent boys and girls differ in factors related
to delinquent behaviour. The second objective is to explain the sex
difference in delinquency during adolescence. Why is it that girls are less
likely to commit offences than boys? This will be examined by testing three
hypotheses that might explain the sex difference in delinquency: the
exposure hypothesis, the vulnerability hypothesis and the threshold
hypothesis.
To demonstrate the extent of the sex difference in delinquency, chapter
2 will start with an outline of how the prevalence of delinquency differs
between boys and girls in the Netherlands. Because it is rather difficult to
measure the prevalence of offending behaviour, both official records as well
as self-reported data are used. In presenting the prevalence, a distinction is
made by type of offence and by age. In addition, the trends of self-reported
delinquency between 1996 and 2005, and for official records until 2007, will
be displayed.
Chapter 3 reviews European research on risk factors for delinquency in
adolescent and young adult females. All research carried out in European
countries between 1984 and 2009 will be reviewed that examined delinquent
behaviour of females aged 12 to 25 and that included a comparison group of
non-delinquents. For inclusion in the review, it is essential that males and
females were examined separately. To analyze whether the factors are
21
unique risks for girls or not, risk factors are compared with those for male
delinquency.
In chapter 4, it is shown which factors in the parents and peer domain
are related to boy and girl delinquency. Furthermore, two hypotheses to
explain the sex difference in the prevalence of delinquency are tested: the
exposure hypothesis – sex differences in exposure to factors explain the
difference in delinquency – and the vulnerability hypothesis – sex
differences in associations between factors and delinquency explain the
difference in delinquency. It is not only examined whether differential
exposure and sensitivity can indeed explain why girls are less often
delinquent than boys, but also whether these operate simultaneously. Data of
the first measurement of the RADAR study is used, an ongoing Dutch
prospective cohort study, to examine the hypotheses. Adolescents were then
12 or 13 years of age.
Chapter 5 investigates the third explanation for the sex difference in
delinquency: the threshold hypothesis, that is whether more risk factors are
present for girls to increase the probability to be delinquent than for boys.
Previous research suffered from many limitations and therefore never
adequately tested the threshold hypothesis. In this chapter the threshold is
assessed in a similar way, by comparing the risk level of delinquent girls
with those of delinquent boys, but improves previous research by taking into
account risk and promotive factors and a non-delinquent comparison group.
Again, data of the RADAR study will be used, only now data of the second
and third waves when adolescents were aged 13 to 15.
Chapter 6 goes a little further into the threshold research by identifying
the exact location of the threshold (i.e. the tipping point) for boys and girls
and by including sex-specific factors in determining the risk level.
Furthermore, age-graded effects are studied in the sense that it is examined
whether there are sex differences in this threshold for delinquency at two
different age periods: in late childhood (ages 10 to 12) and in adolescence
(ages 13 to 16). The chapter will again provide an overview of unique and
shared risk and promotive factors for girls at the developmental stages. Now,
data of two studies from the US are used: the Pittsburgh Youth Study and the
Pittsburgh Girls Study.
Finally, chapter 7 presents an overview of the results found in the
previous chapters. In addition, it shows what the unique and shared risk
factors are for girl delinquency in adolescence and to what extent the three
hypotheses could explain the sex difference in delinquency. The chapter also
reflects on the methodological strengths and on limitations of the dissertation
and implications for practice and theory. This final chapter will also provide
an agenda for future research.
22
What this dissertation adds to the literature on girl delinquency
Not much research has been carried out on girl delinquency and on the
explanations for the sex difference in the prevalence of delinquency.
Existing research was, as explained, limited. This dissertation will improve
on many of these limitations. To make sure that results are not biased
because of selected samples, randomly selected population samples are used
in the empirical chapters and a non-delinquent comparison group is included.
Also, boys and girls are compared without any methodological disparities.
Furthermore, longitudinal data are used whenever possible. Moreover, this
dissertation is the first in examining the exposure hypothesis and the
vulnerability hypothesis concurrently and by testing directly whether they
can explain the sex difference in the prevalence of delinquency. In addition,
this dissertation is the first in examining the threshold properly by including
risk and promotive factors, a non-delinquent comparison group and, lastly,
by identifying the exact location of the threshold.
23
Chapter 2
Prevalence of juvenile delinquency
An extended version of this chapter was published as ‘Wong, T. M. L., Blom, M., &
Van der Laan, A. (in press). De inhaalslag van vrouwen? Omvang, aard en trends in
criminaliteit onder meisjes en vrouwen. In A. Slotboom, M. Hoeve, M. Ezinga & P.
van der Helm (Eds.), Criminele meisjes en vrouwen: Achtergronden en aanpak. Den
Haag: Boom Juridische Uitgevers.’
25
26
This chapter describes the prevalence and time trends of delinquency of 12to-18-year old boys and girls in the Netherlands. Because every method of
determining the prevalence of delinquency has limitations it is difficult to
establish how many juveniles offend. Therefore, it is wise to use more than
one method to determine this prevalence. In this chapter, two methods are
considered: official records and self-reports.
Data sources and limitations
Official records
Official records pertain to juveniles who were arrested because they were
suspects of a crime. Absolute numbers as proportions related to subgroups
(e.g., the proportion of 12-to-18-year old males versus females who were
arrested) are presented here. These relative numbers can be used to compare
groups (such as boys and girls) with each other, as well as to make
comparisons with prevalence numbers from self-reports of delinquency. To
make an optimal comparison possible with self-report, the 2005 records will
be used in this study.
A serious limitation of official records is that not all offences that are
actually committed become known to the police. For instance, those offences
that involve victims who do not report to the police, or offences without
victims are not visible in official records. Moreover, for many crimes, the
police is not able to identify the offender. The average proportion of crimes
that are solved by the police in the Netherlands is 22 per cent (Eggen, 2009).
Besides, police records reflect the priorities and effort of the police and
therefore offenders who commit those offences that are within these
priorities are probably more likely to be identified and arrested than others
(Blom, Van der Laan, & Huijbregts, 2005). Because of all this, there is a
large so-called dark number of crimes. The dark number refers to those
crimes that are committed but not known to the officials.
Self-reports
One way to get a grip on this dark number is to use so-called self-reports of
criminal behaviour. In self-reports, people report themselves which crimes
they have committed in a particular time period. The self-reported
information used in this chapter is obtained from the WODC Youth
Delinquency Survey (WODC YDS). This is a cross-sectional survey,
conducted every three years amongst a representative sample of Dutch
juveniles aged 10 to 17. We used the information gathered in the first three
months of 2005. The sample was randomly selected from the General
Municipal Personal Records Database (GBA). Of all the juveniles
approached, 68 per cent were willing to participate. This resulted in 1,460
juveniles interviewed. The analyses are weighted by a number of
background characteristics (i.e. gender, age, ethnic background, education
level, degree of urbanization, and region) to make the results representative
27
for all juveniles in the Netherlands (for more details, see Van der Laan &
Blom, 2006). For a total of 33 offences, the juveniles reported whether they
had committed them or not in the previous twelve months. Note that the
reference period is the twelve months prior to the first quarter of 2005 and
therefore not completely similar to that of the official records. Two rule
violations (i.e. dodging and using fireworks) are not taken into account in
this chapter, since these are not as serious as the other offences and their
prevalence is relatively high so that their influence on the prevalence
numbers would be (too) large. Furthermore, this way the numbers are more
comparable to those in the official records. Only the prevalence of the
remaining 31 offences will thus be reported on.
The self-report method also has limitations. Firstly, the validity of selfreport is dependent on the honesty and may seriously be jeopardized by
social desirability of the respondents (Farrington, 1973). Also, respondents
might be afraid that their anonymity is not ensured which may lead to
underreporting (Hessing & Elffers, 1995). However, some respondents
might also overreport because they like to brag or because they report about
offences outside the reference period (Nijboer, 1995; Sudman, Bradburn, &
Schwarz, 1996). However, several researchers have shown that self-report is
a sufficiently reliable and valid method for obtaining prevalence data on
delinquency (Thornberry & Krohn, 2000; Van der Heijden, Sijtsma & ’t
Hart, 1995). Another limitation of self-reports is that these are mostly
focused on common crimes and less on serious crimes. Therefore, many
serious offences be under assessed with this method. In the WODC YDS this
limitation is considered by taking more serious crimes into account.
However, the focus is still on more common (less serious) offences.
Prevalence of different offences
In Table 2.1 the prevalence of the 31 self-reported offences is presented. Of
all girls, 37.6% reported to have committed at least one offence in the
previous year. The prevalence of public order, property and minor violent
offences is about equal (around 20%). The offences mostly committed by
12-to-18-year old girls were hitting without injury (13.4%) and spoiling
walls and public transport (12.7%). More boys than girls reported to have
committed an offence, that is 52.3 per cent of the boys indicated to have
committed one offence or more. The prevalence of offending in boys was
therefore 1.5 times the prevalence in girls. Also in case of separate offences,
the prevalence in boys was higher than in girls for almost all offences. For
some offences the prevalence in boys and girls was similar; regarding
changing price tags and shoplifting up to 10 euros, the prevalence in girls
was higher. Just like girls, most boys reported that they have hit somebody
without injuries (19.8%). Besides, many boys reported to have stolen at
school or work (17.9%) in the reporting period.
28
Table 2.1
Prevalence of self-reported delinquency in 2005 in the previous twelve
months amongst 12-to-18-year olds in the Netherlands, by sex (%)
Girls
Boys
(n = 552)
(n = 572)
19.5
29.4*
Vandalism and public order offences
Damaged or destroyed a vehicle
1.1
2.7*
Damaged or destroyed a house
0.9
2.0
Damaged public transport
1.4
2.3
Damaged something else
3.8
8.0*
Soiled walls or public transport
12.7
12.2
Miscall regarding color
8.1
15.8*
Miscall gay
2.7
7.0*
20.6
29.9*
Property offences
Changed price tags
8.0
4.5*
Shoplifting < 10 euro
7.5
4.8*
Shoplifting > 10 euro
0.9
0.3
Stole anything from school or work
9.0
17.9*
Bike or scooter theft
1.7
3.8*
Stole anything from outside car
0.2
1.9*
Stole anything from car
0.0
0.0
Purse snatching
0.4
0.6
Burglary
0.6
1.5
Buying stolen goods
4.0
9.2*
Selling stolen goods
1.8
3.5
20.8
29.6*
Minor violent offences
Threatened somebody
11.0
13.8
Hitted somebody, no injury
13.4
19.8*
Hitted somebody, with injury
6.3
12.9*
0.9
1.6
Serious violent offences
Threatened to steal
0.0
0.2
Robbery
0.0
0.0
Hurt somebody with weapon
0.9
0.5
Forceful sex
0.0
0.9*
2.6
7.8*
Possession of weapon
Weapon when going out
2.6
7.8*
2.2
2.9
Drug offences
Dealing softdrugs
2.1
2.5
Dealing partydrugs
0.4
0.5
Dealing harddrugs
0.0
0.0
6.5
8.9
Internet offences
Sending viruses
1.0
3.0*
Threat through internet
5.9
7.3
Total offences
One of more offences committed
37.6
52.3*
* p < .05
In 2005 more than 5,300 girls aged 12 to 18 were arrested as crime
suspects. This is about 1 per cent of all 12-to18-year old girls in the
Netherlands (See Table 2.2). Relatively speaking, many girls were arrested
because they were suspected of committing a property offence (0.4% of all
Dutch girls). Most often this was because they had committed aggravated
29
thefts (such as burglary or theft with a group of people), although girls were
also arrested for simple thefts. Violent and public order offences were less
common among girls (respectively 0.3% and 0.2%). Regarding the separate
offences, girls appeared to be most often arrested as a suspect of a violent
offence against a person.
The number of boys who were suspects of a crime was much higher
than that of girls, independent of the type of offence. In general, four times
as many boys are as girls were arrested. Differences were most obvious
regarding public order offences: the number of boys who were suspects of a
public order offences was six times bigger than the number of females
arrested for this. With regard to property and violent offences, four times as
many boys as girls were arrested. Within the group of boys, the number of
boys suspected of committing a public order, a property or a violent offence
is almost equal.
Table 2.2
Arrested suspects aged 12 to 18 in 2005, by sex (absolute and %)
Girls
Boys
abs.
%
abs.
%
1,329
0.2
8,440
1.4
Vandalism and public order offences
Public order
921
0.2
5,088
0.8
The authorities
136
0.0
657
0.1
Other vandalism and public order offences
327
0.1
3,613
0.6
2,402
0.4
9,140
1.5
Property offences
Simple theft
884
0.2
2,892
0.5
Aggravated theft
1,333
0.2
5,949
1.0
Other property offences
423
0.1
2,070
0.3
1,751
0.3
8,122
1.3
Violent offences
Sex offence
13
0.0
622
0.1
Threat
350
0.1
1,873
0.3
Violence against persons
1,433
0.2
5,191
0.8
Robbery
132
0.0
1,570
0.3
470
0.1
3,477
0.6
Other offences
Traffic offence
107
0.0
960
0.2
Drug offence
117
0.0
621
0.1
Weapons
47
0.0
858
0.1
Other
204
0.0
1,163
0.2
5,341
0.9
23,130
3.8
Total (at least one offence)
Note. The sum of the sub categories can be larger than the total of the sub subcategories,
since one person can be arrested for more than one offences.
Many differences appear between data obtained from the official
records and those obtained from the self-reports. Although almost 40 per
cent of the girls reported in 2005 to have committed at least one offence in
the previous year, not even 1 percent was arrested in that year. Even though
the self-report survey asked after a number of less serious offences, this still
illustrates the extent of the dark number, as the police only knows about a
very small part of girl delinquency. It has to be noted, however, that self30
reports are especially adequate to measure frequently committed crimes that
are less serious. The numbers therefore may not fully reflect criminal
offences, but the less serious offences less frequently lead to an arrest.
Regarding the type of offences, not many differences were found
between self-reported public order, property and violent offences. Twenty
per cent of the girls reported to have committed at least one such offence.
Arrest rates for these crimes did differ, however. It appeared that more girls
were arrested for a property offences than for a violent offence. Arrest rates
for violent offences were higher though than those for public order offences.
This comparison shows the difference between the two methods. From
the police data, it appears as if girls less often commit violent or public order
offences. The self-reports show, however, that girls are also quite likely to
commit violent and public order offences. An explanation for this might be
that girls are most often arrested for a property offence, since this is seen as a
typical ‘girl offence’. In case of a violent of public order offence, the police
might be less inclined to expect girls to be involved.
Age
The prevalence of delinquency for girls appears to be highest among 15year-olds, according to self-reports (see Figure 2.1). Almost half of the girls
aged 15 in the sample reported to have committed an offence in the previous
twelve months. To see whether there were age differences regarding various
types of offences, a distinction will be made between public order, property
and minor violent offences. It was not possible to look into all separate
offences, since the prevalence then became too low to test for age
differences. For every offence type the prevalence was also highest among
15-year-olds. Girls who were 12 years of age most frequently reported minor
violent offences; those 15 years old minor violent and property offences and
those aged 17 most frequently reported property offences. In comparison
with boys, at every age, fewer girls reported to have committed an offence,
also for separate kinds of offences. At age 12, differences between boys and
girls were small. For boys, the prevalence was highest among 16-year-olds,
which indicates boys have a later prevalence peak age than girls.
Figure 2.2 shows the percentage of female and male juveniles arrested
in 2005 by age. The curve for girls is in line with the general age crime curve
(Farrington, 1986; Junger-Tas, 1992): the prevalence of delinquency
increases with age, peaks in late adolescence and then decreases. Apart from
level differences, the curves are similar for every type of crime. For every
type of crime, the prevalence was highest among 15-year-olds, except for
property crimes where the number of suspects was highest among 18-yearolds. There are a number of noteworthy differences between females and
males. The most apparent sex difference is that, across all ages, the number
of arrested males is much higher than that of females. Furthermore, while the
peak age for females is at age 15, most boys were 19 years old when they
31
Figure 2.1
Prevalence of self-reported delinquency, by age
Boys
80
70
Prevalence (%)
60
50
40
30
20
10
0
12
13
14
15
16
17
15
16
17
Age
Girls
80
70
Prevalence (%)
60
50
40
30
20
10
0
12
13
14
Age
Public order
Property
Minor violence
Total
got arrested. In addition, there at not many differences between the different
types of offences for which boys were arrested, whereas there are differences
for girls: at every age females were most often arrested for property
offences. Thus, the higher arrest rates for girls regarding property crimes
mentioned before are present regardless of age.
32
Number of suspects per 100 peers
Figure 2.2
Arrest rates per 100 peers, by age
Males
8
6
4
2
0
Number of suspects per 100 peers
Age
Females
2
1,5
1
0,5
0
Age
Public order
Property
Violence
Total
Girl offending peaks at the age of 15. This finding is rather certain,
since it was found in the self-reported data as well as in the official records.
After the age of 15, the prevalence of delinquency among women declines.
For the different types of offences it was also found, in both data sources,
that the peak age was 15 years. The one exception was regarding property
crimes in official records which peaks at the age of 18 whereas in selfreported data it peaks at the age of 15. The biggest difference between selfreported data and official records was found for 12-year-old girls: although
33
there are hardly any girls arrested at the age of 12, almost 30 percent of these
girls reported to have committed an offence.
A noteworthy difference between boys and girls is that delinquency
peaks at an earlier age for girls than for boys. This appeared from selfreported data as well as from official records. Although the peak age for girls
is 15 years, the peak age for boys is 19 years according to official records
and 16 years according to self-reported data. Since no self-reported data
existed beyond the age of 17 it cannot be determined what the development
of the prevalence is after this age and therefore there might be another peak
age than 16 years. The difference in official records is rather large. An
explanation might be that because girls have an earlier puberty than boys and
reach adulthood at an earlier age, they might be younger when they start
offending, but also when they desist from delinquent behaviour (Haynie,
2003).
Time trends
After considering sex differences in the prevalence of different (types of)
offences and at different ages, trends in delinquency prevalence over time
will be described. For the trends in self-reported delinquency the prior
measurements of the WODC YDS were used from 1996, 1998 and 2001 (for
more details, see Van der Laan & Blom, 2006). In Table 2.3 the delinquency
prevalence can be seen for these years and for 2005 for girls as well as boys.
The prevalence numbers only consider native Dutch juveniles, since the
measures are not well comparable for juveniles with a different ethnicity, as
for immigrant juveniles non-random samples had been drawn. In addition,
the table only contains those offences that are measured in a consistent way
across years. Only small differences appear between the prevalence numbers
over time. The prevalence of delinquency among girls varies to some extent
between 1996 and 2005. The prevalence of damaging and soiling objects
seems to increase from 1998. The prevalence of changing price tags or of
having injured somebody increased between 2001 and 2005 (please note that
this is only an increase between two measurement points) among girls.
Regarding shoplifting and buying or selling stolen goods the prevalence
seems stable: no clear increase or decrease can be identified. Hence, for
some offences there was a small increase in prevalence, for others it
34
35
15,4
2005f
(N=424)
11,7
13,2
5,3
5,8
9,7
Note. From Jeugddelinquentie: Risico’s en bescherming (p.100), A. M. van der Laan en M. Blom, 2005 , Meppel:
Boom Juridische Uitgevers.
a
One of the offences in the sub category damaged objects. From 1998 to 2001 this category contained six different
question, in 2005 four. All questions consider damaging a car; public transport; windows; a lamppost; phone booth; bus
shelter. b One of the offences in the sub category shoplifting. In 1996 and 1998 juveniles reported on shoplifting of
something with a value lower or higher than 10 guilders; in 2001 on shoplifting in general; in 2005 on shoplifting of
something with a value lower or higher than 10 euros. c In 1998 to 2001 one question about buying or selling something
that has been stolen; in 2005 these were two separate questions. d One of the offences in the subcategory to injure. In
1996 and 1998 it was not assessed. In 2001 and in 2005 two questions were asked: hitted somebody with injury; hurt
somebody with weapon. e In the previous year. f In the previous 12 months.
Table 2.3
Prevalence of self-reported delinquency in the previous 12 months amongst 12-to-18-year olds, by year (%)
Girls
Boys
1996e
1998e
2001f
2005f
1996e
1998e
2001f
(N=413) (N=506) (N=410) (N=411)
(N=477) (N=525) (N=432)
Damaged objecta
4,4
3,5
3,9
4,6
18,4
14,9
12,4
Soiled objects
9,4
7,8
10,8
10,9
9,7
13,1
9,2
Shopliftingb
7,5
5,1
9,4
6,8
11,5
10,1
7,6
Changed price tags
5,9
8,2
4,3
Buying or selling stolen
4,2
3,3
3,8
3,6
12,0
8,2
6,9
goodsc
Injured somebodyd
7,1
8,0
14,6
remained fairly stable. In general, the prevalence of delinquency thus
increased slightly over time for girls.
Delinquency prevalence is always smaller for girls than for boys,
regardless of the assessed year. However, the prevalence of shoplifting
decreased for boys, even below the level of that of girls. The prevalence of
damaging objects also decreased over time. The prevalence of changing
price tags or of injuring somebody increased between 2001 and 2005 among
boys. It thus depends to a great extent on the type of crime how the
prevalence has developed over time for boys. However, although for girls
the prevalence of delinquency appeared to either increase or remain stable
between 1996 and 2005, for boys the prevalence of some offences decreased
over time.
In Figure 2.3 it is shown how the percentage of girls who is arrested
developed between 1996 and 2007. For every year, the increase or decrease
in terms of percentages with reference to 1996 was presented.1
The number of arrested girls seems to be stable between 1996 and 2000
and to strongly increase from 2001. In comparison to 1996, the number of
arrested girls in 2007 has almost doubled. The strongest increase can be
found among the number of girls arrested for a public order offence: two and
a half times more arrests in 2007 compared to 1996. Although the prevalence
of public order offences is relatively low, the proportionate increase
throughout the years was strong and therefore the difference with other types
of crime has decreased. The number of girls that is arrested on suspicion of a
violent offence more than doubled and of a property offence increased with
more than half between 1995 and 2007.
The number of arrested boys also increased after 2001, but less so than
the increase among girls. This indicates that the sex difference in
delinquency according to official records has decreased in this period. This
is especially so for property offences. The number of boys arrested on
suspicion of a property crime hardly changed in the period between 1996
and 2007 (i.e. -9%), whereas the number of girls increased with 66%.
Many differences can be found between the time trends in delinquency
according to self-reports and according to official records. Although the selfreported delinquency prevalence only slightly varied over time for girls, the
official records show a strong increase of arrested girls between 2001 and
2005, that continues in 2006 and 2007 (years that were not available for self-
1
Official record in a particular year are gathered twice: once several months after
the years has passed (the preliminary numbers) and once a year later (the final
numbers). In case of preliminary numbers, there can be errors in registration and
therefore numbers might be incomplete or over complete. For the period 1996 to
2000 only preliminary numbers were available and therefore a dotted line is used to
display the delinquency prevalence in this period.
36
36
Figure 2.3
Time trends in the number of arrested 12-to-18-year olds per 100 residents from 1996 to
2007, by year (increase or decrease in comparison with 1996)
Boys
260
240
220
200
180
160
140
120
100
80
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Year
260
240
220
200
180
160
140
120
100
80
Girls
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Public order
Year
Property
Violence
Total
reported delinquency). Regarding specific types of crime, some similarities
were found between the data sources. Although conclusions should be drawn
with care given the fact that the self-reported data regard a (small) sample,
the results imply that the prevalence of public order offences among girls
somewhat increased between 1996 and 2005. This trend is in line with
official records, although the latter show an increase much larger than the
self-reported increase. It might be that the policy towards female suspects
has been changed, especially from 2001. From 2003 the Dutch government
37
strongly focused on reducing juvenile nuisance (TK, 2002/2003a). It might
very well be that this partly generated the rise in the number of arrested girls
on suspicion of a public order offence. The increased attention for domestic
violence, a policy program that started in 2002 (Ministry of Justice, 2002),
may also be related to the rise in arrest rates of girls. Although this program
started from the idea that females are often victims of domestic violence, the
program might also have led to more female suspects of crime, since
research suggests that violent crimes of girls are often committed within
their own houses (Putallaz & Bierman, 2004). More attention for domestic
violence could therefore have resulted in more female arrests.
Since official records pertain to a select group of delinquents and this
dissertation aims to gain knowledge about delinquent girls in general, in the
following self-reported data will be used, collected in randomly selected
population samples. Data has shown that the prevalence of delinquency
among girls is lower than among boys. This dissertation attempts to explain
this difference between boys and girls.
38
39
Chapter 3
Risk factors for delinquency in
adolescent and young adult females:
A European review
Wong, T. M. L., Slotboom, A., & Bijleveld, C. C. J. H. (2010). Risk factors for
delinquency in adolescent and young adult females: A European review. European
Journal of Criminology, 7, 266-284.
41
This study reviews 30 European studies on risk factors for delinquency of
adolescent and young adult females. Risk factors are compared with those
for (adolescent) male delinquency. Risk factors unique for females are
identified as well as risk factors shared by males and females. Sex
differences in risk factors are mainly found for individual and family factors.
Females and males differ less in school and peer risk factors. As studies
zoom in more at specific domains of risk factors or specific types of
delinquent behaviour, more differences emerge between males and females.
Limitations of studies examining female delinquency are discussed and
future research areas proposed.
42
Introduction
Juvenile delinquency and its causes have been studied extensively. Many
factors that put adolescents at risk of becoming delinquent have been
identified, such as parental monitoring, physical abuse and neglect, having
delinquent friends and low school achievement (Loeber, 1990; Thornberry &
Krohn, 2003). However, such risk factors have been identified
predominantly for males. Much less is therefore known about adolescent
female offending and its aetiology. One reason often put forward for this is
that female offending is generally less serious – so that there would be a less
pressing need to study it – and its prevalence lower – which makes it more
difficult to study.
There are different explanations for the sex difference in delinquency.
First, it is often assumed that female delinquency can be explained in the
same way as male delinquency (Moffitt et al., 2001), but that females are
less exposed to risk factors than males. A second explanation states that male
and female delinquency have different causes. A third explanation is that
females have a higher threshold for becoming delinquent than males. This
means that, while they may be affected by the same risk factors as males,
females have to experience more risk factors in order to start offending.
When the threshold is passed by females, however, it is believed that the
manifestation of delinquency is more severe than in males, the so-called
‘threshold effect’ (Loeber & Keenan, 1994).
A number of meta-analyses and reviews of risk factors of (adolescent
and young adult) female offending (Hoyt & Scherer, 1998; Hubbard & Pratt,
2002; Simourd & Andrews, 1994; Tankersley & Rycraft, 2007) have been
carried out. These generated evidence for individual risk factors such as an
antisocial personality, low intelligence, substance abuse, physical or sexual
abuse, mental health problems, pubertal timing and low levels of self-worth.
In the family domain, important risk factors are the quality of parent-child
relationships, conflicts within the family, parental control, family violence
and having caregivers with a history of substance abuse or delinquency.
Mixed results were found for single parenthood and school achievement.
Delinquent friends, gang membership and the quality of peer relationships
were found to affect female delinquency.
However almost all these reviews and meta-analyses were based on
American studies. It is unlikely that risk factors found for US females can be
generalized to European females. In the US, serious violence rates and gun
use are higher than in Europe (Haen-Marshall, 1996; Junger-Tas et al.,
2003), as is gang involvement (Hill et al., 1999; Klein et al., 2001).
Furthermore, drug law violations figure more prominently in US official
statistics (Tonry, 1999) and contrary to most European countries, in the US
juveniles can be prosecuted for status offences like running away (ChesneyLind, 1999). Moreover, the income distribution in the US is much more
skewed than in Europe, and the US does not have social welfare systems as
43
most European countries do, leading to greater income inequality and to
more poverty (Alesina & Glaeser, 2004).
This paper will provide a systematic overview of studies on adolescent
and young adult female delinquency in Europe. The overview will list risk
factors shared by males and females, as well as risk factors unique for
genders. In that sense, the study aims to provide information that may shed
light on the first and second explanation for the sex difference in offending.
Whereas previous meta-analyses and reviews generally combined specific
risk factors into larger clusters, we will examine risk factors separately.
Method
Sample of studies
Computer searches of the databases Google Scholar, Psych Info and Science
Direct were conducted to locate relevant studies, using the following key
words: female delinquency, female offenders, female juvenile offenders, girl
delinquency, risk factors, juvenile delinquency, gender and sex differences.
We also traced the reference lists of studies. To make sure that every
published and unpublished study considering female delinquency was
included, we also contacted (developmental) criminologists from many
European countries (the Netherlands, Belgium, The United Kingdom,
France, Italy, Germany, Denmark, Finland, Sweden, Poland, Spain,
Switzerland, Slovenia and Hungary) and asked them to refer us to any
published or unpublished studies in their countries with at least an abstract in
English, French or German (or Dutch).
Various criteria were used for the selection of studies. First of all,
studies had to assess delinquency, i.e. behaviour penalized by criminal law,
and not simply ‘risky’ behaviour. Furthermore, if males were also addressed
in the studies, it was essential that females were analyzed separately.
Additionally, it was necessary that studies also considered non-delinquent
females. Another criterion was age: only studies that sampled adolescents
and young adults were selected (aged 12-25). Some studies pertained to
adult females, but these were included if delinquency in adolescence or
young adulthood was explicitly addressed. A last criterion was the year of
publication: only studies published during the last 25 years (from 1984 to
2009) were selected. The search and selection resulted in 30 European
studies (see Table 3.1).
44
45
10.
9.
8.
Hoeve, 2008
The Netherlands
Honkatukia, 1998
Finland
Jolliffe & Farrington, 2007
England
N = 330 (54.8% females)
Age 13-22
N = 1254 (47.5% females)
Age 15-19
N = 720 (47.8% females)
Age 13-17
Table 3.1
European studies on female delinquency and its risk factors
Authors
Sample
1.
Bijleveld & Wijkman, 2009
n = 1540 (48.1% females)
The Netherlands
Average age 20
2.
Campbell, 1987
n = 64 (100% females)
England
Age 15-16
3.
Farrington & Painter, 2004
n = 1013 (51.2% females)
England
Age 40 (risk factors from age 8)
4.
Graham & Bowling, 1996
n = 1648 (about 50%)
England
Age 14-25
5.
Hartjen & Priyadarsini, 2003
n = 387 (62.3% females)
France
Age 13-18
6.
Hauber et al., 1987
n = 308 (57.1% females)
The Netherlands
Age 12-18
7.
Hodgins et al., 2001
n = 13852 (48.7% females)
Sweden
Age 30
Odds ratio
χ²
Pearson correlations
Hierarchical regression
T-tests and χ²
Official records
Self report (23 items)
Self-report (50 items)
Self-report
Official records
4 groups: non-offenders (*),
offenders, violent offenders,
persistent early starters (*)
Self-report
Zero-order correlations
Self report (43 items)
Unclear, most likely
correlations
t-tests
ANCOVA
Strategy of analysis
Odds ratio
Assessment delinquency
Official records
46
Landsheer & Van Dijkum, 2005
The Netherlands
Mansel, 2003
Germany
Martens, 1997
Sweden
McAra, 2005
England
Oberwittler 2003
Germany
18.
17.
16.
15.
13.
14.
n = 270 (58.1% females)
Age 12-14 at T1
n = 2086 (48.2% females)
Age 12-16
n = 502 (43.8% females)
Average age 14
n = 4299 (50% females)
Age 13
n = 5819 (52.1% females)
Age 12-14
Study 2
n = 4500 (about 50% females)
Age 14-15
n = 5056 (100% females)
Age 16
n = 13852 (48.7% females)
Age 30
The Netherlands
Kemppainen et al., 2002
Finland
Kratzer & Hodgins, 1999
Sweden
Study 1
n = 10996 (48.4% females)
Age 14-21
Sample
Junger-Tas et al., 2004
3 different cultural clusters:
Anglo-American (England and
Wales, Northern Ireland, USA);
north-west Europe (Finland,
Germany, Belgium, the
Netherlands, Switzerland) (*);
southern Europe (Spain,
Portugal Italy)(*)
12.
11.
Authors
Self-report
Self-report (violence)
Official records.
4 groups: early starter (*),
adolescent limited (*), adult-onset,
discontinuous, no crime (*)
Official records
Study 2
Self-report
Study 1
χ² and hierarchical regression
Correlation
Longitudinal
Correlation
Logistic regression – Odds
ratio
T-tests
Study 2
Logistic regression – Odds ratio
Study 1
Gamma’s
47
Authors
Piko et al., 2005
Hungary
Rantakallia et al., 1995
Finland
Riley & Shaw, 1985
England and Wales
Ritakallio et al., 2005
Finland
Scholte et al., 2007
The Netherlands
Slotboom et al., 2005
The Netherlands
Smith & McAra 2004
England
Svensson 2004
Sweden
Torstensson, 1990
Sweden
Sample
n = 1240 (47% females)
Age 11-20
20.
n = 11764 (48.9% females)
Age 25
21.
n = 751 (49.7% females)
Age 14-15
22.
n = 50569 (49.9% females)
Age 14-16
23.
n = 214 (49.5% females)
Age 13-17
24.
n = 1978 (45% females)
Average age 13
25.
n = 4144 (about 50% females)
Age 15
26.
N = 979 (48.4% females)
Average age 14
27.
n = 6825 (100% females)
Age 30
4 groups: non-delinquents (*),
delinquents from Child Welfare
Committees (*), females with
police-recorded crimes or serious
drug use and females with police
records ánd records of hard drug
use
28.
Van der Rakt et al., 2005
n = 932 (43.1% females)
The Netherlands
Age 12-14 at T1
29.
Van de Schoot & Wong, in press n = 873 (66% females)
The Netherlands
Age 18-24
(*) In our review we only focused on these groups of the study
(**) Test results are calculated ourselves with the available information
19.
χ²-test
Longitudinal
Correlation
Correlation
Correlation
t-tests (**)
Longitudinal
Bayesian approach: various
theories tested
Official records
Official records
Self-reported (8 items)
Variables
We distinguish four clusters of factors related to delinquency: individual,
family, school and peer factors. The individual factors are: birth
complications, negative life events, personality, moral development,
attitudes towards delinquency and other problem behaviour. Family factors
are: parenting, family relationships and family situation. School factors are:
school achievement, attachment and commitment. The last group of factors
is peer factors: delinquent friends, number of friends and quality of
relationships with peers.
Analytic strategy
To examine the risk factors of female delinquency adequately, the studies
should be as comparable as possible. Therefore, we attempt to focus on
results from bivariate analyses and avoid results from multiple regression
analyses as this type of analysis cannot reveal the independent effects of risk
factors, but instead the effect of a risk factor given the effects of other risk
factors in the model. Results from a regression analysis are therefore only
considered if no other results were reported, and we will indicate whenever
this is the case. If hierarchical regression analyses were used, we will always
discuss the results of the final model with all risk factors.
If factors were inconsistently related to delinquency, in the sense that
one bivariate analysis found an effect for a risk factor but another study did
not, we report this inconsistency but leave the factor out of our summary
tabulations. If however, a large number of studies reported a finding, and
only one study deviated from this, we do report the factor as a risk factor,
noting the diverging study. If a bivariate result for a risk factor in one study
was not replicated in another study in a multivariate regression analysis, we
do report the risk factor as the disappearance of the effect may be due to
multicollinearity. These choices imply that if only one study included a
certain risk factor and found an effect, we do include this risk factor in our
final tabulations.
Although this review focuses on overall delinquency, we have included
results from studies that distinguish between different types of delinquency
as well. We do not include them in our summary tabulations but do report
them in the text.
To discuss differences between risk factors for male and female
delinquency, results for males and females within the same study are
compared with each other, based on the p-value. Factors are considered
similarly relevant to male and female delinquency if the factors are
significantly related, in the same direction, to both male and female
delinquency in the same study. Strictly speaking, we can only decide so after
48
the regression weights have been statistically tested against each other.
Almost none of the studies have carried out such tests, though.2
Results
The studies and a number of their properties are listed in Table 1. On
average, the studies had considerably large sample sizes, ranging from 64 to
over 50,000. Studies used either official records or self-reported
delinquency. Most studies used population-based samples and most of these
contained both female and male respondents; only two had a female-only
sample. Some studies carried out multivariate (regression) analyses, most
employed bivariate analyses. The studies came from 13 mainly West
European countries.
In discussing results, we first consider, for every cluster of factors, the
risk factors found for female delinquency, followed by the differences found
with factors for male delinquency.
Individual factors
Birth complication. No study found an association between birth
complication and female delinquency (Hodgins et al., 2001; Kemppainen et
al., 2002; Rantakallia et al., 1995). Male delinquency, however, was related
to birth complications (Hodgins et al., 2001; Rantakallia et al., 1995).
Negative life events. The number of negative life events predicted female
delinquency (Junger-Tas et al., 2004 study 2). More specifically, female
delinquency was related to crime victimization (Junger-Tas et al., 2004 study
2; McAra, 2005; Smith & McAra, 2004) and to having been harassed3 by an
adult (McAra, 2005; Smith & McAra, 2004). Male delinquency is similarly
related to certain negative life events (Junger-Tas et al., 2004; McAra, 2005;
Smith & McAra, 2004); although the number of negative life events was not
related to male delinquency (Junger-Tas et al., 2004).
Personality. Low levels of self-control (Slotboom et al., 2005) and low
intelligence (Kratzer & Hodgins, 1999; Rantakallia et al., 1995) were found
to be related to female delinquency. Additionally, female delinquency was
associated with high levels of aggression and disobedience (Hauber et al.,
1987). Varying results were found for the relation between female
delinquency and the level of impulsiveness (Hauber et al., 1987; McAra,
2
A complete overview of all results of all studies is available from the authors.
Adult harassment was assessed by questioning: number of times in past year an
adult stared at you so that you felt uncomfortable; followed you on foot; followed
you by car; tried to get you to go somewhere with them; indecently exposed
themselves to you (McAra, 2005).
3
49
2005; Smith & McAra, 2004) and risk-taking behaviour (Hauber et al., 1987;
Smith & McAra, 2004).
Findings for self-esteem were inconsistent, both high (Piko et al., 2005)
and low self-esteem (McAra, 2005; Smith & McAra, 2004) were related to
female delinquency. This is possibly due to the finding by Van de Schoot
and Wong (submitted) that both low and high self-esteem were related to
female and male delinquency.
Regarding male delinquency, a relation was found with low general
self-esteem (McAra, 2005; Smith & McAra, 2004) but not with high general
self-esteem (Piko et al., 2005). Otherwise, delinquent males have the same
personality characteristics as delinquent females (Hauber et al., 1987;
Kratzer & Hodgins, 1999; McAra, 2005; Rantakallia et al., 1995; Slotboom
et al., 2005; Smith & McAra, 2004), except for disobedience on which
delinquent females scored higher than their male counterparts (Hauber et al.,
1987). In addition, the relation between delinquency and low self-control
was stronger for males than for females (Slotboom et al., 2005).
Moral development. Svensson (2004) studied feelings of shame about
delinquency towards friends, parents and teachers (i.e. the level of shame
adolescents would feel towards others if they were caught committing an
offences) and found that it was negatively related to female delinquency as
well as male delinquency.
Jolliffe and Farrington (2007) also distinguished between different types
of delinquent behaviour (i.e. violence, fighting, theft from person, serious
theft, shoplifting, vandalism and drugs use). They found that total empathy
was related to violence and theft from a person, but not to overall
delinquency. Low affective empathy was associated with theft, violence and
shoplifting; low cognitive empathy was associated with theft only. For
males, contrary to females, total and affective empathy were related to
overall delinquency. These factors were also related to violence, fighting and
vandalism. In addition, affective empathy was related to serious theft and
shoplifting amongst males. Cognitive empathy was related to fighting,
serious theft and vandalism. These results show that disaggregating by types
of delinquency and by types of a certain factor can generate complex results
about sex differences. Therefore, no conclusions on unique or shared factors
can be drawn about empathy and delinquency.
Attitudes towards delinquency. A positive attitude towards delinquency
increased the risk of becoming delinquent for females as well as for males
(Hartjen & Priyadarsini, 2003; McAra, 2005; Oberwittler, 2003; Smith &
McAra, 2004; Riley & Shaw, 1985).
Other problem behaviour. Delinquent females were more often depressed
(Piko et al., 2005; Ritakallio et al., 2005), suicidal and substance abusing
50
(Piko et al., 2005) than non-delinquent females. For males, results regarding
the relations between depression and delinquency are inconsistent (Piko et
al., 2005; Ritakallio et al., 2005). Psychological wellbeing was related to
male delinquency, but not to female delinquency (Junger-Tas et al., 2004
study 1). Male delinquency was related to substance abuse, but not to being
suicidal.
Figure 3.1 presents an overview of the unique and shared individual risk
factors consistently related to male and female delinquency. It appears from
this figure that quite a number of factors are shared by males and females.
Victimization, a positive attitude towards delinquency as well as low selfesteem appear to be factors found repeatedly over studies. At the same time,
some differences do emerge as well: females have a marked set of risk
factors regarding internalizing problems. Also, the high number of life
events for females is notable. Finally, only delinquent boys are marked by
birth complications, possibly signalling increased risk for neurological
damage.
Figure 3.1
Unique and shared individual risk factors consistently related to female or male delinquency
Unique factors for females
High number of life
events (1r)
Disobedience (1r)
High self-esteem (2+r)
Depression (2r)
Suicidal behaviour (1r)
Shared risk factors
Victimization (3+r)
Being harassed by an adult
(2)
Low self-control (1)
Aggression (1r)
Low IQ (2+r)
Low self-esteem (3)
Being unashamed for
parents, friends and
teachers (1)
Positive attitude towards
delinquency (5)
Substance abuse (1r)
Unique factors for males
Birth complications (2)
Psychological wellbeing
(1r)
Note. The numbers between brackets indicate the number of studies that took the risk factor
into account. Superscript r: results from all studies come from a regression analysis;
superscript +r: results from some of the studies come from a regression analysis; superscript
l: results from all studies come from a longitudinal analysis. superscript +l: results from some
of the studies come from a longitudinal analysis
Family factors
Parenting. Inadequate parenting was found to be associated with female
delinquency (Hodgins et al., 2001; Mansel, 2003). Hoeve (2008) took four
different parenting styles into account and found that maternal and paternal
51
parenting styles affected female delinquency. Females with a permissive,
poorly responsive mother, father, or both (i.e. moderately attached to
adolescent, little response to adolescent’s needs, low demands on
conformity) had the highest rates of delinquency.
Looking more specifically into parenting practices, we can see that
parental monitoring – parenting behaviour that considers the child’s
whereabouts, activities and friends – was negatively related to female
delinquency (Campbell, 1987; Farrington & Painter, 2004; Graham &
Bowling, 1995; McAra, 2005; Piko et al., 2005; Riley & Shaw, 1985; Smith
& McAra, 2004). Decreasing parental control through adolescence was
positively associated (Honkatukia, 1998). Low parental involvement with
school also puts females at greater risk to become delinquent (McAra, 2005;
Smith & McAra, 2004). Inconsistencies were found for parental knowledge
about activities (Junger-Tas et al., 2004 study 1; Martens, 1997), knowledge
about friends was unrelated to female delinquency (Junger-Tas et al., 2004
study 1).
Parental monitoring is closely linked to what adolescents and young
adults actually tell their parents, i.e. child disclosure (Stattin & Kerr, 2000).
Low child disclosure appeared to be related to delinquency of females (Riley
& Shaw, 1985; Svensson, 2004). Females who are seldom at home, a factor
reflecting parental monitoring, were found to be more likely to be delinquent
(Martens, 1997).
Monitoring also involves rules at home, which was negatively related to
female delinquency (Slotboom et al., 2005) and with harsh discipline
positively related (Campbell, 1987; Farrington & Painter, 2004; Mansel,
2003). Inconsistent results were found for rules about the time until when
adolescents are allowed to go out. Physical abuse by parents was positively
related to female delinquency (Piko et al., 2005) as was low parental trust
towards children (McAra, 2005; Riley & Shaw, 1985).
Some parenting factors were differently associated with male
delinquency. First, males’ delinquency was affected by paternal, but not by
maternal parenting styles. Neglectful, rather than permissive, parenting (i.e.
unattached to adolescents, punish often, lowest demands on conformity)
appeared to be positively related to male delinquency. Males having a father,
or both parents, with a neglectful parenting style were most likely to be
delinquent. Second, high parental knowledge about friends (Junger-Tas et
al., 2004) was related to male delinquency, in contrast to female
delinquency. Third, opposed to their female counterparts, males who were
seldom at home were not more likely to become delinquent (Martens, 1997),
nor were males who had been physically abused by their parents (Piko et al
2005). Inconsistencies were found for the relations between male
delinquency and low child disclosure and low parental trust.
Scholte et al. (2007) examined different types of delinquency and
differential parental treatment among siblings. They found no predictive
52
value of differential parental treatment in their longitudinal analysis on theft,
vandalism and violence. However, males’ vandalism and violent behaviour
could be predicted by a less favourable treatment (compared to their
siblings) by mothers.
Quality of relationship with family members. Low parental warmth was
found to be related to delinquency (Mansel, 2003; Slotboom et al., 2005).
Landsheer and Van Dijkum (2005) longitudinally analyzed parental support,
and found that female delinquency is related to low support of mothers, but
not to that of fathers. The extent to which adolescents and young adults go
out or spend time with family was not consistently related to female
delinquency (Hartjen & Priyadarsini, 2003; Honkatukia 1998; Junger-Tas et
al., 2004 study 1; McAra, 2005; Riley & Shaw, 1985).
Low attachment and commitment to parents is related to female
delinquency if studied bi-directionally (Graham & Bowling, 1995; Slotboom
et al., 2005; Svensson 2004). However, in studies in which the predictive
value of attachment or commitment on delinquency was examined in a
multiple regression analysis, effects always disappeared (Hauber et al., 1987;
Martens, 1997; Van der Rakt et al., 2005). The extent of conflicts that
females have with their parents was related to delinquency (Mansel, 2003;
McAra, 2005; Smith & McAra, 2004; Riley & Shaw, 1985).
When relationships with mothers and fathers were examined separately,
it appeared that both low quality mother-child and father-child relationships
were related to female delinquency (Campbell 1987; Junger-Tas et al., 2004
study 1; Riley & Shaw, 1985).
The associations between male delinquency and family relationship
factors were partly similar to those of females. However, low support from
mothers was not associated with male delinquency (Landsheer & Van
Dijkum, 2005). Furthermore, the quality of mother-child relationship was
inconsistently related to male delinquency.
Disaggregating by type of delinquency, the quality of adolescent sibling
relationships was related to female property offending. No relation was
found with vandalism and violence (Scholte et al., 2007). In addition, no
relation was found between sibling relationships and male offending.
Family situation. Having a convicted father (Bijleveld & Wijkman, 2009;
Farrington & Painter, 2004) or a delinquent sibling (Graham & Bowling,
1995; Farrington & Painter, 2004) was related to delinquent behaviour of
females. Conviction of the mother, however, was not associated with female
delinquency (Bijleveld & Wijkman, 2009; Farrington & Painter, 2004).
Single parenthood was inconsistently associated with female
delinquency as was socio-economic status. Living in a disadvantaged
neighbourhood was related to female delinquency (McAra, 2005). No
association was found between family size and female delinquency (Graham
53
& Bowling, 1995; Martens, 1997; Rantakallia et al., 1995; Riley & Shaw,
1985; Svensson 2004).
Males differ in their family situation risk factors. Having a convicted
father, a delinquent sibling and a convicted mother were related to male
delinquency (Bijleveld & Wijkman, 2009; Farrington & Painter, 2004).
Furthermore, single parenthood was related to male delinquency in almost
all of the studies (except McAra, 2005). Associations with family size and
low SES were comparable with those for females.
Figure 3.2 presents an overview of the unique and shared family risk factors
consistently related to male and female delinquency. We see that there are
numerous differences between males and females, that have been replicated
over several studies. The mother appears to play a unique role in female
delinquency, for both males and females neglect, discipline and conflicts
appear to be prominent.
Figure 3.2
Unique and shared family risk factors consistently related to female or male delinquency
Maternal parenting style
(1)
Low child disclosure (2)
Being seldom at home
(1r)
Physical abuse by parents
(1r)
Low parental trust (2)
Low maternal support
(1r,l)
Low quality mother-child
relationship (3)
Shared risk factors
Inadequate parenting (father
or both parents) (3)
Paternal parenting style (1)
Overall parental monitoring
(7 +r)
Decreasing parental
monitoring (1)
Low involvement parents in
school (3)
Small number of rules at
home (1)
Harsh discipline (3)
Low parental warmth (2)
Conflicts within family (4)
Low quality father-child
relationship (2)
Convicted father (2)
Delinquent sibling (1)
Living in a disadvantaged
neighbourhood (1)
High parental knowledge
about friends (1)
Convicted mother (2)
Single parenthood (9)
superscript +r: results from some of the studies come from a regression analysis; superscript
l: results from all studies come from a longitudinal analysis.
54
School factors
School achievement. School level and its relation with female delinquency
had inconsistent results. The same was found for school achievement –
within school levels. Concerning males, all studies found low school
achievement to be related to delinquent behaviour (Hartjen & Priyadarsini,
2003; Hauber et al., 1987; Junger-Tas et al., 2004 study 2; Rantakallia et al.,
1995; Riley & Shaw, 1985), except one (Piko et al., 2005).
Attachment and commitment to school. Low attachment to school – whether
adolescents like going to school – was related to female delinquency
(Graham & Bowling, 1995; Junger-Tas et al., 2004; Mansel, 2003; Slotboom
et al., 2005; Smith & McAra, 2004). Some studies failed to find an
association though (Hauber et al., 1987; Piko et al., 2005; Van der Rakt et
al., 2005), but these studies all used a regression analysis to test the
association. Low quality of the relationship with teachers (as an indicator of
school attachment) was related to female delinquency (Junger-Tas et al.,
2004; Mansel, 2003; McAra, 2005; Smith & McAra, 2004), as was low
commitment to school – whether adolescents are willing to put effort in
school (Hartjen & Priyadarsini, 2003; Mansel, 2003; McAra, 2005; Riley &
Shaw, 1985; Slotboom et al., 2005; Smith & McAra, 2004; Torstensson,
1990).
No large differences existed in the associations of school factors with
male delinquency compared to female delinquency, although it appeared that
the negative associations of attachment and commitment with delinquency
were somewhat stronger for males (Slotboom et al., 2005). Inconsistent
results were found for the relation between the quality of the relationship
with teachers and male delinquency.
Figure 3.3 presents an overview of the unique and shared family risk factors
consistently related to male and female delinquency. The figure shows that
low school commitment was a shared risk factor for males and females and a
poor relationship with teachers was uniquely related to female delinquency.
These findings have been replicated over many studies.
Peer factors
Delinquent friends. Female delinquency was related to having delinquent
friends (Graham & Bowling, 1995; Hartjen & Priyadarsini, 2003; Riley &
Shaw, 1985; Slotboom et al., 2005) as well as to the extent of these friends’
delinquency (Van der Rakt et al., 2005) and to having ‘negative’ friends,
characterized as having had contact with the police, low school achievement,
and being problematic according to their parents (Hauber et al., 1987).
Problematic youth group membership is also related to female delinquency
(Piko et al., 2005) and more specific, if disaggregated by type of
delinquency, to robbery, assault and severe property offences (Oberwittler,
55
Figure 3.3
Unique and shared school risk factors consistently related to female or male delinquency
Low quality relationship
with teachers (4+r)
Shared risk factors
Low school commitment (7)
Low school
achievement (6+r)
into account. Superscript +r: results from some of the studies come from a regression
analysis.
2003). According to Riley and Shaw (1985), having friends who are positive
towards delinquency was positively related to female delinquency. The
relationship disappeared in a multiple regression analysis (Hauber et al.,
1987).
Male delinquency was comparably and consistently related to
delinquent friends. Only Van der Rakt et al. (2005) found no association
between the extent of delinquency of friends and male delinquency. Results
from this study may deviate from those of other studies not only because the
extent of friends’ delinquency was assessed instead of having delinquent
friends, but also because this was the only study that assessed friends’
delinquency not indirectly through the adolescents but directly through
friends.
Number of friends. The relationship between the number of friends and
delinquency was not clear for females, probably because it depends on
whether these friends are delinquent or not. The number of opposite-sex
friends was related to delinquency (Smith & McAra, 2004). If a regression
analysis was used, the association disappeared (Van der Rakt et al., 2005).
Having a romantic partner was consistently found to be positively related to
delinquency (Smith & McAra, 2004; Riley & Shaw, 1985), as was the extent
to which adolescents and young adults hang out with their friends, or go out
(Hartjen & Priyadarsini, 2003; Honkatukia 1998; McAra, 2005; Smith &
McAra, 2004; Riley & Shaw, 1985).
Males’ delinquency was similarly inconsistently related to the number
of friends. The number of opposite-sex friends was consistently related to
male delinquency (Smith & McAra, 2004; Van der Rakt et al., 2005), but
findings on the direction of this relation were inconsistent.
Quality of relationship with peers. No relation was found between quality of
relationships with peers, or peer support and female delinquency (Junger-Tas
et al., 2004; Landsheer & Van Dijkum, 2005; Slotboom et al., 2005),
56
possibly because no distinction was made between antisocial and prosocial
peers. Friends disliked by parents were a risk factor for female delinquency
(Riley & Shaw, 1985).
For male delinquency, the same associations were found except
regarding the quality of peer relationships for which inconsistencies were
found.
Figure 3.4 presents an overview of the unique and shared peer risk factors
consistently related to female and male delinquency. Here, it appears as if
risk factors are similar as well as unique, replicated over a number of studies.
Figure 3.4
Unique and shared peer risk factors consistently related to female or male delinquency
Extent delinquency of
friends (1r, l)
Shared risk factors
Having delinquent friends
(3)
Having negative friends (1)
Problematic youth group
member (2r)
Having a romantic partner
(2)
Many activities with friends
(5)
Having friends disliked by
parents (1)
superscript l: results from all studies come from a longitudinal analysis.
Discussion
The purpose of this study was to generate an overview of risk factors related
to female delinquency as found in Europe and to identify factors that are
promising as distinctive explanations for offending in females as opposed to
males. This review contains most, if not all, European studies on risk factors
of female delinquency. Given that these number 30, female delinquency
appears definitely understudied.
Overall, many similarities in factors were found between males and
females, although clear differences emerged as well. Females seem to be
more affected by social context factors, such as negative life events and
physical abuse by parents. Also, internalizing problems were more
prominent.
The role of mothers in their daughters’ delinquency is noteworthy.
Maternal support, mother-child relationship and maternal parenting styles
57
were uniquely associated with female delinquency. However, for males,
having a convicted mother was uniquely related to their delinquency. Unique
female delinquency factors seem to be mainly in the relational domain. A
bad relationship with the mother seems to put adolescent and young adult
females at risk of becoming delinquent, but good relationships protect them.
It has to be kept in mind, though, that results have been derived from
crossectional, mostly bivariate results. It is therefore not apparent whether
quality of maternal relationship is a cause or effect of female delinquency (or
both). Also, only for girls it is the relationship with the school teacher that
plays a role.
School factors seem to have affect male delinquency more than female
delinquency. An explanation for this may be that when males fail at school
they may feel they have to compensate by showing delinquent behaviour:
through this compensation they may (re)gain their status (Warr, 2002). For
females, status might be regained by other types of behaviour. School
attachment and commitment might be mediating factors between
achievement and delinquency. If adolescents achieve badly at school they
will probably be less attached and committed to school, leading to more
compensating behaviour. Another explanation for the sex difference may be
that low intelligence is more strongly related to male than to female
delinquency (Moffitt et al., 2001). This might be reflected in stronger
relationships of school achievement, attachment and commitment with male
delinquency.
Even though we reported the delinquency of friends as a risk factor as it
has been established as such repeatedly, it is important to stress that the
relation between friends’ delinquency and delinquency may in fact be
spurious. Most studies assess friends’ delinquency indirectly by asking
respondents and not the friends themselves, and adolescents’ belief about
their friends’ delinquency may be perceived as more similar than it actually
is. The only study that asked the friends directly about their delinquent
behaviour (Van der Rakt et al., 2005), did not find any relation with male
delinquency, although it has to be taken into account that their interest was
different, as they examined the extent of friends’ delinquency, instead of
having delinquent friends.
As studies focused on more specific sub-domains (e.g. of empathy,
disaggregated domains of self-esteem, parental monitoring) or on different
types of offences (e.g. violent, property, public order), more sex differences
emerged. Thus it appears, that females do have a number of different risk
factors for delinquency than males, supporting the second explanation put
forward in the introduction. While our study found clear overlaps as well as
differences in risk factors for males and females, further systematic research
needs to uncover whether overlapping risk factors also have the same
predictive value for male and female delinquency.
58
Additionally, our findings might indicate that females have a higher
threshold to become delinquent; the number of negative life events in
particular affected female delinquency and not male delinquency. Related to
this, the extent of friends’ delinquency was found to be uniquely related to
female delinquency. These findings imply that females may be more resilient
to particular risk factors than males, and that only once females pass a
certain threshold they become delinquent. Previous research on the higher
threshold for females, in which mixed results were found, examined only the
number of risk factors (Baker et al., 1989; Eme 1992; Loeber & Keenan,
1994; Moffitt et al., 2001; Robins, 1966). The current study indicates that
perhaps the level or severity of risk factors might be of importance as well.
Thus, our findings offer some support for the third explanation for the sex
difference in delinquency as well.
Our findings differ from some of the findings from existing (mostly
American) reviews. More extensively studied in the US reviews were
variables such as sexual and physical abuse (Hoyt & Scherer, 1998; Hubbard
& Pratt, 2002; Tankersley & Rycraft, 2007), criminal family members,
substance abuse, mental health problems (Hoyt and Scherer; Tankersley &
Rycraft, 2007) and living in a disadvantaged neighbourhood (Tankersley &
Rycraft, 2007). These variables all appeared to be associated with female
delinquency in the US. In addition, Tankersley and Rycraft (2007) had
different conclusions about single parenthood compared to our findings, as
they found it to be consistently related to female delinquency. This shows
that associations between risk factors for (female) delinquency may indeed
not be generalizable from the US to Europe. An explanation may be that
single parenthood in the US is more strongly associated with poverty than in
European countries, because of the lack of a social welfare system.
Remarkably, neighbourhood effects are less often studied in Europe –
possibly because social disadvantage is less clustered within neighbourhoods
in Europe than in the US. Other factors were not at all studied in European
studies: family substance abuse, family violence and pubertal timing (Hoyt
& Scherer, 1998; Tankersley & Rycraft, 2007). Whether this is coincidental
or whether European studies structurally neglected some factors is not clear.
The current review contained many more variables, on the other hand,
to which previous reviews did not pay attention: birth complications,
empathy, self-control, shame, disclosure, differential parental treatment,
going out and spending time with family, quality of relationship with
teachers, number of (opposite-sex) friends, quality of relationship with
friends and whether parents disliked friends of the adolescent. Crosscontinental studies would be needed to gather evidence on the applicability
of risk factors in European and US contexts.
59
Limitations of the study and need for further research
Due to our choices, it occured that we labelled a factor a risk factor if it had
been identified as such in only one study. As such, risk factors identified by
us should be regarded as ‘not inconsistent’ rather than as ‘consistent’.
Secondly, we were not able to properly compare studies when they used
different methodologies or multivariate models. To make the studies as
comparable as possible, we discussed bivariate results if available. However,
relationships between risk factors and delinquency are obviously more
complicated in reality, and more realistic models, incorporating males as
well as females, are needed to arrive at a comprehensive test of differences.
The selected studies that examined delinquent females were also often
limited in several ways. Overall, there are not many longitudinal studies yet
on female delinquency. More specifically, only three studies were found in
Europe that investigated female delinquency longitudinally; all were from
the Netherlands (Landsheer & Van Dijkum, 2005; Scholte et al., 2007; Van
der Rakt et al., 2005). Longitudinal studies are essential in drawing
conclusions about causality. Therefore, it should be kept in mind that it is
uncertain whether the risk factors identified here are causes, effects or both.
What is more, most studies did not report on the effect sizes of their results.
We could therefore only discuss the significances. Given that many sample
sizes were large, if not huge at times, this also limits the relevance of our
findings.
In most studies, delinquent females were considered as one group.
There can, however, be many differences within groups of offenders, and
therefore it is important to make a distinction between those groups. Not
only in terms of types of offences, but also in terms of developmental
pathways of delinquent behaviour. Females showing delinquent behaviour
only in adolescence, presumably have different risk factors from females
who persist to be antisocial throughout life (Moffitt et al., 2001). These
differences are neglected when all types of offenders are analyzed together.
Given that delinquency is less prevalent in females, more studies concerning
high risk female samples seem warranted.
Furthermore, sex differences in associations of factors with delinquency
were almost never tested as such. Only one study actually performed such a
test (Slotboom et al., 2005). In this review, we had no other option but to
consider factors significant for one sex but not for the other, as being sex
different.
There was almost no overlap in factors that were taken into account in
the self-reported and criminal record studies. This can be explained by the
fact that studies with official records mostly obtain their variables from
existing documents, like obstetric or school files, whereas self-reported
studies are able to choose their own variables. So even if there were enough
studies to compare studies using self-reported delinquency with those using
60
criminal records as an outcome measure, there would be little to compare
since both types of studies cover different risk factors.
Clearly, in order to be able to shed light on the possibly different
mechanisms through which female and male adolescent and young adults
start, continue to and desist from offending, more comprehensive,
longitudinal studies are needed. Such studies would have to include males as
well as females, so that method differences can be disregarded as an
explanation for any sex differences. Ideally, as this review shows, they
would have to include many tentative risk factors not identified previously
found in male-only studies. Such studies would also have to enable explicitly
to model differential exposure to risk factors as an explanation for the
difference in male and female offending. This review shows that it is likely
that females do have a number of unique risk factors for delinquency and
that threshold effects exist. That does not preclude that differences in
exposure further help to explain how females have different pathways into
delinquency.
61
Chapter 4
The relationship between the sex
difference in delinquency and
differential exposure and vulnerability
in the parent and peer domain
Wong, T. M. L., Slotboom, A., Zahn, M., Bijleveld, C. C. J. H., Van Lier, P. A. C.,
Frijns, T., Meeus, W. H. J. & Koot, J. M. (submitted). The relationship between the
sex difference in delinquency and differential exposure and vulnerability in the
parent and peer domain.
63
This study is an attempt to gain insight in the aetiology of the sex difference
in delinquency by testing the risk exposure hypothesis, which states that sex
differences in exposure to risk factors are related to the sex difference in
delinquency, the risk vulnerability hypothesis, which states that sex
differences in associations between risk factors and delinquency are related
to the sex difference in delinquency, and the combination of the two. The
hypotheses are tested by using innovative methodological analyses on a
broad range of factors in the parent and peer domain. A sample of 480
Dutch juveniles from the RADAR study, a prospective cohort study, is
analyzed. Sex differences in exposure and vulnerability were found, but only
differences in exposure were related to the relationship between sex and
offending behaviour. A reason why girls are less often delinquent than boys
is that their mother is more involved in their lives and that they have less
delinquent friends.
64
Introduction
Delinquent behaviour peaks in adolescence. Although this has been widely
acknowledged for both boys and girls, the prevalence of girl delinquency is
always lower than that of boys (Junger-Tas, Marshall, & Ribeaud, 2003; Van
der Laan & Blom, 2006, Steffensmeier & Schwartz, 2009). An explanation
broadly accepted for the peak in adolescence is that adolescents experience a
gap between their biological and social maturity (Moffitt, 1993).
Delinquency is a way of displaying their discomfort about this gap and
demonstrating their autonomy towards their parents. Also, (status among)
peers become(s) more and more important in adolescence and empirical
research has shown that delinquent friends are imperative in explaining
delinquent behaviour (Warr, 2002). As we know, however, not all
adolescents engage in delinquent behaviour. Important factors that affect
whether they do or not are the previous mentioned peer factors, and factors
in the parent domain (Moffitt, Caspi, Rutter, & Silva, 2001; Wong,
Slotboom & Bijleveld, 2010). Parents can affect their children in two ways,
through parenting practices and through the relationship they have with their
children.
The link between parent factors and delinquency can be understood
through the social control theory of Hirschi (1969). This theory claims that
good relationships between parents and children can keep adolescents from
committing crimes. Since adolescents do not want to jeopardize the strong
bond they have with their parents and do not want to go against the wishes of
their parents they will not engage in offending behaviour. Adolescents with a
low quality parent-child relationship do not have this fear of losing the
strong attachment to their parents and are therefore more likely to be
delinquent.
Peers, on the other hand, can influence adolescents’ delinquency
because adolescents imitate each other’s behaviour. An important theory
about the influence of peers on delinquency is the differential association
theory of Sutherland (1947) which states that delinquency is behaviour
learned from others. People learn attitudes, techniques and motives within
intimate personal relationships. As friends are very important during
adolescence, adolescents especially learn behaviour from their friends.
Adolescents who have delinquent friends, are therefore likely to show the
same type of behaviour. However, it can also be a selection effect that
explains why adolescents often show the same behaviour. In other words,
adolescents might also choose their friends because they show similar
(delinquent) behaviour (Weerman, 2007).
Sex differences in parent and peer factors
From a theoretical perspective it can be expected that there are differences
between boys and girls regarding the influence of parents and peers. It has
been postulated, for instance, that girls are more family-oriented than boys
65
(Gecas & Seff, 1990; Gilligan, 1982). Furthermore, parents seem to socialize
their daughters differently than their sons and use gender specific parenting
practices (Hoeve, 2008; Lytton & Romney, 1991): boys are thought to
receive harsher punishment and less supervision than girls, whereas girls are
found to receive more parental warmth (Leaper, Anderson, & Sanders, 1998;
Lytton & Romney, 1991; Zahn-Waxler & Polanichka, 1994). This familyfocused nature of girls and differences in parenting can result in higher
levels of social control in girls compared to boys and therefore may lead to a
higher prevalence of delinquency among boys than among girls. It can also
result in sex differences in exposure and vulnerability regarding parent
factors.
Differential exposure implies that there are differences between boys
and girls in mean levels of risk and promotive factors. Risk factors are those
factors that “have proven or presumed effects that can directly increase the
likelihood of a maladaptive outcome” (Rolf & Johnson, 1990: 387).
Promotive factors are those factors that decrease the likelihood of becoming
delinquent. Promotive and risk factors are actually the opposite end of each
other within the same variable. That is, the same variable may constitute a
risk as well as a protective condition for adolescents (Stouthamer-Loeber,
Loeber, Wei, Farrington, & Wikström, 2002). In general, it is thought that
girls are exposed to fewer risk factors and to more promotive factors than
boys, the so-called risk exposure hypothesis (Moffitt et al., 2001; Murray,
2002). The stronger focus of girls on the family and the different parenting
strategies towards daughters may indeed result in a higher exposure to
promotive factors of girls than of boys. More specifically, girls might
experience more promotive parenting behaviour such as higher levels of
supervision (Leaper, Anderson, & Sanders, 1998), and high quality parentchild relationships (Hoeve et al., 2009). This may lead to higher levels of
social control in girls and therefore in a lower prevalence of delinquency.
Differential vulnerability implies that there are differences in the
association of delinquency with risk and promotive factors between boys and
girls. The risk vulnerability hypothesis states that girls are less vulnerable for
risk factors and more receptive for promotive factors than boys (Lahey et al.,
2006; Moffitt et al., 2001; Murray, 2002). It means that for girls risk factors
less often result in delinquent behaviour as for boys and that promotive
factors have a stronger deterrent effect on girls. This makes sense if we take
into account the family-oriented nature of girls. Compared to boys, girls may
be more afraid to endanger a good bond with their parents and they might be
more compliant towards them (i.e. girls have higher levels of social control).
Parents are therefore expected to have a stronger promotive effect on girls.
Exposure and vulnerability regarding peers may also be gender specific.
To start with exposure, adolescents tend to make friends of the same sex, so
girls have more female friends and boys have more male friends (Rose &
Rudolph, 2006). Since girls are less often delinquent than boys (Junger-Tas,
66
Ribeaud, & Cruyff, 2004; Steffensmeier & Schwartz, 2009; Van der Laan &
Blom, 2006), girls had less delinquent peers than boys. Girls therefore have
less role models from who they can learn delinquent behaviour. Regarding
vulnerability, the nature of boys’ friendships is different from that of girls
which may lead to different influences. While boys focus more on common
activities, competition and status within the group, girls find communication,
social support and taking care of each other more important in their
friendships (Giordano, Cernkovich & Pugh, 1986; Rose & Rudolph, 2006;
Heimer, De Coster & Ünal, 2006). Because girls are less involved with
competition and obtaining status, they seem to be less susceptible for group
pressure and therefore less vulnerable for the (bad) influence of delinquent
peers. Furthermore, boys likely internalize ideas that favour being aggressive
or delinquent, because male peer groups reinforce the idea of masculinity
through competitiveness and physical strengths. Female peer groups, on the
other hand, emphasize the importance of emotional bonds. Therefore, girls
are less likely to favour behaviour that includes harming others and violating
the law (Heimer et al., 2006).
In the present study, sex differences in exposure and vulnerability were
examined in the parent and peer domain. Not only is examined whether these
differences exist, but also whether these differences are related to the lower
prevalence of delinquency in girls.
Previous Research
A number of previous studies have examined whether boys and girls differ
in exposure to and vulnerability for risk and promotive factors regarding
parents and peers. Regarding exposure, sex differences seem to be especially
apparent in the peer domain. It was consistently found that boys had more
delinquent peers than girls (Fagan, Van Horn, Hawkins, & Arthur, 2007; Liu
& Kaplan, 1999; Mears, Ploeger, & Warr, 1998; Moffitt et al., 2001;
Murray, 2002; Rowe, Vazsonyi, & Flannery, 1995). Peer attachment and
peer popularity were higher for girls (Moffitt et al., 2001; Rowe et al., 1995).
In the family domain, sex differences in exposure particularly emerged when
mothers and fathers were considered separately. Moffitt et al. (2001), who
made no distinction between mothers and fathers, showed equal levels of
boys and girls regarding the quality of relationships with parents. However,
when examined more closely, sex differences in the parent-child relationship
appeared to emerge. Boys were found to receive more paternal affection than
girls, whereas girls were found to receive more maternal affection (Fagan et
al., 2007; Rowe et al., 1995). Furthermore, as indicated before, boys and
girls were exposed to different parenting practices (Hoeve, 2008; Leaper, et
al. 1998; Lytton & Romney, 1991; Zahn-Waxler & Polanichka, 1994), i.e
boys were exposed to harsher punishments than girls, less supervision and
less parental support.
67
Regarding vulnerability, previous studies also suggested that sex
differences existed. In the family domain, physical abuse by parents
appeared to have a stronger relationship to girl delinquency than to boy
delinquency (Piko, Fitzpatrick, & Wright, 2005). Boys’ delinquency was
found to be more strongly related to inconsistent parental discipline and
years with single parents (Moffitt et al., 2001; Wong et al., 2010). Family
conflict had a stronger relationship to antisocial behaviour of boys than to
that of girls, but it was related similarly to serious delinquency of both sexes
(Moffitt et al., 2001). The relation between delinquency and maternal factors
seem to be more important for girls than for boys (e.g. low quality motherchild relationship, low maternal support, inadequate parenting and
authoritative parenting; Blitstein, Murray, Lytle, Birnbaum, & Perry, 2005;
Blum et al., 2003; Wong et al., 2010). Parenting of fathers and the quality of
the relationship between father and adolescents appeared however to have a
similar effect on girls’ and boys’ delinquent behaviour (Blitstein et al., 2005;
Wong et al., 2010). Fagan et al. (2007) found the opposite effect between
delinquency and the parent-child relationship of both parents for boys and
girls: attachment to mother had an equal negative effect on boys’ and girls’
delinquency, but attachment to father appeared to be more strongly related to
boys’ delinquency. This might be because, contrary to the other studies that
examined the parent-child relationship, Fagan et al. (2007) examined serious
delinquency instead of general delinquency.
Regarding peers, many studies have shown that boys were more
strongly affected by delinquent peers than girls (Fagan et al., 2007; Giordano
et al., 1986; Mears et al., 1998 Steffensmeier & Haynie, 2000). In addition,
peer rejection appeared to affect boys’ delinquency to a greater extent than
girls’ delinquency (Moffitt et al., 2001). Van der Rakt, Weerman, and Need
(2005) studied self-reported delinquent behaviour of a Dutch school sample
and showed that the extent of friends’ delinquency was uniquely associated
with girl delinquency, but not with boy delinquency. Their results might be
inconsistent with other studies, because they had operationalized peer
delinquency differently. In their study, they had asked all respondents to list
their friends within their school. Since these friends had also participated in
the study and had reported on their delinquent behaviour, the authors could
construct accurately how many delinquent friends all respondents had. It is
therefore not yet understood how and whether peer delinquency
differentially influences boy and girl delinquency.
Limitations in research on differential exposure and vulnerability
Previous studies were limited in some important ways and it is therefore not
clear whether differential exposure and vulnerability is indeed related to the
sex difference in delinquency. Firstly, the few studies that examined
differential vulnerability, mostly did not test whether factors were actually
differently associated with boy and girl delinquency, but only compared the
68
levels of significance. This can lead to wrong conclusions about sex
differences. Although a factor might be related to both boy and girl
delinquency, the influence may still be sex different if the effect is stronger
for one of the sexes.
Secondly, the fact that differences between girls and boys are found in
mean levels of risk and promotive factors and in vulnerability, does not
necessarily mean that differential exposure and vulnerability are related to
the sex difference in delinquency. To test the risk exposure and risk
vulnerability hypotheses properly, it must be shown that these differences in
risk and promotive factors between boys and girls are actually related to the
sex difference in delinquency. So far, four studies have examined whether
differential exposure is related to the gap between boys and girls in
delinquent behaviour. To our knowledge, this has not been investigated for
differential vulnerability. Moffitt et al. (2001) tested the risk exposure
hypothesis on the Dunedin sample, which contained approximately 1000
juveniles, by comparing the sex difference in antisocial behaviour before and
after controlling for sex differences in mean levels of risk factors. They
showed that the sex difference in antisocial behaviour could be partly
explained by differential mean levels of risk factors such as harsh parental
discipline, neuro-cognitive risk, hyperactivity and peer influences. Messer,
Goodman, Rowe, Meltzer, and Maughan (2006) used the same procedure
and found as well, among preadolescent juveniles, that the higher exposure
of boys to harsh punishment, neuro-cognitive risk and peer problems
accounted for the sex difference in delinquency, as well as lower reading
performance of boys and higher prosocial behaviour of girls. These risk
factors were all assessed through parents and teachers though, and therefore
can be already sex biased. Junger-Tas et al. (2004) tested the risk exposure
hypothesis by examining a large European sample of juveniles and also used
the same analytical strategy. They showed that sex differences in parental
supervision could partly explain the difference between boys and girls in
delinquency. Sex differences in school problems and in the quality of
relationship with parents did not contribute to the explanation though. Fagan
et al. (2007) examined whether sex differences in serious delinquency were
mediated by sex differences in exposure to factors such as the parent-child
relationship and found only partial mediation, indicating, again, that
differences in quality of parent-child relationship between boys and girls
could only partially explain the sex difference in delinquency.
Thirdly, previous research focused on separate factors whereas it is also
important to see whether boys and girls are differently exposed to or
vulnerable for domains of risk factors. That is because it is often a
accumulation of risk factors that affect delinquency (Sameroff, Seifer,
Barocas, Zax & Greenspan, 1987). To capture the imperative concurrent
effects of risk and promotive factors on delinquency, that is neglected in
previous univariate analyses, it is necessary to do multivariate research.
69
Fourthly, no study has ever investigated the combination of differential
exposure and vulnerability.
The Current Study
In the current study, we will test the risk exposure and the risk vulnerability
hypotheses, where we deal with the limitations of previous studies. Both
hypotheses will be tested separately, as well as combined. In addition we
investigate whether the hypotheses are related to the sex difference in
delinquency by examining whether the sex difference in delinquency is
related to differential exposure and/or differential vulnerability. Since
previous studies suggested that differential exposure and differential
vulnerability might exist in the family as well as the peer domain, we will
examine a broad range of risk and promotive factors in each of these
domains. Obviously, the exposure and vulnerability hypotheses may hold for
different risk and promotive factors, i.e. it may be that for some factors,
differences in exposure are associated with the sex difference in
delinquency, while for others differences in vulnerability play a larger role.
The two hypotheses and the combination of the two will be analyzed
univariately, as well as multivariately.
Method
Sample
Data for this study come from the RADAR study (Research on Adolescent
Development and Relationships), an ongoing prospective cohort study aimed
at understanding the interplay between adolescent relationships and their
development in many domains (Van Lier et al., submitted). In this study,
adolescents, both their parents, a sibling (between the ages of 10 and 20),
and a best friend are followed across adolescent ages 12-17 years.
Assessments began in 2006 when the adolescents were in the first grade of
secondary school. The current study used data from this first annual wave,
which contained 480 adolescents for whom data on delinquency were
available.
The RADAR study focuses on delinquency development in particular.
Therefore, adolescents at high risk of developing delinquent behaviour were
oversampled. To indicate high risk, teachers in the sixth grade of primary
school rated behaviour problems of the adolescents (See Van Lier et al.
(submitted) for detailed information about this procedure). Of the sample,
200 children were at elevated risk of developing problem behaviour and 280
children were at average risk. The sample contained 56% boys (n = 269) and
44% girls (n = 211). The mean age was 13 years (SD = 0.52).
All respondents were asked to fill in an annual questionnaire during a
home visit by trained interviewers. The questionnaires contained questions
about the adolescent’s behaviour, personal characteristics and relationships
between family members and between the adolescent and the best friend.
70
Measures
Delinquency was measured with a questionnaire covering 25 delinquent acts,
based on the ISRD (Junger-Tas et al., 2003). The items included public order
offences (i.e. destroying means of transport; destroying somebody’s house;
destroying the school building; destroying property; arsony), property
offences (i.e. shoplifting less than five euro’s and more than five euro’s;
theft at school less than five euro’s and more than five euro’s; theft at home
less than five euro’s and more than five euro’s; stealing a bike; theft
from/out of car; burglary; other theft; selling stolen goods), violent offences
(i.e. threatening somebody to scare somebody or to get something; robbery;
hurting with weapon; beating up without injury; beating up with injury;
sexual offences) or drug offences (i.e. soft drug selling; party drug selling;
hard drug selling). For each of the offences, respondents indicated how often
they had committed them in the previous year. We computed a dichotomous
delinquency scale (0 = not delinquent, 1 = delinquent) based on whether or
not the adolescent had committed at least one offence in the previous year.
The same delinquency scale was used for the adolescent’s best friend’s
delinquency, filled in by the best friend him/herself (see also Table 4.1).
Measures of risk and promotive factors are described in Table 1. The
factors comprised five sub domains: 1) mother-child relationship (motherchild conflict, involvement in peer relations, negative interaction, power,
support), 2) father-child relationship (father-child conflict, involvement in
peer relations, negative interaction, power, support), 3) maternal parenting
(parental knowledge, adolescent disclosure, parental solicitation, parental
control), 4) paternal parenting (parental knowledge, adolescent disclosure,
parental solicitation, parental control), 5) peers (extent of best friend’s
delinquency, number of delinquent friends, best friend’s truancy). All the
measures were filled out by the adolescents except for extent of best friends’
delinquency and best friends’ truancy, which were filled out by the best
friend.
71
72
Parental direct interference
in peer relations
Knowledge of parents about
the child’s whereabouts
The extent to which parents
actively ask their children
about their whereabouts
Parenting Practices
(Kerr & Stattin 2000)
Parenting Practices
Parental knowledge
Parental solicitation
Mean of 6 5-Point
Likert items (for
both parents)
Mean of 9 5-Point
Likert items (for
both parents)
Mean of 6 4-Point
Likert items
Sum of 25 items
and dichotomized
total into not
delinquent versus
delinquent
Whether adolescents had
committed particular
offences in the previous 12
months
Parental Management
of Peers Inventory
(Mounts, 2002)
How created?
Description
Involvement in peer
relations
Table 4.1
Description of constructs used in this study
Instruments
Constructs
International SelfDelinquency (of
adolescent and best reported Delinquency
Study (Junger-Tas,
friend)
Haen-Marshall, &
Ribeaud, 2003).
mother =
.74; father
= .75
mother =
.80; father
= .82
.69
Reliability
(alpha)
NA
- How often does your mother/
father talk to your friends?
- How often does your
mother/father start a
conversation about school?
- Does your mother/father
know what you do in your
spare time?
know how you spend your
money?
- I tell my child sometimes
which friends are good for
him/her in my opinion
- I want my child to associate
with peers who work hard at
school
- How often have you
damaged the school in the
previous twelve months?
- How often have you stolen a
bike in the previous twelve
months?
Example of item
73
Controlling the child’s
freedom by setting rules and
restrictions
Parenting Practices
Whether the bond between
parent and child is based on
affection, admiration,
alliance, intimacy, and
companionship
The control parents have in
the relationship with their
child
Network of
Relationships Inventory
(Furman & Buhrmester
1985)
Network of
1985)
Parental support
Parental power
The extent to which
children spontaneously tell
their parents about their
whereabouts
Description
Instruments
Adolescent disclosure Parenting Practices
Constructs
Parental control
Mean of 6 5-Point
Likert items (for
both parents)
Mean of 8 5-Point
Likert items (for
both parents)
Mean of 6 5-Point
Likert items (for
both parents)
Mean of 6 5-Point
Likert items (for
both parents)
How created?
mother =
.88; father
= .82
mother =
.84; father
= .86
mother =
.70; father
= .75
Reliability
(alpha)
mother =
.84; father
= .83
mother/father have his/her
way when you disagree?
mother/father take a decision
for the both of you?
admire and respect you?
- How often do you have fun
with your mother/father?
- How often do you talk with
your mother/father about
school subjects?
- Do you have many secrets
for your mother/father about
what you do in your spare
time?
- Do you need permission of
your mother/father to get
home late on a school night?
demand to tell how you
spend your money?
Example of item
74
Description
Whether the parent-child
bond is characterized by
punishment and irritation
Frequency conflicts
between adolescents and
mother/father
Instruments
Network of
1985)
Parent-child conflicts Interpersonal Conflict
Questionnaire (Laursen,
1993)
Number of friends they
thought had committed an
offence in the previous
three months
How often truant in the
previous three months
Basic questionnaire
peers
School achievement
questionnaire
Delinquent friends
Truancy of best
friend
Parental Negative
interaction
Constructs
1 5-Point Likert
item
Summing 5 items
Sum of 10 5-Point
Likert items (for
both parents
Mean of 6 5-Point
Likert items (for
both parents)
How created?
NA
.72
mother =
.84; father
= .82
mother =
.88; father
= .86
Reliability
(alpha)
- How often have you been
truant in the previous three
months?
- How many of your friends
have damaged something on
the streets
have injured somebody by
hitting or with a weapon?
- How often have you had
conflicts with your
mother/father about money
in the previous week?
conflicts with your
mother/father about
relationships in the previous
week?
- How often do you disagree
and argue with your
mother/father?
- Are you and your
mother/father irritated by
each other’s behavior?
Example of item
Analytic Strategy
Multiple group path analyses were performed in MPlus version 6.1 (Muthén
& Muthén, 1998-2010). In our multiple group comparisons, means and
association paths of the model, depending on the research question, were set
equal or set free across the sexes.
To test our hypotheses, a series of models were run. We started with a
null model in which mean levels of risk and promotive factors as well as the
regression paths of each factor on delinquency were set equal for boys and
girls. This model simply represents that girls and boys differed neither in
exposure nor in vulnerability. In a second model, Model A, we tested the
exposure explanation by setting the mean levels of risk and promotive
factors free across sexes; the regression coefficients on delinquency of each
factor were kept equal for boys and girls. If this model fitted the data better
than the base model, this confirmed that boys and girls differed in mean
levels of the factor. In our third model, Model B, we tested the vulnerability
hypothesis. Mean levels of the factors were set equal for boys and girls, but
the regression coefficients of the factors were allowed to differ across sexes.
If this model fitted the data better than the base model, this would imply that
boys and girls are differently influenced by the risk factor. Finally, we
analysed a model, Model C, in which both mean levels and regression
coefficients were allowed to differ for boys and girls. If this model had a
better fit than the former models, this would imply that boys and girls were
concluded to differ in exposure as well as in vulnerability to that factor. In
all of the models described above (see also Figure 4.1), the probability4 for
delinquency was allowed to differ for boys and girls, because it is well
established that this differs across sexes.
All these models were run for each factor but also multivariately. We
carried out the multivariate analyses in two ways: per domain (by grouping
factors together that belonged to a particular domain) and based on the
univariate results (by grouping factors together for which sex differences
were found in the univariate analyses). After the multivariate analyses, the
best model was tested against a model, the Final Model, with the same
restrictions except regarding the probability for delinquency, which was now
set equal for boys and girls (see Figure 4.2). If this model performed equally
well as the model in which the probability was allowed to differ across
sexes, the sex difference in delinquency was related to differential exposure
and/or vulnerability. If the model performed significantly worse than the
4
In reality, we allowed the threshold for delinquency to differ across sex. The
threshold for delinquency is the statistical cut-off point from which the delinquency
scale changes in value from 0 (not delinquent) to 1 (delinquent). You can compare
this to the intercept of equations concerning continuous outcomes. For a fuller
understanding of the meaning of thresholds, see Long (1997). Thresholds actually
reflect probabilities. Since this term is much easier to understand, we use
probabilities instead of thresholds.
75
model with sex different probabilities, this indicates that the sex difference
in delinquency was not related to differential exposure or vulnerability.
Goodness of fit of the models was evaluated through the comparative fit
index (CFI; acceptable values > .90) (Bentler, 1990) and the root-meansquare error of approximation (RMSEA; acceptable values < .08) (Browne
& Cudeck, 1993). To compare the difference in fit of the different nested
models, we used the chi-square difference test.
Results
Sex differences in delinquency
More boys than girls in our sample had committed at least one offence in the
previous year (respectively 54.6% and 38.9%, p<.05), the boy to girl ratio is
1.4 : 1 (see Table 4.2). Many of these boys and girls reported to have
committed only one type of offence (respectively 40.4% and 53.7%). Most
offences were committed by more boys than girls. The offence committed
mostly by both boys and girls was beating up somebody without injury
followed by shoplifting for items worth less than 5 euros (approximately US
$7).
Sex differences in exposure and vulnerability
Univariate multiple group analyses were carried out for each of the factors.
These results will only be discussed shortly, but the detailed results can be
found in the Appendix. Girls and boys appeared to differ in mean levels
regarding several factors: girls had higher means of adolescent disclosure to
mother and parental solicitation of mother, whereas boys had a higher
number of delinquent friends. Differential vulnerability was found for
maternal support, parental control of mother and the extent of best friends’
delinquency. Maternal support was more strongly related to girl delinquency
than to boy delinquency. Regarding parental control of mother it was found
that it was negatively related to delinquency for boys, but positively for girls.
Additionally, girls but not boys appeared to be affected by the extent of
delinquent behaviour of their best friend. For the other variables boys and
girls no differential exposure or sensitivity was found.
76
77
≠
=
≠
Probability
delinquency
Extraversion
Probability
delinquency
Extraversion
≠
≠
=
Probability
delinquency
Extraversion
Differential vulnerability to factors
♂
♀
Probability
delinquency
Extraversion
≠
≠
≠
Probability
delinquency
Extraversion
77
Differential exposure and vulnerability
factors ♀
♂
Note. ‘=’ between squares indicate that the means are equal for boys and girls, ‘≠’ that they are unequal. ‘=’ between arrows indicate that the association
between the factor and delinquency are equal for boys and girls, ‘≠’ that they are unequal.
Probability
delinquency
Extraversion
Differential exposure to factors
♂
♀
Figure 4.1
Nested model types of differential level and differential vulnerability
Figure 4.2
The sex difference in delinquency explained?
The sex difference in delinquency explained
♂
♀
Extraversion
?
Extraversion
?
Probability
delinquency
=
Probability
delinquency
Note. The best model out of the three models in Figure 1
was tested against the model in Figure 2 in which the mean
score of delinquency in boys and girls is equal. Since the
other restrictions are similar to those of the best previous
model, we have printed ‘?’ in the model.
78
Table 4.2
Prevalence of offences, by sex (%)
Destroying means of transport
Destroying somebody’s house
Destroying the school building
Destroying property
Arson
Shoplifting less than five euro’s
Shoplifting more than five euro’s
Theft at school less than five euro’s
Theft at school more than five euro’s
Theft at home less than five euro’s
Theft at home more than five euro’s
Stealing a bike
Theft from/out of car
Burglary
Other theft
Selling stolen goods
Threatening somebody to scare
somebody or to get something
Robbery
Hurting with weapon
Beating up without injury
Beating up with injury
Soft drug selling
Party drug selling
Hard drug selling
Sexual offence
Total
One or more offences committed
* sex difference at p < .05
Girls
(n = 211)
0.9
0.5
1.4
2.4
5.7
10.9
0.9
3.3
0.0
6.2
2.4
0.5
0.5
0.0
1.4
0.5
6.2
Boys
(n = 269)
8.9*
4.5*
8.2*
10.1*
11.9*
15.2
3.4
10.0*
2.2*
11.2*
3.4
3.7*
1.9
2.6*
5.2*
4.5*
9.7
0.0
0.0
22.3
4.7
0.9
0.0
0.0
0.5
1.1
1.9
31.7*
12.3*
0.7
0.4
0.7
0.7
38.9
54.6*
Consecutively, we carried out multivariate multiple group analysis of
the five sub domains (see Table 4.3) For peers, none of the multivariate
models had an acceptable fit and these are therefore not discussed and not
displayed in Table 3. For most of the other domains (mother-child
relationship, father-child relationship and parenting of father) the best fitting
model was the Null Model in which mean levels and associations were set
equal across sex. So for these factors, boys and girls did not differ in mean
levels or vulnerability. For the sake of simplicity, only the fit indices for the
Null Model are displayed in Table 3 for these models.
79
Table 4.3
Fit indices for multivariate models
Model
Mother-child relationship
Null Model: Fixed
χ2
df
CFI RMSEA
11.54
11
.99
.01
Father-child relationship
Null Model: Fixed
10.03
11
1.00
.00
26.54*
11.34*
9
5
.95
.98
.09
.07
16.63*
5
.96
.10
1.70
1
.99
.05
4.92
2
.91
.08
Maternal parenting
Null Model: Fixed
Model A: Differential exposure
Difference with Null Model
Model B: Differential vulnerability
Model C: Differential exposure and vulnerability
Difference with Model A
Difference with Model B
Final Model: No sex difference in delinquency
Difference with best previous model (C)
Δχ2
15.81*
10.46*
10.46*
15.81*
6.62
Paternal parenting
Null Model: Fixed
11.37
8
.99
.04
Note. CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of Approximation.
* p < .05.
Regarding maternal parenting the best model appeared to be the one in
which boys and girls differed in mean levels and vulnerability. Girls had
higher mean levels of maternal parenting than boys. Additionally, girls were
protected in general more strongly from delinquency through maternal
parenting than boys. Parental knowledge was negatively associated with
delinquency of both boys and girls, but more strongly with delinquency of
girls (see also Table 4.4). However, maternal control was not related to boys’
delinquency, but positively related to girls’ delinquency, indicating that
delinquent behaviour of girls correlated with higher levels of maternal
control. The fit worsened significantly by setting the probability for
delinquency equal for boys and girls. This means that differential exposure
to and sensitivity for maternal parenting factors are not related to the sex
difference in delinquency.
Next to the multivariate analyses per domain, we have done some
multivariate analyses based on the univariate results. The fit indices for these
models are displayed in Table 4.5. First, we took all factors to which boys
and girls were differently exposed univariately and put them in one model.
These were the number of delinquent friends, adolescent disclosure to
mother and parental solicitation of mother. Then we followed the same
analytic strategy as before by comparing different models that represented
differential exposure and vulnerability. As expected, the model in which
boys and girls had different mean levels of these factors had the best fit.
80
Table 4.4
The maternal parenting model with sex differences in mean levels and vulnerability
Means (SD)
B (se)
Boys
Girls
Boys
Girls
Parental knowledge M
8.05
8.03
-0.17*
-0.25*
(1.51)
(1.66)
(0.06)
(0.10)
Adolescent disclosure M
7.11
7.56
-0.13
-0.03
(1.52)
(1.67)
(0.08)
(0.10)
Parental solicitation M
5.16
5.67
0.07
0.03
(1.73)
(1.82)
(0.05)
(0.06)
Parental control M
6.47
6.48
-0.06
0.15*
(2.46)
(2.26)
(0.04)
(0.06)
* significantly related to delinquency (p < .05)
This model was compared with the same model where the probability
for delinquency was set equal for boys and girls. It appeared that the model
in which boys and girls had different mean levels on the factors, but an equal
probability for delinquency was the best model. This indicates that the sex
difference in delinquency is related to sex differences in exposure towards
the number of delinquent friends, adolescent disclosure to mother and
parental solicitation of mother. More specifically, the fact that girls are less
likely to be delinquent is related to girls have higher levels of adolescent
disclosure with mother and parental solicitation of mother and less
delinquent friends than boys. Second, we took all factors for which boys and
girls showed differential vulnerability in the univariate analyses (i.e.
maternal support, maternal control and the extent of best friends’
delinquency), put them in one model, and ran again a series of models. The
best model was the model in which boys and girls were differently
vulnerable for the risk factors, but the model did not improve by setting the
probability for delinquency equal among boys and girls. This indicates that
the sex difference in delinquency was not related to differential vulnerability.
Third, we put all factors in a model boys and girls were differently exposed
to or differently vulnerable for in the univariate analyses. The best fitting
model indeed appeared to be the one in which boys and girls showed
differential exposure to and vulnerability for the factors. However, the model
did not improve if the probability for delinquency was set equal for boys and
girls. In other words, the sex difference in delinquency is not related to this
model of differential exposure and vulnerability.
81
Table 4.5
Fit indices for multivariate models
Model
All differential exposure factors
Null Model: Fixed
Difference with best previous model (A)
All differential vulnerability factors
Null Model: Fixed
Difference with best previous model (B
χ2
df
CFI RMSEA
28.38*
2.59
7
4
.92
1.00
.11
.00
24.68*
4
.92
.15
1.72
1
1.00
.06
5.02
5
1.00
.00
17.77*
14.05*
7
4
.78
.80
.08
.10
5.02
4
.98
.03
1.14
1
1.00
.02
6.95
5
.96
.04
Δχ2
23.48*
0.74
0.74
23.48*
2.74*
4.44
14.26*
14.26*
4.44
2.06
All differential exposure/vulnerability factors
Null Model: Fixed
45.85*
13
.92
.10
17.18*
7
.97
.08
26.93*
28.15*
7
.95
.11
17.17*
1.14
1
1.00
.02
17.17*
3.39
2
1.00
.05
Difference with best previous model (C)
3.48
Note. CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of Approximation.
* p < .05.
Discussion
It is a fact that there are sex differences in offending behaviour. However,
the origin of these sex differences is not yet understood. To better understand
the sex difference, this study addressed the risk exposure hypothesis –
whether the sex difference in delinquency is related to sex differences in
mean levels of parent and peer factors – and the risk vulnerability hypothesis
– whether the sex difference is related to differences in vulnerability of
parent and peer factors. Factors were considered in the sub domains motherchild relationship, father-child relationship, maternal parenting, paternal
parenting, and peers. In addition, we examined the risk exposure and risk
82
vulnerability hypotheses concurrently. Innovative was that we not only
tested whether differences in mean levels or vulnerability between boys and
girls (co-)existed, but also whether these were related to the sex difference in
the prevalence of delinquency. We used an eminently suited data set with
rich and relevant data for the research question at hand. Data had been
gathered – without method differences – for both boys and girls.
Sex differences in Mean Levels and Vulnerability
The sex ratio in delinquency in our sample (i.e. 1.4 : 1) was comparable to
that of 12-and-13-year olds in a Dutch nationally representative self-report
survey (Van der Laan & Blom, 2006). Our study showed that this sex
difference disappeared when differences in mean levels of adolescent
disclosure to mother, parental solicitation of mother, and number of
delinquent friends between boys are taken into account. Put differently, a
reason why boys are more likely to commit offences than girls appears to be
that girls talk more with their mothers than boys, that mothers are involved
more with their daughters than with their sons, and that boys have more
delinquent friends than girls. It makes sense that girls’ involvement with
their mother and vice versa is higher than that of boys, since girls are much
more social and more family oriented. That this is reflected in the reciprocal
behaviour towards their mother and not towards their father is most likely
because mothers are also female and have this same social tendency. The
lower number of delinquent friends of girls was also expected since, as
discussed before, girls more often have female friends whereas boys hang
around more often with male friends.
In light of the social control theory (Hirschi, 1969) and the differential
association theory (Sutherland, 1947) these results might imply that girls are
less likely to be delinquent because they do not want to endanger their high
involvement with their mothers by going against their mothers will and
therefore they have less difficulties with resisting the influence of delinquent
friends. Furthermore, girls do not happen to have a lot of delinquent friends
which makes it even less difficult to get affected by them. This indicates that
if girls had both similar levels of involvement with their mothers as boys ánd
a similar number of delinquent friends, there might not have been a sex
difference in delinquency. We could speculate that to decrease the gender
gap in delinquency, preferably by reducing delinquent behaviour of boys,
both the boys’ level of disclosure towards mothers and the interest of
mothers towards sons should be improved, but boys should also be
encouraged not to hang around with (same sex) delinquent peers anymore.
Then again, we found for maternal parenting, that sex differences
existed in mean levels as well as in vulnerability. This shows that parenting
behaviour of mothers is not only different for boys and girls, but it also has
different effects on boys and girls. Mothers appeared to be involved more
with their daughters than with their sons, in terms of parental solicitation.
83
Moreover, the higher level of adolescent disclosure for girls towards mothers
shows us that girls tell their mothers more than boys do. Girls are therefore
protected more from delinquency than boys. Besides, mother who had higher
levels of parental knowledge protected their daughters more than their sons.
This indicates that girls probably not only absent themselves from
delinquency because they have more to lose in terms of maternal
involvement than boys, but also because they are more vulnerable for the
influence of maternal parenting. The reason for this higher vulnerability
might be that girls are more family-focused than boys and are taught to care
for others. This may result in higher levels of social control: girl do not show
delinquent behavior since they do not want to risk losing the good bond with
their mothers. A finding that could contradict this suggestion is that girls’
delinquency was found to be related to high levels of maternal control,
indicating that maternal control is a risk factors for delinquency in girls.
However, it should be noted that these are cross-sectional results and
therefore, it might very well be that mothers control their daughters more
when they show delinquent behaviour and not the other way around.
Interestingly, while mothers had a differential influence on boys and
girls, the influence of paternal parenting on delinquency was similar across
sexes. This is in line with the studies of Blitstein et al. (2005) and Wong et
al. (2010). Hence, an important addition to previous studies that displayed
that mothers can have different influences on boys’ and girls’ delinquency
development (Moffitt et al., 2001; Pedersen, 1994) is that girls are more
exposed to the influence of mothers.
Strengths and Limitations
This study was limited because of its cross-sectional nature; we could not
distinguish cause and effect or reciprocal effects like in longitudinal studies.
Parenting practices, for instance, might be influenced by the delinquent
behaviour of the adolescent, or the other way around (Kerr & Stattin, 2003).
Also, note that cross-sectional results are generally stronger than
longitudinal results. In addition, it has to be taken into account that
respondents in this study were 12 to 13 years of age and that results cannot
be generalized to adolescents of all ages. It might for instance be that results
are different for older adolescents, for whom parents become less important
and peers more important. Replications for other age groups are therefore
necessary.
A strong point of this study was the direct measurement of best friends’
delinquency. Most studies assessed friends’ delinquency by asking
respondents about their friends. In this way, it is likely that similarity with
friends’ delinquency is overestimated, because adolescents’ belief about
their friend’s delinquency is based on their own assessment. Another strong
point was that not only differences in mean levels and vulnerability were
84
examined, but also whether they co-existed and whether these were actually
related to the sex difference in offending.
Hoyt and Scherer (1998) identified several approaches for explaining
delinquent behaviour of girls. One of these approaches claims that there is no
need for a different theoretical framework to account for female
delinquency, but that existing theories, mostly based on research on males’
delinquent behaviour, can be used to explain female behaviour. The reason
that girls are less often delinquent than boys is simply that they are less
exposed to risk factors than boys are. Another approach states that existing
theories are not adequate for the explanation of female delinquency. Girls
are thought to have different risk factors and therefore need other theories to
clarify their delinquent behaviour. Hence, this approach assumes that girls
and boys differ in their vulnerability for risk factors. Our results indicated
that girls have many similar risk factors as boys, although to some of these
risk factors they were less exposed, but also that girls and boys have some
different risk factors. These results imply that existing (male) theories can be
used to explain girl delinquency, but that these are not sufficient; gender
specific adaptations are probably needed.
85
86
Appendix
None
Vulnerability
Power M
Support M
None
None
NAF
None
Involvement in peer relations F
Negative interaction F
Power F
Support F
None
None
Negative interaction M
Father-child relationship
Father-child conflict
None
None
Model
(sex difference in)
Involvement in peer relations M
Mother-child relationship
Mother-child conflict
Univariate multiple group analyses
2.99
(1.75)
2.52
(2.02)
1.21
(1.34)
5.12
(1.91)
5.12
(1.78)
3.21
(1.89)
2.78
(2.07)
1.75
(1.56)
4.89
(1.76)
5.71
(1.93)
3.01
(1.63)
2.78
(1.93)
1.31
(1.41)
3.04
(1.66)
2.95
(2.03)
1.57
(1.34)
5.05
(1.61)
5.48
(1.77)
Means (SD)
Boys
Girls
B (se)
Girls
-0.21*
(0.04)
-0.05* (0.01)
0.10* (0.03)
0.10* (0.04)
0.16* (0.03)
-0.11*
(0.04)
-0.04 (0.03)
0.14* (0.04)
0.10* (0.04)
0.16* (0.03)
Boys
87
Vulnerability
Parental control M
NAF
Parental control F
7.19
(1.81)
6.75
(1.71)
4.46
(1.74)
8.05
(1.51)
7.11
(1.52)
5.16
(1.73)
6.47
(2.46)
7.11
(1.78)
6.88
(1.94)
4.46
(1.93)
8.03
(1.66)
7.56
(1.67)
5.67
(1.82)
6.48
(2.26)
Means (SD)
Boys
Girls
B (se)
Girls
0.09*
(0.05)
-0.09* (0.03)
-0.18* (0.03)
-0.16* (0.03)
-0.09*
(0.04)
-0.07* (0.03)
-0.20* (0.03)
-0.22* (0.03)
Boys
0.41* (0.03)
1.77
0.67
(2.42)
(1.12)
Best friends delinquency (scale of
Vulnerability
1.10
1.18
0.04
0.15*
variance)
(1.37)
(2.12)
(0.03)
(0.04)
Truancy best friend
None
0.87
0.56
0.12* (0.05)
(1.50)
(1.11)
Note. Means in bold refer to means that are significantly different for boys and girls. When boys and girls do not
differ in their B’s, the single B is shown in the table. When B’s are not equal, two separate B’s are shown.
* p < .05
NAF = model is not of acceptable fit
Mean levels
None
Parental solicitation F
Peers
Number of delinquent friends
None
Adolescent disclosure F
None
Mean levels
Parental solicitation M
Paternal parenting
Parental knowledge F
Mean levels
None
Adolescent disclosure M
Maternal parenting
Parental knowledge M
Model
(sex difference in)
Chapter 5
The sex difference in delinquency: Do
girls need a bigger push?
Wong, T. M. L., Slotboom, A., Bijleveld, C. C. J. H., Van Lier, P. A. C., Meeus, W.
H. J. & Koot, J. M. (submitted). The sex difference in delinquency: Do girls need a
bigger push?
89
This study is an attempt to gain insight in the aetiology of the sex difference
in delinquency by testing the threshold hypothesis which states that girls
have a higher risk threshold for delinquency than boys. In examining the
hypothesis the authors dealt with methodological problems of previous
research by focusing on self-reports of delinquency instead of arrest or
incarceration as indicator of delinquency, by studying general delinquency
as well as serious delinquency, and by taking risk factors and promotive
factors into account. A sample of 495 Dutch juveniles from the RADAR
study, a prospective cohort study, is analyzed. This study did not find
evidence for the threshold hypothesis, but did find some interesting results
regarding the dose-response relationship.
90
Introduction
The sex difference in adolescent delinquency is widely acknowledged. Girl
offending is generally less serious than boy offending and its prevalence is
lower (Junger-Tas, Marshall, & Ribeaud, 2003; Van der Laan & Blom,
2006, Steffensmeier & Schwartz, 2009). An explanation often brought up for
this sex difference is that girls are being socialized against delinquent
behavior, whereas boys are socialized towards it. Boys are taught that it is a
good thing to be competitive and take risks and therefore they may hold
positive beliefs about aggressive and delinquent behavior. In contrast, girls
are taught to care for others, be passive and prudent, which is more likely to
result in negative attitudes towards aggression and delinquency (Agnew,
2009; Mears, Ploeger, & Warr, 1998; Moffitt, Caspi, Rutter, & Silva, 2001).
In other words, the aggressive and delinquent behavior repertoire of girls is
likely to be constrained by their environment through socialization or
modeling. As a result, it is a much bigger step for girls to become delinquent
than for boys, leading to a lower prevalence of delinquency among girls.
In line with this reasoning, it is thought that girls have a higher
threshold than boys for delinquency, the so-called threshold hypothesis, i.e.
that girls pass a higher critical ‘risk level’ in order to become delinquent
(Eme, 1992). This risk level that defines the risk threshold can be
operationalized in two ways : 1) the risk level indicates the severity or level
of a single risk factor, and 2) the risk level considers the number of risk
factors. This last way of operationalizing risk level is derived from the
cumulative risk approach (Rutter, 1979; Sameroff, Seifer, Barocas, Zax &
Greenspan, 1987) : the more risk factors someone experiences, the more
likely he or she is to be delinquent. There is evidence for such a doseresponse relationship between the number of risk factors and the likelihood
of delinquency for boys and girls (Johansson & Kempf-Leonard, 2009;
Loeber, Slot & Stouthamer-Loeber, 2008; Van der Laan & Van der Schans,
2010; Wong et al., submitted). However, the threshold hypothesis states that
for girls it requires more risk factors than for boys to start offending. If girls
are exposed to the same number of risk factors as boys, they are less likely
delinquent. Differences between boys and girls in such a threshold for
delinquency, while often posited, have hardly been studied empirically.
In the current study, we tested the threshold hypothesis following the
cumulative risk approach, since it has been found in general population
studies that it is rather the amount of risks than the level of risk that affects
the development of deviant behavior (Rutter, 1979; Sameroff, Seifer,
Barocas, Zax & Greenspan, 1987). Furthermore, this approach has been
noted for its potential to capture the natural covariation of risk factors
(Luthar, 1993).
A few studies used the same approach to sex differences in
delinquency. In the study by Alemagno, Shaffer-King, and Hammel (2006)
the overall number of risk factors in relation to delinquency was compared
91
among 250 incarcerated boys and girls. In accordance with the cumulative
risk hypothesis they found that detained girls were exposed to more risk
factors than detained boys. However, Van der Laan and Van der Schans
(2010) showed that the sex difference in thresholds may vary by risk
domain. They found that arrested girls were exposed to a higher number of
family risk factors than arrested boys, but that by contrast arrested boys
appeared to have a higher number of risky lifestyle factors compared with
arrested girls. Thus the threshold model may not apply to all risk domains.
Also, since females and males are not equally likely to show delinquent
behavior at different age periods (Junger-Tas et al., 2003; Van der Laan &
Blom, 2006), the sex variant threshold for displaying delinquency may also
be age variant. As Moffitt (1993) suggested, during puberty, it is almost
normative to show delinquent behavior as an expression of the discomfort
experienced due to the discrepancy between biological and social maturity
that is typical for adolescence. If so, it may be that girls in this period, like
boys, are more liable for delinquency and the threshold effect does not show
during this particular age period.
The studies that examined the threshold hypothesis according to the
cumulative risk approach only investigated adjudicated or incarcerated
juveniles. Without a non-delinquent comparison group however, it is
difficult to conclude whether one of the sexes indeed has a higher risk
threshold for delinquency, since it might be that one sex always (also in case
of non-delinquency) has higher risk levels than the other sex. Therefore,
differences in number of risk factors found in samples of delinquent boys
and girls alone are not enough to conclude on a higher threshold for girls.
Another limitation of the studies addressing the threshold hypothesis so
far is that it has been studied in youth who were regarded delinquent based
on being arrested or convicted while the hypothesis was not tested using
self-reports of delinquency. Girls and women, however, are often treated
differently in the juvenile justice system than boys and men (e.g. Daly,
1994). This is illustrated for instance by the bigger gender gap in official
records than in self-reported delinquency, which may result from the fact
that girls are less likely to be arrested than boys (Slotboom, Wong, Swier, &
Van der Broek, 2011). Therefore, the threshold for delinquency in such
studies cannot be separated from the threshold to be arrested, prosecuted or
convicted. Thus it is problematic to attribute sex differences in the number of
risk factors in officially delinquent samples to the threshold for delinquency.
Furthermore, adjudicated or incarcerated adolescents are mostly serious
delinquents, while researchers suggest a higher threshold for girls across the
whole range of delinquency. It should therefore be examined more directly
whether the threshold effect applies to general delinquency, to serious
delinquency or to both. As Moffitt (1993) argues, delinquent behavior seems
to be more normative for adolescents. Therefore, it might be that the
threshold hypothesis only holds for serious delinquency in adolescence.
92
In this study, we try to contribute to the literature on the threshold
hypothesis by focusing on self-reported delinquency, and by considering
general as well as serious delinquency. However, there is one additional
issue that has to be taken into account in threshold research that regards
promotive factors. In examining the threshold, it is essential to include
promotive factors, next to risk factors, since adolescent delinquency is not
only affected by risk factors, but also by promotive factors. Promotive
factors are those factors associated with a decreased probability of
delinquency (Sameroff, Bartko, Baldwin, Baldwin, & Seifer, 1998;
Stouthamer-Loeber, Loeber, Wei, Farrington, & Wikström, 2002).
Promotive and risk factors are actually the opposite end of each other within
the same variable. That is, the same variable can constitute a risk as well as a
protective condition for juveniles. Having a bad relationship with parents,
for instance, can be a risk for the development of adolescent delinquency,
but having a good relationship with parents can serve as protection against it.
The relationship between this factor and delinquency is not necessarily linear
and the relative importance may differ for risk and promotive factors
(Sameroff, et al., 1998; Stouthamer-Loeber et al., 2002). Since the number of
promotive factors can neutralize the influence of the number of risk factors
(Stouthamer-Loeber et al., 2002; Van der Laan & Blom, 2006), not only the
number of risk factors should be considered to identify the risk level in
threshold studies, but also the number of promotive factors. This important
neutralization effect is ignored if promotive factors are not considered,
which might result in overstating the risk and in getting any threshold effect
wrong. To examine the threshold hypothesis properly, we therefore analyze
the relative risk level (i.e., the number of risk factors minus the number of
promotive factors) of boys and girls in relation to delinquency.
Research questions
In this study we tested to what extent 1) the threshold for delinquency is
different for boys and girls; 2) the threshold for serious delinquency is
different for boys and girls; 3) these differential thresholds vary across
different domains.
In examining these research questions we dealt with the methodological
problems of previous research by analyzing delinquent as well as nondelinquent boys and girls, by taking into account promotive factors next to
risk factors, by focusing on self-reported delinquency instead of arrest or
incarceration as indicator of delinquency, and by studying general
delinquency as well as serious delinquency. A broad array of factors are
considered that were shown to predict delinquency in previous studies
(Hoeve et al., 2009; Hubbard & Pratt 2002; Lipsey & Derzon, 1998; Maguin
& Loeber, 1996; Pratt & Cullen, 2005; Simourd & Andrews, 1994; Wong et
al., 2010; Zahn, 2009). These included individual (problem) factors (i.e.
personality characteristics, prosocial behavior, empathy, pubertal timing,
93
levels of self-worth, victimization), family factors (i.e. the quality of parentchild relationships, conflicts within the family, parenting practices, having
caregivers with a history of delinquency), school factors (i.e. school
achievement, truancy) and peer delinquency.
Methods
Sample
Data for this study come from the RADAR study (Research on Adolescent
Development and Relationships), a Dutch ongoing prospective cohort study
aimed at understanding the interplay between adolescent relationships and
their development in many domains (Van Lier et al., submitted). In this
study, adolescents, both their parents, a sibling (between the ages of 10 and
20), and a best friend are followed across adolescence at the age of 12 to 17
years. Assessments began in 2006 when the adolescents were in the first
grade of secondary school (n = 497). The current study used data from the
second and third wave, which contained 495 Dutch adolescents for whom
data on delinquency were available.
The RADAR study focuses on delinquency development in particular.
Therefore, adolescents at high risk of developing delinquent behavior were
oversampled. To indicate high risk, teachers in the sixth grade of primary
school rated behavior problems of the adolescents (for detailed information
about this procedure, see Van Lier et al., submitted). Of the sample used in
this study, 41.4% of the children were at elevated risk of developing problem
behavior. The sample contained 56.6% boys (n = 280) and 43.4% girls (n =
215). In the elevated risk sample, there were 57.1% boys and 42.9% girls.
The mean age was 13.5 years (SD = 0.61) at the second wave.
All respondents were asked to fill out questionnaires during an annual
home visit by trained interviewers. The questionnaires contained questions
about the adolescent’s behavior, personal characteristics and relationships
between family members and between the adolescent and the best friend (for
other published papers that used data from the RADAR study, see
Eichelsheim et al., 2010; Eichelsheim et al., in press; Hale et al., 2011;
Klimstra et al., 2010; Schwartz et al., in press; Yu, Branje, Keijsers, &
Meeus, 2011).
Delinquency measurements
Delinquency was measured in Wave 2 and 3 with a questionnaire covering
30 delinquent acts, among which were five minor offences, based on the
International Self-Reported Delinquency Scale (Junger-Tas et al., 2003).
These minor offences are not taken into account in the analyses, since the
severity of these offences is very low, but the prevalence relatively high and
therefore they would influence the delinquency scale disproportionately
high (see the Appendix for an overview of the 25 included offences). For
each of the offences, respondents and their best friends indicated how often
94
they had committed them in the previous year. For this study we summed all
offences and dichotomized the total of the sum into 0 (no offence committed
– not delinquent) and 1 (1 or more offences committed – delinquent).
Delinquency was not only measured annually in the RADAR study, but
also every three months through internet assessments. In the internet
assessments (of which there were three between the annual measurements)
respondents were asked whether they had committed offences in the
previous three months. During the annual assessments, however, it was
asked whether respondents had committed the offences in the previous
twelve months. It appeared that many adolescents filled out in the annual
assessment that they did not commit any offence while they had reported to
be delinquent some of the three monthly internet assessments. Since it was
clear that they had at least committed one of the offences in the previous
year, regardless of the exact period, these adolescents were assigned to the
delinquent category.
Serious delinquency was defined as all offences carrying a maximum
prison sentence of four years or more in the Netherlands (an overview of all
those offences can be found in the CBS Standaardclassificatie Misdrijven
1993). However, the seriousness of some of the offences was not clear and
therefore we did not include them in the serious delinquency scale5. The
items that were included were (attempt of) arson, theft of a bike, theft from a
car, burglary, handling stolen goods, robbery, weapon injury, violence with
injury, and sex offences. We dichotomized the total sum of all the items into
0 (no serious offence committed) and 1 (at least one serious offence
committed).
Other measurements
The factors used in this study are described in Table 5.1. All the risk and
promotive factors were measured at Wave 2. Previous studies were limited
because most of them were cros-sectional: risks and delinquent behavior
were measured at the same time. By contrast we used risk and promotive
factors of the data preceding the data wave in which delinquency was
assessed.
The factors comprised six domains: individual, negative life events,
parent-child relationship, parenting, school, and peers. Most of the constructs
were assessed through the adolescents. However, mothers had indicated
whether there were family members with contacts with the police or justice.
Besides, adolescents’ best friends reported about their own delinquent
behavior.
5
These included the items: theft with a value of more than five euro’s, threatened
somebody to scare that person or to let him do something.
95
96
The extent to which a
person is trustworthy,
altruistic, kind and
affectionate
person is thoughtful, has
good impulse control is
goal-directed behaviors
person is emotional
instable, anxious,
moody, and sad
person has imagination,
insight and broad range
of interest
Big Five
Big Five
Big Five
Conscientiousness
Neuroticism
Openness to
experience
.72
.85
Sum of 6 7-Point
Likert items
Sum of 6 7-Point
Likert items
.82
.78
.79
Reliability
(alpha)
Sum of 6 7-Point
Likert items
Sum of 6 7-Point
Likert items
Sum of 6 7-Point
Likert items
person is sociable,
talkative, assertive and
expressive
Big Five
How created?
Description
Agreeableness
Table 5.1.
Description of factors used in this study
Instruments
Constructs
Individual
Big Five
Extraversion
- Versatile
- Innovative
- Irritated
- Nervous
- Punctial
- Neat
- Pleasant
- Kind
- Reserved
- Talkative
Example(s) of item
97
Description
Difﬁculties to remain in
control of behavior when
experiencing negative
emotions
Behaviour that regards
helping others with no
thought of reward or
compensation
The tendency to
experience feelings of
sympathy and
compassion for
unfortunate others
The tendency to
spontaneously adopt the
psychological point of
view of others in
everyday life
persons’ self-concept is
clear, confident,
internally consistent and
stable over time
Instruments
Difficulties in Emotion
Regulation Scale (Gratz &
Roemer, 2004)
Prosocial behavior
Interpersonal Reactivity
Index (Davis, 1983)
Interpersonal Reactivity
Index (Davis, 1983)
Self-Concept Clarity
Prosocial behavior
Empathic concern
Perspective taking
Self-concept clarity
Constructs
Impulse control
difficulties
Sum of 12 5-Point
Likert items
Sum of 7 5-Point
Likert items
Sum of 7 5-Point
Likert items
Sum of 11 7-Point
Likert items
Sum of 5 5-Point
Likert items
How created?
.84
.66
.70
.93
Reliability
(alpha)
.76
- My ideas about myself are
often contradictory
- Sometimes I have the feeling
I know other people better
than myself
- I am often worried about
people who are worse off than
me
- If I see that someone is taken
advantage of, I tend to protect
this person
- Sometimes I try to better
understand my friends by
imagining myself in their
position
- During a disagreement, I try
to understand all positions
before taking a decision
- I am in general nice to other
people
- I try to involve other people in
conversations
- When I’m upset, I lose control
over my behaviors
- When I’m upset, I cannot
concentrate anymore
Example(s) of item
98
Victimized or not in the
previous year by 5
different offences
Victim of relational of
physical aggression
Whether a family
member was ever in
contact with the police
Whether a family
member was ever in
contact with justice
Proactive/Reactive
Aggression
NA
NA
Victimization
aggression
Family member with
police contact
Family member
contact with justice
The extent to which the
adolescents had noticed
own physical changes
Pubertal Development
NA
Description
Instruments
Victim offence
Negative life events
Pubertal
development
Constructs
1 dichotomous
item
1 dichotomous
item
Sum of 7 7-Point
Likert items
Dichotomized the
sum of 5
dichotomous
items
1 5-Point Likert
item
How created?
NA
NA
.84
NA
NA
Reliability
(alpha)
Theft
Threat of violence
Threat with a weapon
Violence with assault
Sexual offence
Other people try to let me do
things by physically
intimidating me
- If other people are angry with
me, they exclude me from
common activities
-
- How long is het ago that you
noticed that your body
changed
Example(s) of item
99
Description
bond is based on
affection, admiration,
alliance, intimacy, and
companionship
The control parents have
in the relationship with
their child
bond is characterized by
punishment and irritation
Frequency conflicts
between adolescents and
mother/father
Instruments
Network of Relationships
Inventory (Furman &
Buhrmester 1985)
Inventory (Furman &
Buhrmester 1985)
Inventory (Furman &
Buhrmester 1985)
Interpersonal Conflict
Questionnaire (Laursen,
1993)
Parental power
Parental Negative
interaction
Parent-child conflicts
Constructs
Mother/father
Parental support
Sum of 10 5-Point
Likert items (for
both parents
Mean of 6 5-Point
Likert items (for
both parents)
Mean of 6 5-Point
Likert items (for
both parents)
Mean of 8 5-Point
Likert items (for
both parents)
How created?
mother =
.84; father =
.82
mother =
.88; father =
.86
mother =
.88; father =
.82
mother =
.84; father =
.86
Reliability
(alpha)
conflicts with your
mother/father about money in
the previous week?
conflicts with your
mother/father about
relationships in the previous
week?
- How often do you disagree and
argue with your mother/father?
- Are you and your mother/father
irritated by each other’s
behavior?
mother/father have his/her way
when you disagree?
mother/father take a decision
for you both?
- Does your mother/father admire
and respect you?
- How often do you have fun with
your mother/father?
Example(s) of item
100
Description
Knowledge of parents
about the child’s
whereabouts
The extent to which
children spontaneously
tell their parents about
their whereabouts
The extent to which
parents actively ask their
children about their
whereabouts
Controlling the child’s
freedom by setting rules
and restrictions
Instruments
Parenting Practices (Kerr &
Stattin 2000)
Stattin 2000)
Stattin 2000)
Stattin 2000)
Parental knowledge
Adolescent disclosure
Parental solicitation
Parental control
Constructs
Mean of 6 5-Point
Likert items (for
both parents)
Mean of 6 5-Point
Likert items (for
both parents)
Mean of 6 5-Point
Likert items (for
both parents)
Mean of 9 5-Point
Likert items (for
both parents)
How created?
mother =
.84; father =
.83
mother =
.74; father =
.75
mother =
.70; father =
.75
mother =
.80; father =
.82
Reliability
(alpha)
- Do you need permission of
your mother/father to get
home late on a school night?
demand to tell how you
spend your money?
mother/father talk to your
friends?
mother/father start a
conversation about school?
- How often do you talk with
your mother/father about
school subjects?
- Do you have many secrets
for your mother/father about
what you do in your spare
time?
- Does your mother/father know
what you do in your spare time?
- Does your mother/father know
how you spend your money?
Example(s) of item
101
Peers
Delinquent friends
Truancy
Constructs
School
School achievement
How often truant in the
School achievement
questionnaire
Number of friends they
thought had committed
an offence in the
How well performed at
school in previous week
compared to class mates
School achievement
questionnaire
Basic questionnaire peers
Description
Instruments
Summing 5 items
1 5-Point Likert
item
1 10-Point Likert
item
How created?
.72
NA
NA
Reliability
(alpha)
have damaged something on
the streets
have injured somebody by
hitting or with a weapon?
- How often have you been
truant in the previous three
months?
-
- In comparison to your class
mates, how well did you
perform at school?
Example(s) of item
We performed preliminary analyses to identify the risk versus promotive
nature of the factors. To examine whether the factors were risk factors,
promotive factors or both we partly used the same method as StouthamerLoeber et al. (1993; 2000) and Van der Laan and Blom (2006). All factors
were trichotimized into a promotive, a neutral and a risk component using
the sex-specific 25th and 75th percentiles of the variable distributions as cutoffs, if applicable. For some variables it was more appropriate to
trichotimize according to the categories (see notes of Table 2). All these
factors were recoded into two dummy variables: a risk variable and a
promotive variable. The reference category in each variable was the neutral
component. For some other variables it was only possible to dichotomize the
values (i.e., victimization of delinquency, family members with police
contacts, family members with justice contacts, having a delinquent best
friend) given the content of the variable. These were all considered to be risk
factors and had therefore only risk variable.
The risk and promotive variable that were computed for each factor
were analyzed in a logistic regression. If the risk variable was related to
delinquency (p less than 0.05), we considered the variable to be a risk factor.
If the promotive variable was related, the variable was considered to be a
promotive factor. If both variables were related to delinquency, the factor
was both a risk as well as a promotive factor.
To examine whether girls have a higher threshold for delinquency than
boys, we created three types of variables: one that indicated the number of
risks, one that indicated the number of promotive factors, and one that
indicated the relative level of risk. The relative risk level indicated the
number of risk factors minus the number of promotive factors. These
variables were created for each domain.
All the previous mentioned preliminary analyses were carried out for
general delinquency as well as for serious delinquency.
After these analyses we carried out a logistic regression to investigate
whether the risk level variable, sex and the interaction between risk level and
sex could predict delinquency (to avoid multicollinearity, the interaction
terms were constructed with the centered version of the risk level variable).
In the regressions we controlled for delinquency at Wave 2. If the interaction
term significantly predicted delinquency, this would indicate that there is a
different relation between the number of risk (or promotive) factors and
delinquency for boys and girls. All these analyses were performed for the
number of risk factors, the number of promotive factors, and for the relative
risk level. Furthermore, we carried out the analyses for all the factors
together and for the separate domains of factors. The same analyses were
performed for serious delinquency.
102
Results
Delinquent behavior
In Wave 3 48.2 per cent of the boys and 38.1 per cent of the girls reported to
have committed at least one offence in the previous year, while 14.6 per cent
of the boys and 7.5 per cent of the girls reported at least one serious offence.
Risk and promotive factors
In Table 5.2 the risk and promotive factors for delinquency and serious
delinquency are presented. Empty cells in this table indicated no risk or
promotive effect. All the risk factors were summed to indicate the number of
risk factors for each adolescent. The same was done with the promotive
factors. For general delinquency, the risk domain of school and peers
consisted of only one factor. For serious delinquency, both the school and
peer domain consisted of only one risk and one promotive factor.
In Table 5.3 the descriptive statistics of the risk and promotive factor
counts are presented. Girls in general had more promotive factors than boys.
Boys showed to have more peer risk factors than girls.
Threshold for general and serious delinquency
Table 5.4 shows the results of the logistic regression of Wave 3 general and
serious delinquency on sex, the risk level, and the interaction between sex
and risk level, controlling for delinquency at Wave 2. These analyses were
carried out for the number of risk factors, for the number of promotive
factors, and for the relative risk level (number of risk factors minus the
number of promotive factors). The general number of risks (OR = 1.47), as
well as the number of specific risks of the negative life events (OR = 2.69),
father (OR = 1.62) and peer (OR = 3.39) domain were associated with an
increased risk of general delinquency. However, none of the levels of
promotive factors were significantly related to general delinquency.
Regarding the relative risk levels, the general risk level, as well as the
negative life events risks and the peer risks were related to delinquency (OR
= 1.15, 2.69, and 3.01, respectively). All associations were similar for boys
and girls, as indicated by non-significant sex by risk level interaction terms.
For serious delinquency we found an increased risk originating from the
number of risks in the individual domain (OR = 3.74). The number of
promotive factors (general nor in specific domains) was not associated with
serious delinquency. Finally, the relative level of risk in the individual
domain was positively associated (OR = 3.74). As was the case for general
delinquency, all associations again were similar for boys and girls, as
indicated by non-significant sex by risk level interaction terms.
103
Table 5.2
Significant risk and promotive factors for general delinquency and for serious delinquency
General delinquency
Serious delinquency
Factors
Risk
Promotive
Risk
Promotive
Individual
Extraversion
Agreeableness
Low
Conscientiousness
High
Neuroticism
Openness to experience
Impulse control difficulties
Prosocial behavior
Low
Empathic concern
Low
Low
Perspective taking
Low
High
Low
Self-concept clarity
Low
High
Low
Pubertal developmentc
Victim offence (theft,
Yes
Yes
violence, threat, sexual)d
Victimization (physical or
High
High
relational) aggression
Family member with police
Yes
Yes
contactsd
Family member contact with
justiced
Mother
Support mother
Low
Low
Power mother
High
Negative interaction mother
High
Low
High
Low
Conflict with mother
Higha
Lowa
Higha
Lowa
Parental knowledge mother
Low
High
Low
Adolescent disclosure mother
Lowa
Higha
Lowa
Parental solicitation mother
Parental control mother
Father
Support father
Low
High
Low
High
Power father
High
Negative interaction father
High
Low
High
Conflict with father
Higha
Lowa
Higha
Low
b
Parental knowledge father
Low
High
Low
Adolescent disclosure father
Lowa
Higha
Lowa
High
Parental solicitation father
Parental control father
Highb
School
School achievementc
Truancyc
High
Low
High
Low
Peers
Delinquent best friendd
Yes
Delinquent friendsc
Many
None
Many
None
104
Note. Cell entries indicate whether the factor constituted a risk or protection based on
logistic regressions with general and serious delinquency, respectively. High = highest
level; Low = lowest level; Yes = factor present; Many = many delinquent friends; None =
no delinquent friends;
Most factors were trichotimized based on the sex specific 25 th and 75th percentiles of the
variable distributions as cut-offs. Factors with a superscript d were already dichotomous,
and only represented risk dummies. Factors with a superscript c had cut-offs based on the
existing response categories. Regarding puberty, adolescents were at risk if their body
was mature already or when it had started changing more than a year ago. No body
changes were regarded as a promotive factor. For school achievement, it was classified a
risk when adolescents responded that they achieved worse to much worse than their peers
and as a promotive factor when they achieved better to much better. Regarding truancy it
was a risk if adolescents had stayed away from school at least once a month and a
promotive factor if they had never been truant in the last three months. With regards to
delinquent friends, it was a risk factor if adolescents responded about at least one of the
offences that most of their friends had committed them or about at least two of the
offences that some of their friends had committed them. It was a promotive factor if none
of their friends had committed any of the offences.
Superscript a indicates that the factor was too highly correlated with factors from the same
domain and therefore excluded from all risk and promotive analyses. Superscript b
indicates that the factor was too highly correlated with factors from other domains.
Therefore they were included in the analyses where separate domains were analyzed and
excluded from the risk and promotive analyses in which the overall risk level was
examined.
105
106
Table 5.3
Risk level means of delinquent and non-delinquent boys and girls
All
Non-delinquent
Delinquent
Boys
Girls
Boys
Girls
Boys
Girls
(n = 280)
(n = 215)
(n = 145)
(n = 133)
(n = 135)
(n = 82)
Risk
General (all factors)
3.44a (2.60)
3.51a (2.94)
2.47a (2.02)
2.58a (2.32)
4.45b (2.76)
5.05b (3.21)
a
a
a
a
b
Individual
1.03 (1.14)
1.14 (1.17)
0.81 (1.03)
0.91 (1.04)
1.26 (1.20)
1.54b (1.28)
0.61a (0.75)
0.57a (0.80)
0.36a (0.61)
0.40a (0.65)
0.86b (0.79)
0.88b (0.93)
a
a
a
a
b
Mother
0.84 (0.97)
0.85 (1.06)
0.62 (0.86)
0.64 (0.94)
1.06 (1.02)
1.20b (1.16)
a
a
a
a
b
Father
0.94 (1.00)
0.85 (1.14)
0.64 (0.81)
0.60 (0.99)
1.26 (1.09)
1.27b (1.26)
School
0.05a (0.21)
0.07a (0.26)
0.02a (0.14)
0.03a (0.17)
0.07ab (0.26)
0.15b (0.36)
a
b
a
a
b
Peers
0.87 (0.67)
0.67 (0.69)
0.64 (0.55)
0.50 (0.62)
1.11 (0.70)
0.97b (0.72)
Promotive
3.10a (2.01)
3.49b (2.22)
3.74a (2.03)
4.14a (2.14)
2.43b (1.75)
2.41b (1.91)
Individual
0.70a (0.85)
0.64a (0.78)
0.84a (0.93) 0.75ab (0.85)
0.55bc (0.73)
0.46c (0.59)
a
a
a
a
b
Mother
0.52 (0.67)
0.50 (0.67)
0.64 (0.70)
0.64 (0.73)
0.40 (0.61)
0.27b (0.50)
a
b
ac
b
a
Father
0.61 (0.73)
1.04 (0.96)
0.70 (0.75) 1.15 (0.98)
0.53 (0.70)
0.84bc (0.89)
School
0.65a (0.48)
0.64a (0.48)
0.74a (0.44)
0.77a (0.42)
0.56b (0.50)
0.53b (0.50)
a
a
a
a
b
Peers
0.58 (0.49)
0.64 (0.48)
0.86 (0.43)
0.79 (0.41)
0.39 (0.49)
0.39b (0.49)
Risk minus promotive
0.43a (3.94)
0.06a (4.56)
-1.16a (3.36) -1.53a (3.78)
2.09b (3.84)
2.70b (4.55)
a
a
a
a
b
Individual
0.33 (1.65)
0.51 (1.59)
-0.03 (1.16) -0.15 (1.48)
0.71 (1.61)
1.09b (1.60)
0.61a (0.75)
0.57a (0.80)
0.36a (0.61)
0.40a (0.65)
0.86b (0.79)
0.88b (0.93)
a
a
a
a
b
Mother
0.32 (1.36)
0.36 (1.47)
-0.01 (1.29) 0.01 (1.40)
0.66 (1.35)
0.94b (1.42)
Father
0.32a (1.36)
-0.19b (1.72)
-0.07a (1.20) -0.55b (1.56)
0.73c (1.39)
0.43ac (1.81)
School
-0.61a (0.58)
-0.57a (0.62)
-0.72a (0.49) -0.74a (0.50)
-0.48b (0.63)
-0.28b (0.71)
a
b
a
a
b
Peers
0.17 (0.97)
-0.06 (0.98)
-0.22 (0.73) -0.34 (0.86)
0.59 (1.02)
0.40b (1.00)
Note. Means of total group of boys and girls are compared with t-tests. Means of non-delinquent and delinquent boys and girls are
compared with ANOVA’s. Within each domain, for both the total group of boys and girls, as for the non-delinquent and delinquent
boys and girls, means with similar superscripts are equal to each other. In between brackets are the standard deviations.
Discussion
This study examined the threshold hypothesis, which states that girls have a
higher risk threshold for delinquency than boys. Accumulated risk levels in
general and in several of the risk domains under study (individual, father,
peers) were related to general delinquency. However, associations were not
different for boys and girls, thus disconfirming the threshold hypothesis.
Serious delinquency was only related to the risk level in the individual
domain with, again, no differences in associations between boys and girls.
Roughly the same results were found when levels of risk were corrected for
levels of promotive factors, with the exception that the risk level in the
father domain was not related anymore to delinquency.
The fact that despite careful analyses no differences in risk-delinquency
associations for boys and girls were found, thereby refuting the proposed
threshold hypothesis, suggests that other mechanisms might account for the
always found sex difference in levels of delinquency. An explanation more
consistently found is that girls are less exposed to risk factors than boys. For
example, girls seem to be less exposed to risk in the peer domain (Moffitt et
al., 2001; Murray, 2002; Wong et al., submitted), which was confirmed in
the present study. So although girls and boys have the same risk threshold
for delinquency, girls may to be less likely than boys to reach this threshold,
and therefore they are less likely delinquent.
Regarding the dose-response relationship, there were some interesting
findings. We found that the dose-response relationship for general
delinquency existed in other domains than for serious delinquency. General
delinquency was affected by the general risk level and the risk level in the
domains of negative life events, father and peers, whereas serious
delinquency was only affected by the risk level in the individual domain.
These results imply that risk factors from the environment increase the risk
for delinquency in general, but individual risk factors (such as low empathy,
low perspective taking and an unclear self-concept) increase the risk for
serious delinquency. Results on serious delinquency seem to be in line with
Moffitt’s taxonomy (1993) in which she states that seriously delinquents
(defined as life course persistent offenders) become delinquent because they
have neurological deficits manifested as problems in the individual domain,
such as low self-control, hyperactivity and difficult temperament.
The difference between our findings and previous investigations, where
girls were found to have a higher threshold for delinquency than boys
(Alemagno et al., 2006; Van der Laan & Van der Schans, 2010), might be
explained by the fact that previous studies examined a population that was
already arrested or incarcerated for committing an offence. It could very
well be that girls do not have a higher threshold for delinquency than boys,
but that girls do have a higher threshold for being arrested. Because of the
high prevalence of problems among arrested and convicted girls (e.g.
Chamberlein & Reid, 1994; Emeka & Sorensen, 2009), it seems like girls
107
108
Table 5.4
Logistic regression of general and serious delinquency on sex, risk level and the sex x risk level interaction
Risk
Promotive
B
SE
Odds ratio
B
SE
Odds ratio
(95% CI)
(95% CI)
General delinquency
Sex
0.38
0.22 1.47 (0.95 – 2.26)
0.27
0.22
1.29 (0.84 – 2.00)
Risk level
0.31* 0.14 1.37 (1.12 – 1.89)
-0.38*
0.18
0.68 (0.48 – 0.98)
Sex × Risk level
-0.04
0.09 0.96 (0.81 – 1.14)
0.06
0.12
1.06 (0.84 – 1.36)
Individual
Sex
0.30
0.22 1.35 (0.89 – 2.06)
0.28
0.22
1.32 (0.87 – 2.01)
Risk level
0.18
0.29 1.20 (0.68 – 2.12)
-0.39
0.41
0.68 (0.30 – 1.51)
Sex × Risk level
0.10
0.19 1.10 (0.76 – 1.60)
0.05
0.29
1.05 (0.60 – 1.84)
a
Sex
0.30
0.22 1.35 (0.88 – 2.09)
Risk level
0.99* 0.47 2.67 (1.08 – 6.70)
Sex × Risk level
-0.28
0.30 0.75 (0.42 – 1.34)
Mother
Sex
0.29
0.21 1.34 (0.88 – 2.03)
0.31
0.22
1.36 (0.89 – 2.07)
Risk level
0.41
0.33 1.50 (0.78 – 2.89)
-0.46
0.51
0.63 (0.24 – 1.71)
Sex × Risk level
-0.03
0.21 0.97 (0.64 – 1.46)
-0.01
0.35
0.99 (0.50 – 1.98)
Father
Sex
0.28
0.22 1.32 (0.86 – 2.03)
0.21
0.22
1.24 (0.81 – 1.90)
Risk level
0.77* 0.35 2.15 (1.09 – 4.25)
-0.44
0.42
0.65 (0.28 – 1.49)
Sex × Risk level
-0.17
0.21 0.83 (0.54 – 1.26)
0.19
0.26
1.21 (0.73 – 2.02)
Peers
b
Sex
0.18
0.23 1.19 (0.76 –1.87)
Risk level
1.22* 0.55 3.39 (1.16 – 9.91)
Sex × Risk level
-0.23
0.35 0.79 (0.40 – 1.58)
0.22
0.10
0.06
0.22
0.21
0.14
0.22
0.47
0.30
0.22
0.24
0.16
0.22
0.25
0.15
0.22
0.38
0.24
0.32
0.16
0.07
0.30
0.99*
-0.28
0.32
0.33
-0.02
0.17
0.57*
-0.19
0.18
1.10*
-0.27
1.20 (0.79 – 1.83)
3.01 (1.43 – 6.31)
0.77 (0.48 – 1.23)
1.19 (0.87 – 1.82)
1.77 (1.09 – 2.89)
0.83 (0.62 – 1.12)
1.37 (0.90 – 2.09)
1.39 (0.86 – 2.23)
0.98 (0.72 – 1.32)
1.35 (0.88 – 2.09)
2.67 (1.08 – 6.70)
0.75 (0.42 – 1.34)
1.38 (0.90 – 2.11)
1.17 (0.78 – 1.75)
1.07 (0.81 – 1.42)
1.39 (0.90 – 2.16)
1.28 (1.06 – 1.55)
0.96 (0.85 – 1.09)
SE
Odds ratio
(95% CI)
0.33
0.25*
-0.04
B
109
SE
Risk
Odds ratio
(95% CI)
B
Promotive
SE
Odds ratio
(95% CI)
B
2.03 (0.87 – 4.37)
1.19 (0.57 – 2.45)
1.12 (0.67 – 1.86)
1.60 (0.76 – 3.39)
1.27 (0.56 – 2.89)
1.09 (0.61 – 1.95)
1.38 (0.62 – 3.07)
2.57 (0.77 – 8.57)
0.81 (0.36 – 1.84)
1.48 (0.71 – 3.06)
3.74 (1.05 – 13.35)
0.52 (0.21 – 2.30)
2.27 (0.85 – 6.05)
1.27 (0.92 – 1.74)
1.02 (0.82 – 1.26)
SE
Odds ratio
(95% CI)
Serious delinquency
Sex
0.69
0.45 2.00 (0.82 – 4.86)
1.21
0.62
3.37 (0.99 – 11.41)
0.82
0.50
Risk level
0.35
0.21 1.42 (0.94 – 2.14)
-0.05
0.49
0.95 (0.36 – 2.50)
0.24
0.17
Sex × Risk level
-0.03
0.14 0.97 (0.74 – 1.28)
-0.45
0.39
0.64 (0.30 – 1.38)
0.02
0.11
Individual
a
Sex
0.39
0.37 1.48 (0.71 – 3.06)
0.39
0.37
Risk level
1.32* 0.65 3.74 (1.05 – 13.35)
1.32* 0.65
Sex × Risk level
-0.66
0.47 0.52 (0.21 – 1.30)
-0.66
0.47
a
Sex
0.32
0.41 1.38 (0.62 – 3.07)
0.32
0.41
Risk level
0.94
0.61 2.57 (0.77 – 8.57)
0.94
0.61
Sex × Risk level
-0.21
0.42 0.81 (0.36 – 1.84)
-0.21
0.42
b
Mother
Sex
0.57
0.39 1.77 (0.82 –3.79)
0.47
0.38
Risk level
-0.09
0.52 0.92 (0.33 – 2.53)
0.24
0.42
Sex × Risk level
0.27
0.36 1.31 (0.65 – 2.66)
0.08
0.30
Father
Sex
0.67
0.41 1.95 (0.88 – 4.35)
0.71
0.44
2.03 (0.85 – 4.84)
0.71
0.41
Risk level
0.29
0.51 1.33 (0.49 – 3.61)
0.08
0.84
1.08 (0.21 –5.55)
0.17
0.37
Sex × Risk level
0.09
0.35 1.09 (0.55 – 2.16)
-0.43
0.65
0.65 (0.18 – 2.33)
0.12
0.26
Note. In all the analyses we controlled for delinquency at Wave 2. * p < .05.
Superscript a: there were no promotive factors in this domain. Superscript b: there was only 1 promotive factor in this domain.
B
are not arrested if they do not have many problems. The threshold of risks
may be even bigger for being convicted. Previous research indeed has
shown that juvenile females are relatively less often arrested than juvenile
males for crimes (Sealock & Simpson, 1998). Also, if they are arrested, they
are relatively less often prosecuted compared to males (Jeffries, Fletcher, &
Newbold, 2003; Spohn & Beichner, 2000; Williams, 1999). It would be
interesting to examine in a future study if there indeed are different
thresholds for being arrested or convicted between boys and girls.
Our results do not necessarily have to implicate that the threshold
hypothesis is not true. It might be, for instance, that girls do have to exceed
a different risk level than boys for delinquency, but not at all ages. The
sample in our study was 13 to 15 years old when they were examined and as
Moffitt et al. (2001) suggest, in this age period delinquent behavior might be
sort of normative, also for girls. It may be that girls in this period, like boys,
are therefore more liable for delinquency and as a result do not need such a
big push anymore in terms of risk factors. Also, the peak age of delinquent
behavior is earlier for girls than for boys (15 and 16 years, respectively)
(Slotboom et al., 2011). This indicates that delinquency is relatively seen
even more common for girls than for boys at the age of our sample. Future
research should examine the threshold hypothesis at different ages,
preferably before, during, and after adolescence.
Another possibility is that girls do have to surmount a higher threshold
for serious delinquency, but that there were not enough seriously delinquent
boys and girls included in our study. Because of the low prevalence of
serious delinquency, especially among girls, self-reported data may not be
truly suitable to examine risk for serious delinquency. Therefore, most
studies that intend to investigate these delinquents analyzed adjudicated or
incarcerated samples. However, with these samples there is the problem that
one cannot disentangle the threshold for arrest from the threshold for
delinquency.
In this study we considered promotive factors next to risk factors to
specify the risk level. Although our results did not find large differences in
results regarding the number of risk factors alone or regarding the relative
risk level, the results in the father domain did differ. This indicates that it is
important to take into account promotive factors.
The current study was limited in some ways. A relevant limitation of
this study is that in the preliminary analyses, we determined whether a factor
was a risk (or promotive) factor for delinquency by analyzing the entire
sample and not by distinguishing between girls and boys. However,
previous studies (Wong et al., 2010) have shown that girls and boys share
many risk and promotive factors, but also have some different risk and
promotive factors. Also, we regarded each risk or promotive factor as
equally important, whereas it is possible that some factors are more strongly
related to delinquency than others. We have chosen not to make a distinction
110
between risk factors for boys and girls or – when aggregating risk factors
within domains - between less and more important factors, as no a priori
assumptions could be made to guide such choices. As studies on the
threshold hypothesis are very limited, we decided to deal with the issues of
promotive factors, overlapping effects of factors and different risk domains
first. It would be interesting for future studies to additionally consider
different risk factors for boys and girls and/or weight them in their study on
the threshold hypothesis.
Furthermore, the study was limited by the sample size. If the sample
would have been bigger, there would have been more power to carry out the
complicated analyses of this study.
To conclude, this study showed that there is a clear dose-response
relationship between the number of risks – overall and in several different
domains – and general as well as serious delinquency. However, we could
not demonstrate that girls need a higher number of risks than boys for
(serious) delinquency. Thus, for now, the threshold hypothesis does not
seem to provide an adequate explanation for the always demonstrated sex
differences in levels of delinquency.
111
Appendix
Delinquency items
General delinquency (included serious delinquency)
Destroying means of transport
Destroying somebody’s house
Destroying the school building
Destroying property
Shoplifting less than five euro’s
Shoplifting more than five euro’s
Theft at school less than five euro’s
Theft at school more than five euro’s
Theft at home less than five euro’s
Theft at home more than five euro’s
Other theft
Threatening somebody to scare somebody or to get
something
Beating up without injury
Soft drug selling
Hard drug selling
Party drug selling
112
Serious delinquency
Arson
Stealing a bike
Theft from/out of car
Burglary
Handling stolen goods
Robbery
Hurting with weapon
Beating up with injury
Sexual offence
113
Chapter 6
Sex and age differences in the risk
threshold for delinquency
Wong, T. M. L., Loeber, R., Slotboom, A., Bijleveld, C. C. J. H., & Koot, J. M.
(submitted). Sex and age differences in the risk threshold for delinquency.
115
This study examines sex differences in the risk threshold for adolescent
delinquency. Analyses were based on longitudinal data from the Pittsburgh
Youth Study (n = 503) and the Pittsburgh Girls Study (n = 856). The study
identified risk factors, promotive factors, and accumulated levels of risks as
predictors of delinquency and nondelinquency, respectively. The risk
thresholds for boys and girls were established at two developmental stages
(late childhood: ages 10-12 years, and adolescence: ages 13-16 years) and
compared between boys and girls. Sex similarities as well as differences
existed in risk and promotive factors for delinquency. ROC analyses
revealed only small sex differences in delinquency thresholds, that varied by
age. Accumulative risk level had a linear relationship with boys’
delinquency and a quadratic relationship with girls’ delinquency, indicating
stronger effects for girls at higher levels of risk.
116
Introduction
Many girls involved in the juvenile justice system – those who are arrested,
adjudicated or incarcerated – have been exposed to trauma or abuse, have
mental health as well as academic problems, and come from multi-problem
families (Chamberlain & Moore, 2002; Kataoka et al., 2001; Lederman,
Dakof, Larrea, & Li, 2004; Slotboom, Wong, Swier, & Van der Broek,
2011). Compared to arrested, adjudicated, or incarcerated boys, girls in the
juvenile justice system have more problems and are exposed more to known
risk factors (Belknap & Holsinger, 2006; Emeka & Sorensen, 2009; Gavazzi,
Yarchek, & Chesney-Lind, 2006; Gover, 2004; Johansson & KempfLeonard, 2009). This has been interpreted as delinquent girls having a more
problematic background than delinquent boys, which has also been
rephrased as the ‘threshold’ hypothesis, i.e. that girls pass a higher critical
‘risk level’ in order to become delinquent. This hypothesis was initially
defined for antisocial personality disorder (Cloninger & Gottesman, 1987)
and later expanded to other developmental disorders (Eme, 1992).
A threshold has been defined as the point that must be exceeded to
begin producing a given effect or result (www.thefreedictionary.com).
Thresholds are encountered in many areas of (social) science and generally
denote a critical value, under which a certain effect is not present and above
which it is, such as the absolute hearing threshold in medicine, or the
extinction theshold in ecology. In the manner in which the ‘threshold’hypothesis has been phrased in criminology, it denotes the ‘risk level’ above
which the probability to be delinquent is larger than the probability not to be
delinquent.
This ‘risk level’ that defines the risk threshold can, however, be
operationalized in two ways. Firstly, it can be operationalized as the severity
or level of a single risk factor: having a problematic relationship with parents
is a risk factor for delinquency, and only those youth with a very problematic
parent-child relationship have a risk level that is high enough to pass the
threshold to offend. The other way of operationalizing risk level is derived
from the cumulative risk approach (Rutter, 1979; Sameroff, Seifer, Barocas,
Zax & Greenspan, 1987) and defines the risk level as the number of risk
factors. Thus, according to this operationalization the more risk factors
someone experiences, the more likely he or she is to be delinquent. There is
evidence for such a dose-response relationship between the number of risk
factors and the likelihood of delinquency for boys and girls (Johansson &
Kempf-Leonard, 2009; Loeber, Slot & Stouthamer-Loeber, 2008; Van der
Laan & Van der Schans, 2010; Wong et al., submitted).
A key issue, that is unresolved in the literature and that is the focus of
this study, is whether there are sex differences in the risk threshold for
delinquency: differences between boys and girls in such a threshold for
delinquency, while often posited, have hardly been studied empirically..
117
Sex difference in risk thresholds
Alemagno, Shaffer-King, and Hammel (2006) examined the number of risk
factors of 250 detained boys and girls and found that incarcerated girls were
exposed to more risk factors than their male counterparts. Van der Laan and
Van der Schans (2010) showed, using a similar analytical strategy, that
arrested girls were exposed to more risk factors in the family domain than
arrested boys. Although the results of these studies concur with the
differential risk threshold hypothesis, they do not show that such a
differential threshold exists for delinquency, since all studies investigated
samples of adjudicated or incarcerated juveniles. Given that girls and women
are often treated differently in the juvenile justice system, the threshold for
delinquency cannot be separated from the threshold to be arrested,
prosecuted or convicted (e.g., Daly, 1994). Thus it is problematic to attribute
sex differences in the number of risk factors in officially delinquent samples
to the threshold for delinquency. This may also explain seemingly
incompatible findings, such as that arrested boys have in fact a higher
number of risky lifestyle factors compared to arrested girls (Van der Laan &
Van der Schans, 2010). Self-reported delinquency studies tend not to have
the confounding effect of justice processing.
Wong et al. (submitted) investigated sex differences in the delinquency
threshold using self-reported data of a Dutch population-based sample, and
did not find support for a sex-related threshold. The use of a comparison
group is necessary, as without such a group it is impossible to determine
whether delinquent girls have a higher risk level than delinquent boys or vice
versa. In contrast to the previously mentioned studies, the authors included
such a comparison group of nondelinquents, and in addition to risk factors,
the authors also examined the extent to which promotive factors influenced
the risk of later delinquency. Promotive factors are those factors associated
with a decreased probability of delinquency (Sameroff, Bartko, Baldwin,
Baldwin, & Seifer, 1998; Stouthamer-Loeber, Loeber, Wei, Farrington, &
Wikström, 2002). Since promotive factors can neutralize risks (StouthamerLoeber et al., 2002; Van der Laan & Blom, 2006), ignoring these factors
might result in overstating the importance of risk factors and might make it
impossible to assess any accurate threshold effect.
Although the study by Wong et al. (submitted) had fewer limitations
than previous studies, the authors did not investigate the threshold as such as
they compared risk levels of delinquents with those of nondelinquents. The
present study will improve upon previous research firstly by actually
assessing the threshold itself, i.e. identifying the exact cut off value, for boys
and girls. Secondly, this study will improve on previous studies by
investigating whether the threshold varies with age and/or sex. Boys’ and
girls’ involvement in delinquency changes with age, and criminal careers
develop differently for boys and girls (Junger-Tas, Haen-Marshall, &
Ribeaud, 2003; Wong, Blom & Van der Laan, in press). Girls’ delinquency
118
tends to peak earlier than that of boys, i.e. at age 15 versus at age 16 (JungerTas et al., 2003; Slotboom, Wong, Swier, & Van der Broek, 2011). It
remains to be seen whether delinquency thresholds vary with age for each
sex. As Moffitt (1993) suggested, during puberty, it is almost normative to
show some delinquent behavior. Thirdly, this study will add to previous
research by incorporating sex-shared as well as sex-specific risk factors for
delinquency (Wong, Slotboom, & Bijleveld, 2010; Zahn, 2009).
We will address the following research questions: 1) Is the age-crime
curve for girls lower than that of boys? 2) Which shared and sex-specific risk
and promotive factors measured in middle childhood (ages 7 to 9) and late
childhood (ages 10 to 12), respectively, predict self-reported delinquency in
late childhood (ages 10 to 12) and adolescence (ages 13 to 16)? 3) Are there
sex differences in exposure to risk and promotive factors? 4) Are there linear
or quadratic differences in the relationship between cumulative risk and
protective factor score and delinquency for each sex? 5) Are there
differences by sex and age in the optimal cumulative threshold to predict
delinquency?
The questions are addressed using data from the Pittsburgh Youth Study
(PYS) and the Pittsburgh Girls Study (PGS) using self-reported delinquency
as outcomes at late childhood and adolescence. The studies contain a broad
array of risk and promotive factors known to predict delinquency in previous
studies (e.g., Hoeve et al., 2009; Hubbard & Pratt 2002; Lipsey & Derzon,
1998; Maguin & Loeber, 1996; Pratt & Cullen, 2005; Simourd & Andrews,
1994; Wong et al., 2010; Zahn, 2009). These include individual (problem)
factors (i.e., birth problems, early disruptive behavior disorder, callous
unemotional behavior, anxiety, early puberty), family factors (i.e., poor
education of parents, single parent household, physical punishment,
communication with parents, positive parenting, supervision, parent-child
relationship), school factors (i.e., truancy, school motivation, school
achievement), peer delinquency, and neighborhood problems.
Methods
Sample
The PYS is a longitudinal study that started in 1987 (Loeber, Farrington,
Stouthamer-Loeber, & White, 2008), consisting of three samples of boys
who were in grades one, four, and seven, respectively, at the start of the
study. Boys who attended public schools in Pittsburgh participated in the
study. In the initial screening assessment, information about the boys’
antisocial behavior was collected through the boys themselves, the
caretakers, and their teachers. On the basis of this information, a risk score
was calculated and all of the boys with the highest scores on antisocial
behavior (n = c. 250, for every sample) were selected for follow-up, while a
random sample of the remaining boys (N = c. 250) were also included in the
follow-ups. Only boys from the youngest sample (n = 503) were included in
119
the present study. In the first four years of the follow-ups, interviews were
conducted by trained interviewers every half year with the boys and one or
both caretakers. In the same period, one of the boys’ teachers was asked to
rate the boys’ behavior. Subsequently, interviews were held every year. For
the current analyses, information about grades was transformed in agespecific data.
The PGS is also a longitudinal study, but is based on a stratified,
random sample from all households in Pittsburgh with a girl between the age
of 5 and 8 (Keenan et al., 2010). Disadvantaged neighborhoods were
oversampled. The final sample consists of 2,451 families. To make the
samples of PGS and PYS youth comparable, the current study included only
girls aged 7 or 8 at the initial assessment, who attended public schools at the
first assessments in 2000 (n = 856). Follow-ups in the PGS consisted of
yearly interviews with the girls, their caretaker and teacher ratings.
Measurements
To achieve comparability between the sexes, only measurements were
included that were comparable across the PYS and the PGS.
Delinquency. Delinquency was measured at ages 11-16 through the 40-item
Self-Reported Delinquency Scale (SRD; Loeber, Farrington, StouthamerLoeber, & Van Kammen, 1998) which was based on an adaptation of the the
National Youth Survey (Elliott, Huizinga, & Ageton, 1985). For each of the
offenses, respondents indicated whether they had committed a delinquent
act, and if so, how often in the previous year. For this study we focused on
moderate to serious delinquency (see details in Loeber et al., 1998), which
included breaking-and-entering, stealing things worth more than 5 dollars,
purse snatching, stealing from a car, dealing in stolen goods, joyriding,
vehicle theft, attacking with intent to injure, forcible robbery, and gang
fighting. All offences were summed and dichotomized into 0 (no offence
committed – nondelinquent) and 1 (1 or more offences committed –
delinquent). At age 11 the dealing in stolen goods item was accidentally not
assessed in the PGS, so we did not include this item in the delinquency
construct for both boys and girls.
The SRD was judged to be too difficult to understand for the youngest
respondents. For that reason, the Self-Reported Antisocial Behavior Scale
(SRA) instead of the SRD was administered at age 10. Since boys were
selected in the first wave by grade and therefore had different ages, and since
the switch from SRA to SRD was made in one phase for all boys, some of
the 10-year-old boys filled out the SRA en some the SRD. For girls, the
switch was made after the age of 10 and therefore all 10-year-old girls
reported on the SRA. The SRA consisted of 27 items of delinquent behavior
that were appropriate to younger children (Loeber et al., 1989). For the
current study, only those items were used to construct the delinquency scale
120
that were comparable to the SRD items used in the construct: theft from
building, theft from a car, and purse snatching.
After the creation of the moderate and serious delinquency constructs
for each age, we prepared summary constructs for age blocks in late
childhood (ages 10 to 12) and adolescence (ages 13 to 16), contrasting
nondelinquents with delinquents (1 or more offences committed at this age).
Risk and promotive factors. Table 6.1 lists all constructs used in this study
based oncomparable measures in the PYS and PGS. For most factors, we
created two age blocks: for late childhood and adolescence. Birth problems
and early disruptive behavior disorder were only assessed in the first
assessment and regarding early pubertal development only the measurements
prior to the delinquency age blocks were included (i.e. age 9 and age 12). In
the PGS, no information about single parent households was available at the
age of 7, so the late childhood age block regarding single parent households
only contained age 8 and 9. Truancy was only measured at age 11 and 12, so
the late childhood age block was not created.
In creating the constructs from reported waves, missing constructs were
coded as missing if more than 33% was missing. If fewer were missing, the
mean of the available responses was substituted for the missing data. In
creating the age blocks, only the non-missing ages were used to calculate the
age blocks for a respondent. The age block was set to be missing if the
construct was missing at all ages.
To identify the risk versus promotive effect of the factors we used the
same method as Stouthamer-Loeber et al. (1993). All age blocks were
trichotimized into a promotive, a neutral and a risk component using the sexspecific 25th and 75th percentiles of the age block distributions as cut-offs.
The age blocks were recoded into two variables: a risk variable and a
promotive variable. The reference category in each variable was the neutral
component (the 26th to the 74th percentile of the distribution). The exceptions
were birth problems, early disruptive behavior disorder, poor education of
the parents, and child’s truancy, because these were inherently dichotomous.
Another exception was the age block for single parent households. In this
case, it was more appropriate to trichotimize according to the number of
years the household consisted of a single parent (i.e. risk: single parent in all
years of age block; promotive: both parents in all years of age block; neutral
see Table 6.1).
121
122
Psychopathy
Screening
Device
Child Behavioral
Checklist
(CBCL)
Callous unemotional
behavior7
Parent
Parent
Assessed by
Parent
7-9, 10-12
First
assessment
Ages
First
assessment
32 items
(alpha
from .90 to
.93)
ADHD: 27
items;
ODD: 18
items; CD:
18 items
6 items
(alpha
from .56 to
.69)
ADHD:
14 items;
ODD: 8
items; CD:
12 items
Reliability
Boys
Girls
15 items
7 items
Highest 25%
At least one
of the
following
disorders:
ADHD,
ODD, CD
Risk
Any pre- or
perinatal birth
problem
Lowest
25%
NA
Promotive
NA
Due to the time of the assessment, the diagnoses of ADHD, ODD, and CD in the PYS were based on the DSM-III-R, whereas the diagnoses
in the PGS were based on the DSM-IV. To make diagnoses comparable, we only included those symptoms that were assessed in both
studies. For ADHD, the age of onset, that is usually part of the diagnosis, could not be taken into account since it was not assessed in the
PGS. To reach the diagnosis of ADHD, boys and girls had to have 9 symptoms or more. For the diagnosis of CD, 3 or more symptoms were
required, and for the diagnosis of ODD, 4 or more symptoms.
7
In the PYS a construct is created that measures psychopathic features in childhood, assessed by the CBCL. Examples of items are ‘lying or
cheating’ ‘sudden changes in mood or feelings’, and ‘behaving irresponsibly’. In the PGS, items from the PSD were used to create a similar
construct for girls. The following items are included: concerned about school or tasks, keeps promises, feels bad about doing wrong,
concerned about others’ feelings, shows feelings and emotions, keeps the same friends.
6
Child Symptom
Inventory (CSI)
Diagnostic
Interview
Schedule for
Children (DISC)
Instruments
Boys
Girls
Birth and
Pre and Perinatal
developmental
Risk Factors
history
Early disruptive
behavior disorder6
Constructs
Birth problems
Table 6.1
Constructs used in this study
123
Girls
Scared
Highest degree
of education
Petersen
Pubertal
Development
Scale (PPDS)
How many
caretakers?
Parent-Child
Conflict Tactics
Scale (CTSPC)
Highest degree of
education
Petersen Pubertal
Development
Scale (PPDS)
How many
caretakers?
Discipline
Poor education of
parents
Early pubertal
development
Single parent
household
Physical punishment
of both parents
Instruments
Boys
CBCL
Constructs
Anxiety
Child
Parent
Child
Parent
Assessed by
Parent
7-9, 10-12
8-9, 10-12
9, 12
7-9, 10-12
Ages
7-9, 10-12
1 item
1 item
5 items
(alpha
from .56 to
.75)
1 item
1 item
1 item
5 items
(alpha
from .50 to
.69)
1 item
Reliability
Boys
Girls
7 items
29 items
(alpha
(alpha
from .54 to from .90 to
.61)
.92)
Highest 25%
Living with
one parent at
all ages
Highest 25%
No diploma
or a General
Education
Diploma
(GED) for
both parents
at all ages
Risk
Highest 25%
No
physical
punishment
at all ages
Living
with both
parents at
all ages
Lowest
25%
NA
Promotive
Lowest
25%
124
SIS
Parent-Child
Relationship
Survey (PCRS)
SRD
Teacher
Works not hard
compared to peers
SIS
Parent-Child
Relationship
Survey (PCRS)
SRD
Works not hard
compared to peers
Low supervision
Bad quality
relationship with
primary caretaker
Truancy
Low school
motivation
Child
Child
Child
Child
Parent Practices
Scale (PPS)
Child
Assessed by
Parent Practices
Scale (PPS)
Girls
Low positive
parenting of both
parents
Instruments
Supervision and
Involvement
Scale (SIS)
Boys
Low communication Supervision and
about activities with Involvement
Scale (SIS)
both parents
Constructs
7-9, 10-12
11-12
7-9, 10-12
7-9, 10-12
7-9, 10-12
7-9, 10-12
Ages
1 item
1 item
16 items
(alpha
from .83 to
.91)
4 items
(alpha
from .54 to
70)
14 items
(alpha
from .71 to
.97)
10 items
(alpha
from .64 to
.84)
1 item
1 item
16 items
(alpha = )
4 items
(alpha
from .45 to
.61 )
14 items
(alpha
from .71 to
.97)
10 items
(alpha
from .52 to
.87)
Reliability
Boys
Girls
Highest 25%
Truant at both
ages
Highest 25%
Highest 25%
Highest 25%
Highest 25%
Risk
Lowest
25%
NA
Lowest
25%
Lowest
25%
Lowest
25%
Lowest
25%
Promotive
125
Parent
Your
Neighborhood
Your
Neighborhood
Neighborhood
problems
7-9, 10-12
7-9, 10-12
Ages
7-9, 10-12
17 items
(alpha
from .93 to
.96)
17 items
(alpha
from .94 to
.96)
Reliability
Boys
Girls
9 items
9 items
(alpha
(alpha = )
from .64 to
.71)
5 items
7-9: 5
(alpha
items
from .68 to (alpha
.84)
from .78 to
.80); 1012: 6 items
(alpha
from .75 to
.78)
Highest 25%
Highest 25%
Risk
Highest 25%
Lowest
25%
Lowest
25%
Promotive
Lowest
25%
For 7-to-9-year-olds, exactly the same offences were included (vandalism, shoplifting, stealing at school, stealing from building, violence
against adult) in the PYS and the PGS. For 10-to-12-year-olds, the peer delinquency scale was similar in the PGS, but included more serious
offences in the PYS. Therefore, we only took those offences of the PYS into account that were comparable to those of the PGS (and which
are also similar to the offences considered at earlier ages), i.e. vandalism, stolen something up to $100, stealing from building, and hitting
someone with intent to hurt. We corrected for the number of possible items.
8
Child
Peer
Delinquency
Scale (PDS)
Assessed by
Parent and
teacher
Peer Delinquency
Scale (PDS)
Instruments
Boys
Girls
CBCL & TRF
CBCL & TRF
Peer delinquency8
Constructs
Low school
achievement
Analyses
First, we established which risk and promotive factors predicted delinquency
at late childhood and adolescence, respectively. These analyses were carried
out separately for boys and girls and separately for the two age periods. If a
factor predicted delinquency (p < 0.05), this was regarded as a risk effect; if
a factor predicted low or nondelinquency, this was regarded a promotive
effect. If both variables were related to delinquency, this was regarded both a
combined risk and a promotive effect. Some risk factors predicted
delinquency in boys and girls and were labeled shared risk factors. The same
applied to factors predicting nondelinquency in boys and girls and were
labeled shared promotive factors. Factors that were only related to
delinquency in either boys or girls were labeled sex-specific risk and
promotive factors. Odds Ratios were calculated for the risk and promotive
factors: an Odds Ratio larger than 1 with a p-value < .05 indicates that the
presence of the risk factor significantly increased the prediction of
delinquency, while an Odds Ratio smaller than 1 with a p< .05 indicates a
promotive factor that significantly predicted nondelinquency.
Next, we created three types of cumulative risk level indexes. The first
index consisted of the number of significant risk factors in the data set. A
second index indicated the number of significant promotive factors in the
data set. The third, called the combined risk index indicated the number of
significant risk factors minus the number of significant promotive factors.
Because the three risk indexes were created by taking into account shared
factors as well as sex-specific factors, each risk index consisted of slightly
different risk and promotive components for boys and girls.
Thresholds were studied at two levels. First, we studied whether the
distribution of the relationships between cumulative risk were similar for
boys and girls; for this we carried out a curve fitting analysis to see whether
cumulative risk indexes predicted delinquency in a linear or quadratic way
for boys and girls. If, for example, a quadratic function applied to one but
not the other sex, this indicated that the risk of future delinquency
accelerated faster for one sex compared to the other.
In a second set of analyses, we examined whether a threshold could be
empirically established by means of signal detection theory (Swets, 1964).
Receiver Operating Curves (ROC) were calculated with Area Under the
Curve (AUC) indicating how well a cumulative risk index predicted
delinquency. The analyses also allow the identification of optimal prediction
thresholds in which, for every possible cut-off, the trade-off between the
false negative and false positive rates is calculated. AUC values can range
from 0 (total inaccuracy) to 1 (perfect accuracy). A value of 0.5 indicates
that the model is not better than chance, a value between 0.5 and 0.75 is
regarded as fair, between 0.75 and 0.92 as good, between 0.92 and 0.97 as
very good and between 0.97 and 1 as excellent (McFall & Treat, 1999). The
Youden’s index, a function of sensitivity (number of true positives) and
126
specificity (number of true negatives), was used to identify the optimal cutoff point (Youden, 1950). The optimal cut-off is the value with the highest
combination of sensitivity and specificity. This cut-off point is the threshold
for delinquency. We carried out these analyses separately for late childhood
and adolescence and for boys and girls.
Results
Table 6.2 shows the descriptive results. The average number of measured
risk and promotive factors are presented for boys and girls in middle and late
childhood as well as the number of delinquents in late childhood and
adolescence. Girls appeared to have on average more promotive factors than
boys. No other sex differences were found regarding the average number of
measured risk and promotive factors. The prevalence of delinquency differed
by gender in both late childhood as well as in adolescence.
Table 6.2
Descriptive results
Boys
Girls
Average
Middle childhood
Average number of risk factors
Average number of promotive factors
3.43 (2.33)
2.92 (2.11)
3.29 (2.28)
3.00 (2.37)
3.34 (2.30)
2.97 (2.28)
Late childhood
Average number of risk factors
Average number of promotive factors*
3.41 (2.24)
2.95 (2.17)
3.19 (2.34)
3.42 (2.51)
3.27 (2.31)
3.24 (2.40)
% delinquent*
24.5%
9.7%
15.2%
Adolescence
% delinquent*
42.6%
21.2%
29.2%
Note. Standard deviations are in parentheses. With t-tests it was tested whether boys
and girls differed in number of risk and promotive factors. Crosstabs were used to test
the difference in delinquency prevalence.
* significantly different for boys and girls at p < .05
The first question we addressed was: Is the age-crime curve for girls
lower than that of boys? Figure 6.1 shows that at age 10 there was only a
small, although significant (3.6% vs. 1.8%; p<.05) sex difference in the
prevalence of moderate to serious delinquency, but at all other older ages the
prevalence of delinquency was higher for boys than girls (for all ages p<.01).
However, the peak age of the age-crime was the same for the two sexes (age
14).
127
Figure 6.1
Age crime curve for moderate to serious delinquency by sex
25%
% delinquents
20%
15%
10%
5%
0%
10
11
12
13
Age
% delinquent boys
14
15
16
% delinquent girls
The second question that we posed was: Which shared and sex-specific
risk and promotive factors measured in middle childhood (ages 7 to 9) and
late childhood (ages 10 to 12), respectively, predict self-reported
delinquency in late childhood (ages 10 to 12) and adolescence (ages 13 to
16) Table 6.3 shows the odds ratios of the risk and promotive factors for
boys and girls in the two age periods. An empty cell indicates that there is no
statistically significant risk (or promotive) effect of a given factor.
The results showed that delinquent behavior of boys and girls is related
to many different factors. As Table 6.3 shows, many risk and promotive
factors are shared by boys and girls, but some differences were found
between boys and girls, and between age periods as well.
Next we asked: Are there sex differences in exposure to risk and
promotive factors? Table 6.4 shows the average number of (significant) risk
factors and (significant) risk minus promotive factors for nondelinquent and
delinquent boys and girls during middle and late childhood. Delinquent boys
and girls averaged higher risk scores than nondelinquent boys and girls,
respectively. Furthermore, delinquent girls averaged a higher number of risk
factors than delinquent boys at each age period. When average of risk and
promotive factors were considered, delinquent girls compared to delinquent
boys scored higher at middle childhood only. At late childhood, average
exposure to risk and promotive factors was similar for of delinquent boys
and girls.
128
129
Table 6.3
Odds ratios of risk and promotive factors for delinquency at ages 10 to 12 and ages 13 to 16, by sex
Delinquency (10 to 12 years)
Delinquency (13 to 16 years)
Risk
Promotive
Risk
Promotive
Boys
Girls
Boys
Girls
Boys
Girls
Boys
Girls
Factors
Birth problems
1.82**
Disruptive behavior
3.22**
2.64**
Callous unemotional behavior
2.39**
3.21**
0.35**
0.18**
2.15** 3.16**
0.41** 0.35**
Anxiety
0.62*
2.28**
0.47**
Early pubertal development
Poor education of parents
3.37**
Single parent household
0.64*
0.62*
0.53**
Physical punishment of parents
0.37**
1.82*
0.64*
Positive parenting
1.88*
1.63*
0.55*
Supervision
1.83*
0.41*
2.44** 2.31**
0.41** 0.35**
Relationship with primary caretaker
2.20**
2.80**
0.20**
2.83** 2.86**
0.44** 0.29**
Truancy
4.18** 6.10**
School motivation
2.59**
2.32**
0.35**
0.44*
2.29** 2.13**
0.46** 0.28**
School achievement
1.93**
2.20**
0.48*
0.38*
1.93**
0.50** 0.42**
Peer delinquency
2.19**
2.57**
0.32**
0.12**
3.78** 4.96**
0.23** 0.18**
Neighborhood problems
1.97**
0.52*
1.60*
1.79**
0.54** 0.58*
* p < .05, ** p < .01
130
Delinquent
Difference
within boys
Non
delinquent
Girls
Delinquent
Difference
within girls
Sex difference
between delinquents
Middle childhood
Average Number of risk
1.15 (1.27)
1.95 (1.41)
2.19 (1.76)
3.82 (1.81)
t (468) = 5.66**
t (804) = 7.85** t (192) = 8.11**
factorsa
-0.89 (2.41)
0.65 (2.40)
0.61 (2.61)
2.91 (2.11)
t (468) = 4.96**
t (804) = 7.58** t (192) = 6.77**
minus promotive factorsa
Late childhood
1.13 (1.25)
2.32 (1.62)
1.63 (1.50)
3.39 (1.89)
t (444) = 8.83**
t (747) = 12.39** t (347) = 5.70**
factorsa
-1.39 (2.72)
0.92 (2.69)
-1.68 (3.18) 1.43 (2.61)
t (444) = 8.91**
t (747) = 11.35** t (347) = 1.80
minus promotive factors
Note. Standard deviations in parentheses. Means of nondelinquent and delinquent boys and of nondelinquent and delinquent girls are compared with T-tests
as well as those of delinquent boys and girls.
* p < .05, ** p < .01
Non
delinquent
Table 6.4
Means and standard deviations of risk levels for nondelinquent and delinquent boys and girl
Boys
The fourth question we asked was: Are there linear or quadratic
differences in the relationship between cumulative risk and protective factor
score and delinquency for each sex? Curve fitting analyses showed that for
both age periods positive linear relationships between the risk levels and
delinquency were found for boys (with R² of .07 and .15 respectively; other
relationships had a worse fit to the data), but positive quadratic relationships
for girls (with R² of .06 and .17 respectively, again other relationships had a
worse fit to the data; see the modeled relationships in Figures 6.2 and 6.3).
This indicates that, regardless of sex, the more risk factors boys and girls are
exposed to, the more likely they are to be delinquent. However, for boys the
increase in likelihood for delinquency is similar across risk levels, whereas
for girls the increase in likelihood is amplified at every next risk level. More
specifically, for boys every increase in the number of risk factors results in
5.2% more delinquents in late childhood and 7.3% more delinquents in
adolescence. For girls, this increase depends on the risk level. An increase
from 3 to 2 promotive factors (in middle and late childhood respectively), for
instance, leads to 0.6% more delinquents in late childhood and to 3.3% more
delinquent in adolescence, whereas an increase from 3 to 4 or more risk
factors (in middle and late childhood) results in 5.4% and 10.5% more
delinquents in late childhood and adolescence. Thus, for girls we see that the
effect of a one-step risk increase becomes ever stronger: the higher the risk
level, the larger the corresponding shift in delinquency of an increase in risk.
The final question concerned: Are there differences by sex and age in
the optimal cumulative threshold to predict delinquency? The results
regarding the predictive power of the combined risk levels on late childhood
delinquency for boys and girls are in Figure 6.2: girls have slightly higher
AUC values than boys (0.74 vs. 0.68). Furthermore, the optimal cut-off point
for girls is higher than for boys (1 vs. 0 risk factors) which indicates that
girls have a higher threshold for delinquency in late childhood than boys.
Next, adolescent delinquency was predicted from risk levels at the age
of 10 to 12 (see Figure 6.3). Girls had slightly higher AUC values (0.77 vs.
0.72), but boys had a higher optimal cut-off point than girls (1 vs. 0 risk
factors). Boys therefore have a higher threshold than girls to become
delinquent in adolescence. Thus, we see that there are no consistent
differences in the delinquency threshold for boys and girls: the thresholds
differ by age period. The differences are also small; however, as the
threshold is a group-value and not the average of a set of individual-level
values, we cannot test whether it differs significantly for boys and girls.
131
Figure 6.2
Combined risk levels (number of risk factors minus number of promotive factors) at the age of
7 to 9 predicting moderate to serious delinquency at age 10 to 12, for boys and girls.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
% delinquent boys
(AUC: 0.68)
Prevalence of boy
delinquency
Modeled relationship
Threshold
Promotive
risk factors
100%
% delinquent girls
(AUC: 0.74)
90%
80%
70%
Prevalence of girl
delinquency
60%
50%
40%
Threshold
30%
20%
10%
0%
Promotive
132
risk factors
Figure 6.3
Combined risk levels (number of risk factors minus number of promotive factors) at the age of
10 to 12 predicting moderate to serious delinquency at age 13 to 16, for boys and girls.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Threshold
Promotive
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
% delinquent boys
(AUC: 0.72)
Prevalence of boy
delinquency
risk factors
% delinquent girls
(AUC: 0.77)
Prevalence of girl
delinquency
Threshold
Promotive
risk factors
133
Discussion
This study examined whether boys and girls had different risk thresholds for
delinquency at two age periods (late childhood and adolescence). Using data
from the PYS and PGS studies, we first tested which factors (at ages 7 to 9
and 10 to 12) had a risk effect, a promotive effect, or both. Boys and girls
appeared to share many risk and promotive factors, but sex differences and
differences between age periods were found as well.
Not surprisingly, boys and girls who were delinquent appeared to have
higher risk levels than their nondelinquent counterparts. Within the
delinquents, girls on average had higher number of risk factors than boys
when only risk factors were considered. When promotive factors were taken
into account as well, girls compared to boys had on average a higher risk
levels in middle childhood. In late childhood, the risk level of delinquent
boys and girls was similar.
The relationship between the risk level and delinquency was linear for
boys, indicating that every extra risk factor resulted in a similar step-wise
increase regarding delinquency probability. For girls, however, this
relationship turned out to be non-linear, with the increase in the probability
of delinquency larger at the higher risk level ranges than in the lower part.
Thus, at low risk levels, an additional risk factor gives but a small increase in
the delinquency probability. However, at higher risk levels, one additional
risk factor augments this probability substantially. Due to this amplification,
delinquent girls would – even with a same delinquency threshold - have
higher average risk levels than boys. Therefore, previous studies that focused
on the average risk level for boys and girls found higher risk levels among
delinquent girls than among delinquent boys (Alemagno et al., 2006; Van
der Laan & Van der Schans, 2010.
While higher risk levels are associated with a stronger increase in
likelihood of delinquency in girls than in boys, this study has shown that
girls do not have a higher threshold for delinquency, and that any differences
in the threshold are small and fluctuate with age. All in all, in this study –
that was appropriately designed with a control group, and sex-specific risk as
well as promotive factors – no evidence for a sex-specific delinquency
threshold emerged.
The threshold hypothesis was examined using two complementary
approaches: curve fitting and ROC analyses. The curve estimation analyses
showed a linear association between risk level and delinquency for boys and
a curvilinear relationship for girls. The ROC analyses examined the location
of the threshold and did not show sex differences. While there appears to be
no different threshold as such, increases of the risk level beyond this
threshold impact differently on girls than on boys. That is, from the
threshold onwards, risks contribute more and more to the delinquency risk
for girls, but not for boys.
134
Differences with previous studies
Several explanations can be put forward for the fact that most previous
studies on the threshold had such different results than the present study.
These explanations regard differences between previous studies and the
present study regarding the sample, the definition of the threshold, and
regarding the operationalization of risk. With regard to sample differences,
previous studies mainly examined adjudicated or incarcerated samples. In
these samples the threshold for delinquency is confounded with the threshold
for criminal justice system involvement. The fact that our study showed that
the threshold for delinquency differs minimally for boys and girls, these
studies probably picked up on arrest, prosecution or incarceration thresholds.
Concerning differences in the definition of the threshold, previous
studies based their conclusions about sex different thresholds on risk levels
of delinquent boys and girls (Alemagno et al., 2006; Belknap & Holsinger,
2006; Emeke & Sorensen, 2009; Johansson & Kempf-Leonard, 2009; Van
der Laan & Van der Schans, 2010), whereas the current study identified the
location of the threshold. Because delinquent girls had on average higher
risk levels than boys and because delinquency is less prevalent in girls,
previous studies concluded that girls have a higher threshold for
delinquency. However, the threshold as such was not assessed.
Regarding the operationalization of risk, there are two main differences
between previous studies and the present study. First, previous studies did
not include promotive factors to measure risk. However, since the number of
promotive factors can buffer the influence of risk factors only (StouthamerLoeber et al., 2002; Van der Laan & Blom, 2006), it is inadequate to
examine only risk factors. To see how the results would differ if we would
have considered risk factors only, the analyses of the present study were
carried out as well for the risk index that only considered the number of risk
factors9. Just like in previous studies (Alemagno et al., 2006; Van der Laan,
& Van der Schans, 2010), we found a higher threshold for girls when we
focused solely on risk factors, for both age periods. Slightly better AUC
values showed, however, that models that included both risk factors and
promotive factors were more adequate than models that considered risk
factors only. Not including promotive factors can lead to overestimation of
the risk and therefore of the threshold. This indeed turned out to be the case
for girls.
Second, the present study included sex-specific factors while other
studies did not (see Moffitt, Caspi, Rutter, & Silva, 2001; Junger-Tas,
Ribeaud, & Cruyff, 2004). Again, for the sake of comparison, the analyses of
the present study were also carried out with models that only considered
9
For reasons of simplicity, these results are not presented here, but are available
from the author.
135
shared factors10. Models that considered sex-specific factors as well resulted
in better predictions of delinquency at puberty for girls than models with
solely shared factors. In these latter models, that were utilized in previous
studies, girls’ risks are underestimated and their risk threshold cannot be
examined properly.
Our study showed that girls and boys do not differ in their delinquency
‘threshold’, i.e. the risk level beyond which the probability to be delinquent
is greater than the probability to be not delinquent. It is likely that the
threshold that was picked up in previous studies among criminal justice
samples may actually have been a criminal justice-involvement threshold.
Difference in the average risk levels of delinquent boys and girls are
generated by the increasing impact of risk factors on girls beyond the
delinquency threshold.
Strengths and limitations
This study had several limitations. First, only moderate to serious
delinquency was taken into account. It might be, however, that although no
sex differences were found in the threshold for delinquency in general, boys’
and girls’ thresholds do differ for violent or serious delinquency. As Moffitt
(1993) claimed, during puberty, delinquent behavior is more normative,
which as we argued may explain the lack of a clear differential threshold.
For less normative behavior, such a threshold may well emerge. This is
difficult to test, however, since serious (violent) delinquent behavior is a rare
phenomenon in juvenile females and therefore such analyses would have
suffered from a lack of power.
Another limitation is that not all factors that have an important risk or
promotive effect on delinquency could be taken into account. This is because
two different studies (the PYS and the PGS) were combined and we were
strict in our decision not to consider factors that were not conistently
measured in both studies. For instance, negative life events (i.e. crime
victimization, abuse, neglect), that have been shown to be important in
predicting delinquency especially for girls (Wong et al.. 2010), could not be
included because of assessment differences.
Despite these limitations, the present study improved on previous
studies by identifying thresholds for delinquency, and by taking into account
promotive factors. Furthermore this study focused on self-reports of
delinquency, and included shared as well as sex-specific risk and promotive
factors, and examined thresholds longitudinally at two age periods.
Moreover, we showed that some of our design improvements actually
improved predictions compared to previously studies.
10
See footnote 13.
136
137
Chapter 7
General discussion
139
140
This dissertation aimed to increase our understanding of the correlates of
girls’ delinquency, how these differ from those of boys’ delinquency and
how the sex difference in the prevalence of delinquency can be explained.
As we showed in chapter 1, quite a number of theoretical frameworks have
been proposed to explain differences in the prevalence of offending and the
criminal careers of girls and boys, but very few of these explanations have
been properly tested. The practical importance of gaining knowledge about
factors related to girl delinquency lies in the fact that increasingly more girls
enter the juvenile justice system the past years. Although it is the question
whether girls are indeed becoming more delinquent than in the past or
whether the police is more focused on (offences usually committed by) girls
nowadays, there is an increasing group of girls present in the juvenile justice
system to deal with. If girls have similar risk factors as boys, it is likely that
interventions used to handle boys’ delinquency and theories to explain it, can
also be applied to girls. If not, however, different programs and theories
should be developed for girls.
The first aim of this dissertation was therefore to generate an overview
of risk factors related to delinquent behaviour of adolescent girls and to see
which of these were shared with boys and which of them were sex-specific.
The second aim of this dissertation was to test explanations for the
difference in prevalence between boys and girls. Using these risk factors, the
exposure hypothesis, the vulnerability hypothesis and the threshold
hypothesis were examined. The results regarding the first aim are described
and discussed first, followed by the results regarding the second aim.
Subsequently, the implications for interventions are discussed, limitations
and strengths of this dissertation are outlined and an agenda for future
research is proposed.
Unique and shared risk factors of delinquency
In Table 7.1 the risk factors found in chapter 3 and those that appeared to be
univariately related to delinquency in chapter 4 are combined. Note that in
these chapters it was not directly tested whether the factors related to
delinquency were risk factors, promotive factors or both. For the sake of
simplicity, the factors in the table are defined or rephrased in such a way that
they all had a positive relationship with delinquency and therefore referred to
as risk factors. The results of chapter 5 and 6 were not included in the table,
since in chapter 5 no distinction was made between risk factors of boys and
girls and since chapter 6 measured risk and promotive factors before
adolescence during childhood.
There was an overlap between the factors found to be related to
delinquency in chapter 3 (that provided a synthesis from previous research)
and those examined in chapter 4 (that analysed new data). It should be noted
that these included only factors from the parent (i.e. mothers and fathers
separately) and peer domain. Almost no inconsistent results were found
141
between chapter 3 and chapter 4, but also not many findings from chapter 3
were replicated by chapter 4. This is because previous studies mainly
examined parents in general, whereas chapter 4 distinguished between
mothers and fathers. Therefore, results from this dissertation are more
detailed than previous studies. Findings that were replicated concerned
maternal support and the number of delinquent friends. Low levels of
maternal support and having many delinquent friends were consistently
found to be unique factors for girls.
Taking all the results together, boys and girls appeared to have many
risk factors for adolescent delinquency in common, especially regarding
personality, specific parenting characteristics (such as parental monitoring,
harsh discipline, number of rules at home, negative interaction, parental
solicitation and involvement of parents in school) and peers. However, some
factors were uniquely associated with either boys’ or girls’ delinquent
behaviour. Regarding the individual domain, unique factors for boys were
mainly birth problems, whereas for girls unique risk factors referred to
mental health problems. In the family domain, unique factors – for both boys
and girls – were found with regard to maternal factors in particular though
not with regard to paternal factors. Furthermore, it is noteworthy that unique
factors for girls often reflected the severity of risks, such as the number of
life events, the extent of friends’ delinquency and the extent of best friend’s
delinquency in specific.
Implications for the girl delinquency approaches
A first conclusion of this thesis is therefore that some unique risk factors for
girl delinquency exist, but that boys and girls predominantly share risk
factors. Moreover, it should be noted that, for boys as well as for girls, a
wide range of factors is related to delinquent behaviour. This indicates that a
theoretical framework that only includes specific risk domains cannot offer a
complete explanation for delinquent behaviour. This also implies that the
cumulative risk approach – which claims that delinquency is not caused by a
single risk factor, but that multiple factors are always involved – is necessary
for the study on boy and girl delinquency and thus for the development of
theories.
A further reason for the inclusion of multiple factors or risk domains is
that only then it is possible to integrate the add-and-stir and the female-only
focused approach. As we integrated the approaches in this dissertation
shared as well as sex-specific factors could be identified. Otherwise,
important results would probably have been overlooked. For instance,
interesting sex differences about maternal factors and mental health
problems would not have been noticed if the focus had been on factors from
existing theories. Besides, in mainstream theories, the distinction between
mothers and fathers is never made and differences between maternal and
paternal factors – that fathers have a similar effect on boys’ and girls’
142
delinquency, but that mothers had different effects – could thus not have
been detected. Focusing on sex-specific factors only, on the other hand,
would have been poor as well, since all the risk factors that boys and girls
share would not have been detected. Therefore, factors and domains from
existing theories as well as sex-specific factors must be included in order to
include the relevant factors for girl delinquency. The cumulative risk
approach seems a suitable approach to do so.
Explanations for the sex difference in the prevalence of delinquency
Chapter 2 showed that girls are clearly less likely to commit offences than
boys. The second aim of this dissertation was to explain this sex difference
in the prevalence of delinquency. Three different possible explanations were
examined for this universal phenomenon: the exposure hypothesis, the
vulnerability hypothesis and the threshold hypothesis.
The exposure hypothesis
The exposure hypothesis – which states that girls have lower levels of risk
factors and higher levels of promotive factors than boys and are therefore
less likely to be delinquent –was to a large extent confirmed in this
dissertation. Girls reported higher levels than boys of adolescent disclosure
towards mothers and of parental solicitation from mothers. These were both
factors that had an inhibiting effect on delinquency. On the other hand, boys
were exposed to more delinquent friends than girls, while having many
delinquent friends was a risk factor for juveniles to be delinquent
themselves. It was demonstrated that sex was not a predictor of delinquency
anymore when differences in exposure regarding these three factors were
taken into account. Differential exposure to adolescent disclosure towards
mother, regarding parental solicitation of mother and regarding peer
delinquency together therefore appeared to be related to – and might even
explain – the higher prevalence of delinquency among boys than among
girls.
Future research should examine why the sex difference in the
prevalence in delinquency is associated with the sex differences in exposure
regarding the two mother factors and peer delinquency. In addition, causal
effects should be investigated. The study on differential exposure used
crossectional data and therefore it is uncertain whether girls are less often
delinquent than boys because they are differently exposed to particular
factors or that differential exposure to these factors was a result of
differences in delinquent behaviour. As previous longitudinal research has
already shown that parenting and peer factors can influence antisocial
behaviour or delinquency of girls over time (Landsheer & Van Dijkum,
2005; Neumann, Koot, Barker, & Maughan, 2010; Van der Rakt, Weerman,
& Need, 2005), it is likely that the first explanations will have more
143
144
Table 7.1
Shared and sex-specific factors related to delinquency
Shared risk factors
Victimization
Individual
Being harassed by an adult
Low self-control
Aggression
Low IQ
Low self-esteem
Being unashamed for parents, friends
and teachers
Positive attitude towards delinquency
Substance abuse
Inadequate parenting
Family
Overall parental monitoring
Decreasing parental monitoring
Low involvement parents in school
Small number of rules at home
Harsh discipline
Low parental warmth
Conflicts within family
Delinquent sibling
Mother
Low parental knowledge*
Low involvement in peer relations*
Low parental solicitation*
Mother-child conflict*
Negative interaction*
Low adolescent disclosure*
Unique factors for boys
Birth complications
Psychological wellbeing
High parental knowledge
about friends
Single parenthood
Low support (but less
strong than for girls)*
Convicted mother
Unique factors for girls
High number of life events
Disobedience
High self-esteem
Depression
Suicidal behaviour
Being seldom at home
Physical abuse by parents
Low parental trust
Parenting style
Low support**
Low quality mother-child
Relationship
145
Shared risk factors
Inadequate parenting
Parenting style
Low parental knowledge
Low involvement in peer relations*
Low parental solicitation*
Low quality father-child
relationship
Father-child conflict*
Negative interaction *
Low adolescent disclosure*
Convicted father*
Having delinquent friends
Truant best friend*
Having negative friends
Problematic youth group member
Having a romantic partner
Many activities with friends
Having friends disliked by parents
Low commitment
Low child-teacher
relationship
Many delinquent friends**
Highly delinquent best friend*
Unique factors for girls
Disadvantaged neighbourhood
Community
*Factor added by this dissertation to the results of previous European studies
**Replicated by this dissertation
School
Peers
Father
Low achievement
Unique factors for boys
predictive strength, although a (partial) reverse association cannot be
outruled.
The vulnerability hypothesis
The vulnerability hypothesis was not confirmed in this dissertation.
According to this hypothesis, boys and girls have different associations
between risk factors and delinquency (i.e. they are differentially vulnerable
to risk factors) and these differences are associated with the sex difference in
delinquency. More specifically, girls are supposed to be less vulnerable for
risk factors and more ‘vulnerable’ for promotive factors than boys and
therefore less likely to be delinquent. Although our study found that boys
and girls were differentially vulnerable for some factors, these differences
were not always in the direction expected by the hypothesis. As expected,
support from mothers appeared to have a stronger promotive effect on girls
than on boys. Nevertheless, maternal control had a promotive effect on boys
but a risk effect on girls. Also, the level of delinquency of best friends was
positively related to girls’ delinquency whereas no relation was found with
boys’ delinquency. It should be kept in mind though that – again - results
were based on cross-sectional data and therefore no inferences could be
drawn about causal relationships.
As with our data girls were not found to be less vulnerable for risk
factors than boys, but instead more vulnerable for risks, therefore differences
in vulnerability could not explain the sex difference in delinquency. In
chapter 6 it was found as well that girls’ delinquency was in general related
to more risk factors than boys’ delinquency. However, in line with the
vulnerability hypothesis, girls had more promotive factors than boys. It
might therefore be that the reason for the sex difference in delinquency is not
that girls are less vulnerable for risk factors, but only that they are more
vulnerable for promotive factors than boys. This should be examined by
future research, since it was not examined in chapter 4 whether factors had a
risk or a promotive effect on delinquency.
Differential exposure and vulnerability appeared to operate
concurrently, at least regarding maternal parenting factors. The higher
exposure of girls in terms of promotive maternal parenting factors (i.e.
especially to adolescent disclosure and parental solicitation) than boys
coincided with differential vulnerability for maternal parenting (i.e.
especially for maternal knowledge and control). This combination of
differential exposure and vulnerability for maternal parenting factors did
however not appear to be able to explain why girls less often commit
offences than boys. Again, this was not surprising, given the direction of the
previous mentioned vulnerability differences (i.e. the risk effect of maternal
control for girls and the promotive effect for boys).
146
146
The threshold hypothesis
The threshold hypothesis – claiming that girls compared to boys become
delinquent only after they are exposed to a higher number of risks – was also
not confirmed in this dissertation. In the first study (i.e. chapter 5) on the
threshold hypothesis, the threshold for adolescent delinquency was examined
using the same analytical strategy as previous studies (Alemagno et al.,
2006; Van der Laan & Van der Schans, 2010), that is, comparing the risk
levels of delinquent girls with those of delinquent boys. However, the study
improved on previous studies by taking into account a non-delinquent
comparison group, by examining self-reported data of a random populationbased sample and by considering risk as well as promotive factors. With
these improvements, no differences in risk levels appeared to exist between
delinquent boys and girls, suggesting that there was no sex difference in
thresholds for delinquency in adolescence.
It is important to realize though that in order to examine the threshold, it
is not entirely accurate to analyse the average risk level of delinquent
juveniles. This way, conclusions can only be drawn about the average
number of risk factors of already delinquent boys and girls, but not whether
girls actually have a higher threshold to become delinquent. The second
threshold study of this dissertation (i.e. chapter 6) therefore attempted to
examine the threshold itself by identifying a ‘tipping point’ for boys and
girls. This was studied for two age periods: for delinquency in late childhood
(ages 10-12) and in adolescence (ages 13-16). For delinquency in late
childhood a higher threshold was found for girls than for boys, whereas for
delinquency in adolescence boys appeared to have a higher threshold. In
addition to this fluctuation, the threshold differences were small. The
differences would most probably be insignificant if there was any way to test
it. Therefore, as in chapter 5, it can be concluded that no true sex differences
in thresholds exist. This indicates that it is not a difference in thresholds for
delinquency between boys and girls that accounts for the sex difference in
the prevalence of delinquency.
However, chapter 6 also showed that the relationship between the
probability to be delinquent and the level of risk is quadratic for girls –
indicating that the probability of delinquency amplifies when girls have
higher levels of risk – whereas for boys the probability had a linear
relationship with the risk level. Because of this amplification, delinquent
girls on average have more risk factors than boys once they are delinquent.
This might be one reason that previous research found higher levels of risk
among delinquent girls than among delinquent boys. Another reason is that
previous studies examined official records of crime. Therefore, the threshold
that was picked up in previous studies was mostly a criminal justiceinvolvement threshold. This is in line with what was found in chapter 2: the
sex difference in delinquency appeared much larger in official records, than
in self-reported data. This illustrates that the police are relatively less
147
focused on delinquent girls than on delinquent boys. Besides, chapter 2
suggested that official records strongly reflect the priorities and effort of the
police. This was demonstrated in the increase of arrest of particularly those
offences that gained attention in policy programs of the Dutch government.
The reason for the higher threshold to be arrested for girls, might therefore
be related to the policy of the police.
Implications for interventions
Since increasing numbers of girls are involved in the juvenile justice system,
there is more and more demand for suitable interventions for girls. The
studies in this dissertation have given some important indications for such
interventions. Firstly, interventions should consider many different factors
and domains. Girl delinquency (but boy delinquency as well) appeared to be
related to many factors from the individual, parent, school, and peer domain.
Secondly, interventions may focus on risk factors that are also important for
boys. These include mainly factors in the peer domain and regarding
personality and specific parenting factors. In the Netherlands, the
interventions currently offered to delinquent girls are similar to those offered
to delinquent boys. As the studies in this dissertation have shown that girls
share many risk factors with boys, these interventions are probably partly
effective for delinquent girls. ‘Probably’, because no effect studies have
been carried out yet in the Netherlands for these interventions for delinquent
girls. Thirdly, this study identified a number of unique factors for girls that
interventions should take into account, such as maternal factors and mental
health problems.
In line with the studies in this dissertation, a recently published report of
Slotboom, Wong, Swier, and Van der Broek (2011), that examined the
suitability of interventions now offered to Dutch delinquent girls, also
concluded that broad-focused interventions for girls are most likely to be
successful. In addition it was recommended that mental health problems,
traumatic experiences and sexuality are of importance in tackling girl
delinquency. Sexuality was not investigated in this dissertation, but the
number of negative life events was indeed found to be a unique risk factor
for girls.
Before it can be concluded with certainty, however, that boys’
interventions can be applied to girls as well, research should examine the
effectiveness of interventions for girls. In addition more research should be
carried out regarding the mechanisms behind the association between risk
factors and delinquency. If it is known how risk factors affect delinquency
exactly, interventions could focus on these mechanisms in particular which
can enhance the effectiveness of interventions. More detailed
recommendations for future research are given in a later paragraph.
148
Limitations and strengths of the studies in this dissertation
Limitations
As with all research, a number of limitation to the studies in this dissertation
must be noted. One limitation is the sole reliance on self-reported data. As
mentioned in chapter 2, self-report is limited to some extent. Delinquency
concerns a sensitive topic. Most adolescents are too aware that committing
offences is not desirable behaviour and might therefore be reluctant to report
it (Farrington, 1973). In addition, respondents might have forgotten about
committed offences (Glasner & Van der Vaart, 2009; Tourangeau, Rips, &
Rasinski, 2000) or they might conceal criminal behaviour because they fear
that their anonymity is not ensured (Hessing & Elffers, 1995). On the other
hand, respondents may overreport because they like to brag about their
behaviour or because they report about offences outside the reference period
(Nijboer, 1995; Sudman, Bradburn, & Schwarz, 1996). However, it has been
shown that self-reports are a sufficiently reliable and valid means to measure
delinquency (Thornberry & Krohn, 2000; Van der Heijden, Sijtsma & ’t
Hart, 1995).
It may also be that differences in accuracy or honesty exist between
boys and girls. Piquero, MacIntosh, and Hickman (2002) have shown that
males and females vary in the way they report on delinquency items. Males
and females had different items on which they found it easier to report
positively on. Males had less difficulty, for instance, to admit that they had
broken into a house or carried a hidden weapon, whereas girls more easily
reported about being loud or unruly in public place or about hitting their
parent. Differences might thus exist in the way boys and girls report about
various items. However, also in general sex differences might exist
regarding reporting behaviour. It might be, for example, that girls are more
hesitant to admit that they had committed offences than boys, since this
might not match with their idea of feminity. On the other hand, it might also
be that girls, obedient as they usually are, (too) honestly fill out every
question. If girls were overreporting, this implies that the differences
between non-delinquents and delinquents are not that clear-cut and therefore
the effects found in the studies of this dissertation would likely be
underestimated. The consequence of underreporting of girls depends on
which of the delinquent girls underreport. If only serious delinquent girls
underreported, this would have the same consequence as overreporting, since
there would then be relatively little serious delinquents among girls leading
to fewer differences between non-delinquents and delinquents. If only minor
delinquent girls underreported, differences between non-delinquents and
delinquents are exaggerated leading to overstated effects. In any case, we are
not sure whether there are differences between boys and girls in how they
report on delinquency questionnaires and it is therefore uncertain whether
our results are confounded with reporting biases. However, as will be
149
discussed in the next paragraph, the use of official records is even more
problematic; the use of self-reported data are therefore most adequate.
Another limitation of this dissertation was that it was not examined
consistently whether the factors related to delinquency of girls were risk
factors, promotive factors, or both. Without testing whether factors had a
risk effect, a promotive effect, or both, factors were either considered as risk
or as promotive factors, depending on how the factor was defined and on the
direction of the association. For instance, the negative relationship between
adolescent disclosure and delinquency as found in chapter 4was presumed to
represent a promotive effect of adolescent disclosure. Put differently, we
assumed that higher levels of adolescent disclosure would protect
adolescents against becoming delinquent. However, it might very well be
that it is not a promotive factor, but instead a risk factor, which would imply
that low levels of adolescent disclosure put adolescents at risk for becoming
delinquent.
One more limitation, particular to the study on differential exposure and
vulnerability in particular (see chapter 4), was the use of cross-sectional data.
Using such data, it is not clear whether a factor that is related to delinquency
is a cause of delinquency, an effect, or both. To disentangle these effects,
longitudinal data are essential. However, longitudinal data were not available
in the period that the study was carried out. Moreover, the exposure and
vulnerability hypotheses were only examined in early puberty (when the
adolescents were aged 12 to 13). It might very well be that different results
regarding exposure and vulnerability will be found at later ages. After all,
the older adolescents become, the more important peers become and the less
important parents. This may have an influence on the level of exposure to
parent and peer factors, but also on the associations between the factors and
delinquency. As the RADAR study from which we analysed data was on
going as the research was carried out, we had to make do with the waves that
were available.
Furthermore, in the studies that addressed sex differences in the
threshold no attention was paid to the differential importance of various
factors in measuring the threshold. Instead, every risk and promotive factor
was considered to have an equal impact on delinquency. As we know that
some factors are more important than others (and these may differ for males
and females), future research should examine the threshold by considering
the relevance of each factor.
Strengths
Despite of these limitations, the studies in this dissertation also had clear
strengths. First of all, self-reported measures of population-based samples
were used. Regardless of the previous mentioned possible drawbacks of selfreported measures, it is the least problematic method to examine girl
delinquency in such a way that results can be generalized to all delinquent
150
girls. Another way to measure delinquency is to use official records.
However, this will always result in a select sample. As already discussed, not
all delinquents are arrested or incarcerated and therefore official records only
represent those delinquents the (policy of the) police is focused on. These are
usually the most serious or obvious delinquents. The measures used in this
study - self-reported data of population samples – are therefore the least
biased and thus the most adequate.
A second strength of this study is the inclusion of a non-delinquent
comparison group. Including such a group is very difficult or even
impossible in studies using official records and this possibility is therefore
another advantage of population-based samples. Without such a comparison
group no proper claims can be made about risk or promotive factors.
Third, boys and girls were compared without method differences. To be
able to conclude anything about sex differences in delinquency and its
correlates, this is essential. Otherwise, differences found can be due to
dissimilarities in the methods and not to actual differences.
Fourth, it was directly tested whether differential exposure and
vulnerability were related to the sex difference in the prevalence of
delinquency. Previous studies only examined whether differences in
exposure and vulnerability existed, but not whether these could account for
the difference in prevalence. The studies in this dissertation have shown that
if boys and girls are differently exposed to or vulnerable for factors that this
did not necessarily mean that this was related to the sex difference in
delinquency. Furthermore, this dissertation concurrently tested whether
differential exposure and vulnerability were related to the sex difference in
delinquency.
Fifth, if possible, maternal and paternal factors were separated instead
of merged into parental factors. The finding that mothers had a unique
influence on boys and girls whereas fathers played a similar role regarding
delinquent behaviour of the two sexes, shows that it is important to
disentangle mother and father influences on delinquency. However, most
previous studies – also on boys’ delinquency – combined maternal and
paternal factors or only examined factors in the mother domain, since
mothers are often the main caregiver.
Sixth, attention was paid to promotive factors in examining the
threshold hypothesis. Since the threshold is contingent on the level of risk,
and promotive factors can neutralize risks, considering promotive factors in
threshold research is essential. This has – to our knowledge – not been
implemented before. Lastly, the threshold hypothesis was studied
extensively, with the exact location of the threshold identified and compared
between boys and girls.
151
Agenda for future research
Similar factors, similar mechanisms?
Many risk factors appeared to be similar for boys and girls. An important
next step is to examine whether the shared factors for boys and girls assert
influence in the same manner for both sexes. We need to know whether the
mechanisms behind the associations are similar between boys and girls. If
so, than we can be certain that particular existing theories developed to
explain boys’ delinquency can be applied to girls’ delinquency as well. For
example, adolescents with delinquent friends are more likely to become
delinquent themselves because they learn attitudes, techniques and motives
from them (Sutherland, 1947). It is the question though whether girls – when
they have delinquent friends – learn delinquent behaviour as well in this
way. It may be that different processes underlie the influence of delinquent
friends on adolescents’ delinquency for boys and girls. For theoretical
advancement, it is essential to examine this. Besides, the sex of friends
should also be taken into account as processes might dependent of the sex of
delinquent friends too. It might very well be that a group of delinquent male
friends has a different influence on a girl than a group of delinquent female
friends for instance. Hence, future studies should investigate whether and
how the mechanisms behind the association between factors and delinquency
vary by sex.
The unique role of mothers
Research should also examine the unique factors of girl delinquency more
closely. One interesting domain in which girls appeared to experience quite
some unique influences is the mother domain. The relationship between
daughters and mothers appeared to have an important and distinctive effect
on girls’ delinquency. Mothers therefore seem to have a stronger inhibiting
effect on delinquent behaviour of girls than on that of boys. This possibly
not only because girls interact more with their mothers, but also because
maternal knowledge (multivariately) and support (univariately) has a
stronger restraining effect on girls. Girls are perhaps more afraid than boys
to endanger the bond they have with their mother by committing crimes. The
social control theory (Hirschi, 1969) postulates that people naturally have the
tendency to commit crimes, but that good relationships with other people can
inhibit this propensity. Future research should examine whether this indeed
is particularly so for girls with respect to their mothers, and why. On the
other hand, it was found that high levels of maternal control had a risk effect
on girls’ delinquency, whereas maternal control was a promotive factor for
boys. It should be noted that cross-sectional data were used for these results,
but it might be that different processes are involved here. Prospective studies
should examine the nature of this sex difference.
Furthermore, since quite some differences were found between mothers
and fathers, future research should always distinguish between mothers and
152
fathers. It should also be investigated why it is that mothers have a
differential effect on boys and girls, whereas fathers do not. It might be that
because girls are more relationship-focused than boys, the relationships with
both parents (and even the teacher) affect their delinquent behaviour whereas
for boys this only holds for the relationship with the parent of the same sex.
This is, however, only speculation; again, future research should examine
this.
The finding that mothers have such an important influence on girls
makes it even more relevant to study girl delinquency. This is because many
delinquent girls will, most likely, eventually become mothers as well. The
question is what will happen with the children of these mothers if their
delinquent behaviour is not intervened on during adolescence. Previous
research has shown that seriously delinquent girls are at increased likelihood
to suffer problems in adulthood, such as various psychological or psychiatric
problems, substance abuse, being in a dysfunctional or violent relationships,
poor educational achievement and unstable employment (Fontaine et al.,
2008; Giordano, 2010; Pajer, 1998). These features likely interfere with
adequate parenting (Cummings, Davies, & Campbell, 2000).
The pathology of the delinquent girl
Next to mothers, girls also appeared to be uniquely affected by mental health
disorders. Previous studies that examined characteristics of arrested or
detained boys and girls also found that a larger number of delinquent girls
than boys had mental health problems, such as depression, anxiety and anger
(Alemagno et al., 2006; Cauffman, Lexcen, Goldweber, Shulman, & Grisso,
2007; Gavazzi et al., 2006; Johansson & Kempf-Leonard, 2009; Kataoka et
al., 2001). However, these results concern incarcerated youth and, again, it
might be that girls are at relatively higher risk than boys to become arrested
when they have mental health problems. Furthermore, girls in general suffer
more from mental health problems than boys (Gray, 2002) and since no
control group was included in these studies, it could not be tested whether
delinquent girls indeed had more mental health problems than the average
girl. More research should thus be carried out in better suited designs to
investigate why delinquent girls and not delinquent boys are affected by
mental health problems. It is important for future research not to study this in
a setting were selection effects can play a role, but instead to use randomly
selected populations-based samples. In addition, it is essential to take a nondelinquent comparison group into account.
The sex difference in delinquency
The studies in this dissertation have shown that the sex difference in
delinquency might be explained by differential exposure between boys and
girls. Sex differences in vulnerability appeared to be less important for the
explanation and sex differences in thresholds for delinquency were found to
153
be (almost) non-existent and therefore not related to the sex difference in the
delinquency prevalence. Hence, future research should focus mostly on sex
differences in exposure to account for the sex difference in delinquency.
In specific, it was found that sex differences in exposure to parental
solicitation of mothers, to adolescent disclosure toward mothers and to the
number of delinquent friends were related to the difference in delinquency
prevalence between boys and girls. These results imply that if girls had a
similar number of delinquent friends as boys and similar levels of interaction
with their mothers, their delinquency prevalence might be as high as that of
boys. However, this is probably not true. Although the association between
sex and delinquency in our analysis became insignificant when differential
exposure to the previous mentioned factors was included, this might not be
the entire explanation of the sex difference. It should not be ignored, for
instance, that boys and girls are biologically seen different types of beings.
Sex differences in biological factors therefore most probably also play a role
in explaining in prevalence differences. It is even possible that biological sex
differences are the reason of the differences found in exposure. Future
research should pay more attention to including and combining biological
and environmental factors that might influence (the sex difference in)
delinquency. In the current dissertation this was not possible, since data on
biological factors were not available.
Besides, more research is needed that attempts to explain why it is so
that differential exposure to adolescent disclosure to mothers, to maternal
solicitation and to peer delinquency are associated with the lower prevalence
of delinquency among girls. Put differently, it should be investigated how
these factors influence delinquency in boys and girls. Furthermore, research
should try to explain the differences between boys and girls in exposure to
these factors. Why is it that girls tell their mothers more than boys? And why
do mothers put more effort in talking to their daughters than to their sons?
And are girls less exposed to delinquent peers simply because they have less
friends of the opposite sex who are thus less often delinquent, or is there
another explanation?
Sex differences in exposure should also be examined in future research
for different age periods, since different factors may be important during
different ages. In this dissertation it was only possible to examine the
hypotheses in early adolescence (when the adolescents were aged 12 to 13).
Another important issue for prospective exposure (and vulnerability)
studies is that risk and promotive effects should be disentangled. In order to
know what the consequence is of differential exposure, it should be known
how the factors affect delinquency to which boys and girls are differently
exposed to. Besides, the results of this dissertation suggested that differential
vulnerability for risk factors cannot account for the sex difference in
delinquency, but that differential vulnerability for promotive factors maybe
can.
154
Higher arrest threshold
The studies in this dissertation have shown that there is no higher threshold
for delinquency for girls than for boys. Previous studies did however suggest
a higher threshold for girls than for boys, because they found that girls in
juvenile justice institutions had more problematic characteristics than their
male counterparts (Belknap & Holsinger, 2006; Emeka & Sorensen, 2009;
Gavazzi, Yarcheck, & Chesney-Lind, 2006; Gover, 2004; Johansson &
Kempf-Leonard, 2009) or because they found that delinquent girls had a
higher number of risk factors than delinquent boys (Alemagno et al., 2006;
Van der Laan & Van der Schans, 2010). It should be noted that all of these
studies based their results on selective samples of arrested or incarcerated
youth. However, different studies have shown that girls are less often
arrested and incarcerated than boys (Jeffries, Fletcher, & Newbold, 2003;
Sealock & Simpson, 1998; Spohn & Beichner, 2000; Williams, 1999) which
implies that it is a selected group of girls that ends up in juvenile justice
institutions. This suggests that there is no higher threshold for girls than for
boys to become delinquent – as is shown by the self-reported data used in
this dissertation – but that there is a higher threshold for girls to become
arrested. It might thus be that girls should have many problems and risk
factors before the police notice a girl or maybe the police is more lenient
towards girls. Whether and why this occurs should be examined in future
research. It might be that it is a deliberate choice of the police only to take in
problematic girls. However, this would imply that no juvenile justice
interventions are offered to delinquent girls that are less problematic and
therefore no attempt is made to diminish their delinquent behaviour.
To adequately examine a higher threshold for girls to become arrested,
it is essential to include two comparison groups for both boys and girls: a
non-delinquent group and a delinquent-but-not-arrested group. With the nondelinquent group typical characteristics of boys and girls can be
distinguished from characteristics of delinquent boys and girls (i.e. risk
factors). Subsequently, with the delinquent-but-not-arrested group, it can be
investigated whether there is indeed a higher threshold for girls to become
arrested. Next to including these comparison groups, it should be noted again
that including promotive factors is of great importance in examining the
threshold. Furthermore, future research should consider that different factors
can have different weights of influence on delinquency. Otherwise, factors
that only affect delinquency to a small extent are regarded as important as
factors that have a large influence on delinquency.
All in all, this dissertation has shown that the relationship with mothers and
the number of delinquent friends come across as important ingredients of the
sex difference in the prevalence of delinquency. In addition, maternal factors
and mental health problems appeared to be uniquely related to delinquency
of girls. Moreover, no sex differences in thresholds for delinquency appeared
155
to exist. This dissertation therefore provided a next stepping stone for the
development of theories about girl delinquency and about sex differences in
delinquency.
156
References
157
Agnew, R. (2009). The contribution of ‘mainstream’ theories to the
explanation of female delinquency. In M. A. Zahn (Ed.), The
delinquent girl (pp. 147-163). Philadelphia: Temple University Press.
Alemagno, S. A., Shaffer-King, E., & Hammel, R. (2006). Juveniles in
detention: How do girls differ from boys? Journal of Correctional
Health Care, 12, 45-53.
Alesina, A. & Glaeser, E. L. (2004). Fighting poverty in the US and Europe:
A world of difference. Oxford: University Press.
Baker, L. A., Mack, W., Moffitt, T. E., & Mednick, S. (1989). Sex
differences in property crime in a Danish adoption cohort. Behavior
Genetics, 19, 355-370.
Belknap, J., & Holsinger, K. (2006). The gendered nature of risk factors for
delinquency. Feminist Criminology, 1, 48-71.
Bentler, P. M. (1990). Comparative fit indexes in structural models.
Psychological Bulletin, 107, 238-246.
Bijleveld, C. C. J. H. & Wijkman, M. (2009). Intergenerational continuity in
convictions: A five-generation study. Criminal Behaviour and Mental
Health, 19, 142-155.
Blitstein, J. L., Murray, D. M., Lytle, L. A., Birnbaum, A. S., & Perry, C. L.
(2005). Predictors of violent behavior in an early adolescent cohort:
Similarities and differences across genders. Health Education &
Behavior, 32, 175-194.
Blom, M., Van der Laan, A. M., & Huijbregts, G. L. A. M. (2005). Monitor
Jeugd terecht 2005. Den Haag: WODC Cahier 2005-17.
Blum, J. B. A., Ireland, M., & Blum, R. W. M. D. (2003). Gender
differences in juvenile violence: A report from add health. Journal of
Adolescent Health, 32, 234-240.
Bontekoe, E. H. M. (1984). Criminaliteit en geslacht. [Criminal behaviour
and sex] Tijdschrift voor Criminologie, 26, 18-31.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model
fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation
models (pp. 136-162). Newbury Park, CA: Sage
Bruinsma, G. & Lissenberg, E. (1987). Vrouwen in de theoretische en
empirische criminologie. [Women in the theoretical and empirical
criminology] Tijdschrift voor Criminologie, 29, 111-114.
Campbell, A. (1987). Self-reported delinquency and home life: Evidence
from a sample of British females. Journal of Youth and Adolescence,
16, 167-177.
Cauffman, E., Lexcen, F. J., Goldweber, A., Shulman, E. P., & Grisso, T.
(2007). Gender differences in mental health symptoms among
delinquent and community youth. Youth Violence and Juvenile Justice,
5, 287-307.
158
Calhoun, G. B. (2001). Differences between male and female juvenile
offenders as measured by the BASC. Journal of Offenders
Rehabilitation, 33, 87-96.
Chamberlain, P. & Moore, K. (2002). Chaos and trauma in the lives of
adolescent females with antisocial behavior and delinquency. Journal of
Aggression, Maltreatment & Trauma, 6, 79-108.
Chapple, C. L., McQuillan, J. A., & Berdahl, T. A. (2005). Gender, social
bonds, and delinquency: a comparison of boys' and girls' models. Social
Science Research, 34, 357-383.
Chesney-Lind, M. (1989). Girl's crime and woman's place: Toward a
feminist model of female delinquency. Crime and Delinquency, 35, 529.
Chesney-Lind, M. (1999). Challenging females’ invisibility in Juvenile
Court. The ANNALS of the American Academy of Political and Social
Science, 564, 185-202.
Cloninger, C. R. & Gottesman, I. I. (1987). Genetic and environmental
factors in antisocial behavior disorders. In S. A. Mednick, T. E. Moffitt,
& S. A. Stack (Eds.), The causes of crime: New biological approaches
(pp. 92-109). Cambridge: Cambridge University Press.
Cummings, E. M., Davies, P. T., & CAmpbell, S. B. (2000). Developmental
psychopathology and family process: Theory, research, and clinical
implications. New York: The Guilford Press.
Daigle, L. E., Cullen, F. T., & Wright, J. P. (2007). Gender differences in the
predictors of juvenile delinquency: Assessing the generality-specificity
debate. Youth Violence and Juvenile Justice, 5, 254-286.
Daly, K. (1994). Gender, crime, and punishment. New Haven : Yale
University Press.
Davis, M. 1983. Measuring individual differences in empathy: Evidence for
a multidimensional approach. Journal of Personality and Social
Psychology, 44, 113–126.
Dijksterhuis, F. P. H. & Nijboer, J. A. (1987). Meisjes- en
jongensdelinquentie: Dezelfde etiologische processen. [Girl and boy
delinquency. Similar etiological processes] Tijdschrift voor
Criminologie, 29, 104-110.
Dowden, C. & Andrews, D. A. (1999). What works for female offenders: A
meta-analytic review. Crime and Delinquency, 45, 438-452.
Eggen, A.Th.J. (2009). Geregistreerde criminaliteit. [Official crime]. In S. N.
Kalidien & A. Th. J. Eggen (red.), Criminaliteit en rechtshandhaving
2008. Den Haag: Boom Juridische uitgevers.
Eichelsheim, V., Buist, K., Deković, M., Wissink, I., Frijns, T., van Lier, P.,
Koot, H., & Meeus, W. (2010). Associations among the parentadolescent relationship, aggression and delinquency in different ethnic
groups: A replication across two Dutch samples. Social Psychiatry and
Psychiatric Epidemiology, 45, 293-300.
159
Eichelsheim, V. I., Buist, K., Deković, M., Cook, W. L., Manders, W.,
Branje, S., Frijns, T., Van Lier, P. A. C., Koot, H. M., & Meeus, W. (in
press). Negativity in problematic and non-problematic families: A
multigroup Social Relations Model analysis with structured
means. Journal of Family Psychology.
Elliott, D. S., Huizinga, D., & Ageton, S. S. (1985). Explaining delinquency
and drug use. Beverly Hills, CA: Sage.
Eme, R. F. (1992). Selective females affliction in the developmental
disorders of childhood: A literature review. Journal of Clinical Child
Psychology, 21, 354.
Emeka, T. Q. & Sorensen, J. R. (2009). Female juvenile risk: Is there a need
for gendered assessment instruments? Youth Violence and Juvenile
Justice, 7, 313-330.
Fagan, A. A., Van Horn, M. L., Hawkins, J. D., & Arthur, M. W. (2007).
Gender similarities and differences in the association between risk and
protective factors and self-reported serious delinquency. Prevention
Science, 8, 115-124.
Farrington, D. P. (1973). Self-reports of deviant behavior: Predictive and
stable ? Journal of Criminal Law and Criminology, 64, 99-110.
Farrington, D. P. (1986). Age and crime. Crime and justice: An annual
review of research (Vol. 7, pp. 189-250). Chicago: University of
Chicago Press.
Farrington, D. P., Jolliffe, D., Hawkins, J. D., Catalano, R. F., Hill, K. G., &
Kosterman, R. (2003). Comparing delinquency careers in court records
and self-reports. Criminology, 41, 933-958.
Farrington, D. P. & Painter, K. A. (2004). Gender differences in risk factors
for offending: Implications for risk-focused prevention: Home Office.
Fontaine, N., Carbonneau, R., Barker, E. D., Vitaro, F., Hébert, M., Côté, S.,
et al. (2008). Girls’ hyperactivity and physical aggression during
childhood and adjustment problems in early adulthood. Archives of
General Psychiatry, 65, 320-328.
Furman, W. & Buhrmester, D. (1985). Children’s perceptions of the
personal relationship in their social networks. Developmental
Psychology, 21, 1016-1024.
Gaarder, E. & Belknap, J. (2002). Tenuous borders: Girls transferred to adult
court. Criminology, 40, 481-517.
Gavazzi, S. M., Yarcheck, C. M., & Chesney-Lind, M. (2006). Global risk
indicators and the role of gender in a juvenile detention sample.
Criminal Justice Behavior, 33, 597-612.
Gecas, V. & Seff, M. A. (1990). Families and adolescents: A review of the
1980s. Journal of Marriage and the Family, 52, 941–958.
Gilligan, C. (1982). In a different voice. Cambridge, MA: Harvard
University Press.
160
Giordano, P. C. (2010). Legacies of crime. A follow-up of the children of
highly delinquent girls and boys. New York: Cambridge University
Press.
Giordano, P. C., Cernkovich, S. A., & Pugh, M. D. (1986). Friendships
and delinquency. American Journal of Sociology, 91, 1170-1202.
Glasner, T. & van der Vaart, W. (2009) Applications of calender instruments
in social surveys: A review. Quality and Quantity, 43, 333-349
Gover, A. R. (2004). Childhood sexual abuse, gender, and depression among
incarcerated youth. International Journal of Offender Therapy and
Comparative Criminology, 48, 683-696.
Graham, J. & Bowling, B. (1996). Young people and crime (Research study
no. 145). London: Home Office.
Gratz, K. L. & Roemer, L. (2004). Multidimensional assessment of emotion
regulation and dysregulation: Development, factor structure, and initial
validation of the Difficulties in Emotion Regulation Scale. Journal of
Psychopathology and Behavioral Assessment, 26, 41-54.
Gray, P. (2002). Psychology, 4th edition. New York: Worth Publisher.
Haen-Marshall, I. (1996). How exceptional is the United States? European
Journal on Criminal Policy and Research, 4, 7-35.
Hale, W., Keijsers, L., Klimstra, T., Raaijmakers, Q., Hawk, S., Branje, S.,
Frijns, T., Wijsbroek, S., Van Lier, P., & Meeus, W. (2011). How does
longitudinally measured maternal Expressed Emotion affect adolescent
antisocial and depressive symptoms? Journal of Child Psychology and
Psychiatry, 52, 1174–1183.
Hartjen, C. A. & Priyadarsini, S. (2003). Gender, peers, and delinquency: A
study of males and females in rural France. Youth & Society, 34, 387414.
Hauber, A., Toornvliet, L., & Willemse, H. (1987). Achtergronden van
criminaliteit bij meisjes. [Backgrounds of criminality in girls]. In G.
Bruisma (ed.), Vrouw en criminaliteit: Vrouwen als plegers en
slachtoffers van criminaliteit (pp. 41-52). Meppel: Boom.
Heimer, K., De Coster, S., & Unal, H. (2006). Opening the black box:
Understanding the social psychology of the gender gap in delinquency.
Sociology of Crime, Law, and Deviance, 7, 109-135.
Hessing, D. J. & Elffers, H. (1995). De validiteit van de self-report methode
in onderzoek naar regelovertredend gedrag. Tijdschrift voor
Hill, K. G., Howell, J. C., Hawkins, J. D., & Battin-Pearson, S. R. (1999).
Childhood risk factors for adolescent gang membership: Results from
the Seattle Social Development Project. Journal of Research in Crime
and Delinquency, 36, 300-322.
Hirschi, T. (1969). Causes of Delinquency. Berkeley, CA: University of
California Press.
161
Hodgins, S., Kratzer, L., & McNeil, T. F. (2001). Obstetric complications,
parenting, and risk of criminal behavior. Archives of General
Psychiatry, 58, 746.
Hoeve, M. (2008). Parenting and juvenile delinquency. (Doctoral
dissertation, Radboud University, Nijmegen.
Hoeve, M., Dubas, J. S., Eichelsheim, V. E., Van der Laan, P. H., Smeenk,
W., & Gerris, J. R. M. (2009). The relationship between parenting and
delinquency: A meta-analysis. Journal of Abnormal Child Psychology,
37, 749-775.
Holsinger, K. (2000). Feminist perspectives on female offending: Examining
real girls’ lives. Women and Criminal Justice, 12, 23-51.
Honkatukia, P. (1998). Gender, social control and crime: The conforming
behaviour of females. In: P. Honkatukia, Sopeutuvat tytöt? Sukupuoli,
sosiaalinen kontrolli ja rikokset. Helsinki: Hakapaino OY.
Hoyt, S. & Scherer, D. G. (1998). Female juvenile delinquency:
Misunderstood by the juvenile justice system, neglected by social
science. Law and Human Behavior, 22, 81-107.
Hubbard, D. J. & Pratt, T. C. (2002). A meta-analysis of the predictors of
delinquency among females. Journal of Offender Rehabilitation, 34, 113.
Jeffries, S., Fletcher, G. J. O., & Newbold, G. (2003). Pathways to sex-based
differentiation in criminal court sentencing. Criminology, 41, 329-354.
Johansson, P. & Kempf-Leonard, K. (2009). A gender-specific pathway to
serious, violent, and chronic offending? Exploring Howell's risk factors
for serious delinquency. Crime & Delinquency, 55, 216-240.
Jolliffe, D. & Farrington, D. P. (2007). Examining the relationship between
low empathy and self-reported offending. Legal and Criminological
Psychology, 12, 265-286.
Junger-Tas, J. (1992). Criminaliteit en leeftijd. [Crime and age]. Justitiële
Verkenningen, 18, 66-89
Junger-Tas, J., Haen-Marshall, I., & Ribeaud, D. (2003). Delinquency in an
international perspective: The International Self-Reported Delinquency
Study (ISRD). Amsterdam: Kugler.
Junger-Tas, J., Ribeaud, D., & Cruyff, M. (2004). Juvenile delinquency and
gender. European Journal of Criminology, 1, 333-375.
Kataoka, S. H., Zima, B. T., Dupre, D. A., Moreno, K. A., Yang, X., &
McCracken, J. T. (2001). Mental health problems and service use
among female juvenile offenders: Their relationship to criminal history.
Journal of the American Academy of Child and Adolescent Psychiatry,
40, 549-555.
Keenan, K., Hipwell, A., Chung, T., Stepp, S., Stouthamer-Loeber, M.,
Loeber, R., & McTigue, K. (2010). The Pittsburgh Girls Study:
Overview and initial findings. Journal of Clinical Child and Adolescent
Psychology, 39, 506–521.
162
Keenan, K., Loeber, R., & Green, S. (1999). Conduct disorder in girls: A
review of the literature. Clinical Child and Family Psychology Review,
2, 3-19.
Kemppainen, L., Jokelainen, J., Isohanni, M., Järvelin, M.-J., & Räsänen, P.
(2002). Predictors of female criminality: Findings from the northern
Finland 1966 birth cohort. Journal of the American Academy of Child
and Adolescent Psychiatry, 41, 854-859.
Kerr, M. & Stattin, H. (2000). What parents know, how they know it, and
several forms of adolescent adjustment: Further support for a
reinterpretation of monitoring. Developmental Psychology, 36, 366-380.
Kerr, M. & Stattin, H. (2003). Parenting of adolescents: Action or reaction?.
In A.C. Crouter, & A. Booth (Eds.), Children’s influence on family
dynamics: The neglected side of family relationships. Mahwah, NJ:
Erlbaum.
Klein, M. W., Kerner, H.-J., Maxson, C. L., & Weitekamp, E. G. M. (2001).
The Eurogang paradox: Street gangs and youth groups in the US and
Europe. Dordrecht: Kluwer Academic Pub.
Klimstra, T. A., Luyckx, K., Hale, W. W., Frijns T., Van Lier, P., & Meeus,
W. (2010). Short term fluctuations in identity: Introducing a micro-level
approach to identity formation. Journal of Personality and Social
Kratzer, L. & Hodgins, S. (1999). A typology of offenders: A test of
Moffitt’s theory among males and females from childhood to age 30.
Criminal Behaviour and Mental Health, 9, 57-73.
Lahey, B. B., Van Hulle, C. A., Waldman, I. D., Rodgers, J. L., D’Onofrio,
B. M., Pedlow, S., et al. (2006). Testing descriptive hypotheses
regarding sex differences in the development of conduct problems and
delinquency. Journal of Abnormal Child Psychology, 34, 730-748.
Landsheer, J. A. & Van Dijkum (2005). Male and female delinquency
trajectories from pre through middle adolescence and their continuation
in late adolescence. Adolescence, 40, 729-748.
Laursen, B. (1993). The perceived impact of conflict on adolescent
relationships. Merrill Palmer Quarterly, 39, 535-550.
Leaper, C., Anderson, K. J., & Sanders, P. (1998). Moderators of gender
effects on parents’ talk to their children: A meta-analysis.
Developmental Psychology, 34, 3–27.
Lederman, C. S., Dakof, G. A., Larrea, M. A., & Li, H. (2004).
Characteristics of adolescent females in juvenile detention.
International Journal of Law and Psychiatry, 27, 321-337.
Lipsey, M. W. & Derzon, J. H. (1998). Predictors of violent or serious
delinquency in adolescence and early adulthood: A synthesis of
longitudinal research. In R. Loeber & D. P. Farrington (Eds.), Serious
and violent juvenile offenders: Risk factors and successful interventions
(pp. 86-105). Thousand Oaks, CA: Sage.
163
Liu, X. & Kaplan, H. B. (1999). Explaining the gender difference in
adolescent delinquent behavior: A longitudinal test of mediating
mechanisms. Criminology, 37, 195.
Loeber, R. (1990). Development and risk factors of juvenile antisocial
behavior and delinquency. Clinical Psychology Review, 10, 1-41.
Loeber, R., Farrington, D. P., Stouthamer-Loeber, M., & Van Kammen, W.
B. (1998). Antisocial behavior and mental health problems:
Explanatory factors in childhood and adolescence. Mahwah, NJ:
Erlbaum.
Loeber, R., Farrington, D. P., Stouthamer-Loeber, M., & White, H. R.
(2008). Violence and serious theft: Development and prediction from
childhood to adulthood. New York: Routledge.
Loeber, R. & Keenan, K. (1994). Interaction between conduct disorder and
its comorbid conditions: Effects of age and gender. Clinical Psychology
Review, 14, 497–523.
Loeber, R., Slot, N. W., & Stouthamer-Loeber, M. (2008). A cumulative
developmental model of risk and promotive factors. In R. Loeber, N.
W. Slot, P. H. Van der Laan & M. Hoeve (Eds.), Tomorrow’s criminals.
The development of child delinquency and effective interventions (pp.
133-161). Farnham: Ashgate Publishing.
Luthar, S. S. (1993). Annotation: Methodological and conceptual issues in
research on childhood resilience. Journal of Child Psychology and
Psychiatry, 34, 441-453.
Lytton, H. & Romney, D. M. (1991). Parents’ differential socialization of
boys and girls: A meta-analysis. Psychological Bulletin, 109, 267-296.
Maguin, E. & Loeber, R. (1996). Academic performance and delinquency.
In M. Tonry (Ed.), Crime and Justice (pp. 259-281). Chicago, Ill:
University of Chicago Press.
Mansel, J. (2003). Die Selektivität strafrechtlicher Sozialkontrolle: Frauen
und Delinquenz im Hell- und Dunkelfeld, als Opfer und Täter, als
Anzeigende und Angezeigte. In S. Lamnek & M. Boatca (eds.),
Geschlecht - Gewalt – Gesellschaft, 384-406. Opladen: Leske u.
Martens, P. L. (1997). Parental monitoring and deviant behaviour among
juveniles. Studies on Crime and Crime Prevention, 6, 224-240
McAra, L. (2005). Negotiated order: Gender, youth transitions and crime.
British Society of Criminology, e-journal 6.
McFall, R. M. & Treat, T. A. (1999). Quantifying the information value of
clinical assessments with signal detection theory. Annual Review
Mears, D. P., Ploeger, M., & Warr, M. (1998). Explaining the gender gap in
delinquency: Peer influence and moral evaluations of behavior. Journal
of Research in Crime and Delinquency, 35, 251.
Megens, K. & Day, S. (2007). Delinquentie van meisjes: Over de relatie
tussen risicofactoren en delinquent gedrag na behandeling in een JJI.
164
[Delinquency of girls. On the relationship between risk factors and
delinquent behavior after treatment in a Juvenile Justice Institution]
Tijdschrift voor Criminologie, 49, 370-385.
Mertens, N. M., Grapendaal, M., & Docter-Schamhardt, B. J. W. (1998).
Meisjescriminaliteit in Nederland. [Girl delinquency in the
Netherlands] Den Haag: WODC
Messer, J., Goodman, R., Rowe, R., Meltzer, H., & Maughan, B. (2006).
Preadolescent conduct problems in girls and boys. Journal of the
American Academy of Child and Adolescent Psychiatry, 45, 184.
Miller, J. & Mullins, C. W. (2009). Feminist theories of girls’ delinquency.
In M.A. Zahn (Ed.), The delinquent girl (pp. 50-83). Philadelphia:
Temple University Press.
Ministry of Juctice (2002). Privé geweld, publieke zaak. Een nota over de
gezamenlijke aanpak van huiselijk geweld. [Private violence, public
case. An annotation about the collective approach of domestic
violence]. Den Haag: Hega Offset.
Moffitt, T. (1993). Adolescence-limited and life-course-persistent antisocial
behavior: a developmental taxonomy. Psychological Review, 100, 674701.
Moffitt, T. E., Caspi, A., Rutter, M., & Silva, P. A. (2001). Sex differences in
antisocial behaviour: Conduct disorder, delinquency, and violence in
the Dunedin longitudinal study. Cambridge: Cambridge University
Press.
Mounts, N. S. (2002). Parental management of adolescent peer relations in
context: The role of parenting style. Journal of Family Psychology, 16,
58-69.
Murray, J. (2002). Explaining sex differences in offending behaviour. Thesis,
University of Cambridge.
Muthén, L. K. & Muthén, B. O. (1998-2010). Mplus user’s guide. Sixth
edition. Los Angeles, CA: Muthén and Muthén.
Neumann, A., Barker, E. D., Koot, H. M., & Maughan, B. (2010). The role
of contextual risk, impulsivity, and parental knowledge in the
development of adolescent antisocial behavior. Journal of Abnormal
Nijboer, J. A. (1995). Het meten van delinquentie door middel van ‘selfreport’. [Measuring delinquency through self reports]. Tijdschrift voor
Oberwittler, D. (2003). Geschlecht, Ethnizität und sozialräumliche
Benachteiligung. Überraschende Interaktionen bei sozialen
Bedingungsfaktoren von Gewalt und schwerer Eigentumsdelinquenz
von Jugendlichen. In S. Lamnek & M. Boatca (eds.), Geschlecht Gewalt – Gesellschaft, 269-294. Opladen: Leske u.
165
Pajer, K. A. (1998). What happens to" bad" girls? A review of the adult
outcomes of antisocial adolescent girls. American Journal of
Pedersen, W. (1994). Parental relations, mental health, and delinquency in
adolescents. Adolescence, 29, 975-990.
Piko, B. F., Fitzpatrick, K. M., & Wright, D. R. (2005). A risk and protective
framework for understanding youths’ externalizing problem behaviour
in two different cultural settings. European Child & Adolescent
Piquero, A. R., Macintosh, R., & Hickman, M. (2002). The validity of a selfreported delinquency scale: Comparisons across gender, age, race, and
place of residence. Sociological Methods and Research, 30, 492–529.
Pratt, T. C. & Cullen, F. T. (2005). Assessing macro-level predictors and
theories of crime: A meta-analysis. Crime and Justice, 32, 373- 450.
Putallaz, M. & Bierman, K. (2004). Aggression, antisocial behavior, and
violence among girls: A developmental perspective. New York:
Guilford.
Rantakallia, P., Myhrman, A., & Koiranen, M. (1995). Juvenile offenders,
with special reference to sex differences. Social Psychiatry and
Psychiatric Epidemiology, 30, 113-120.
Riley, D. & Shaw, M. (1985). Parental supervision and Juvenile
delinquency. Home Office Research Study 83, HMSO.
Ritakallio, M., Kaltiala-Heino, R., Kivivuori, J., & Rimpelä, M. (2005).
Brief report: Delinquent behaviour and depression in middle
adolescence: A Finish community sample. Journal of Adolescence, 28,
155-159.
Robins, L. N. (1966). Deviant children grown up. Baltimore, MD: Williams
and Wilkins.
Rolf, J. & Johnson, J. (1990). Protected or vulnerable: The challenges of
AIDS to developmental psychopathology. In J. E. Rolf, A. S. Masten,
D. Cicchetti, K. H. Nuechterlein, and S. Weintraub (Eds.), Risk and
protective factors in the development of psychopathology. New York:
Cambridge University Press.
Rose, A. J. & Rudolph, K. D. (2006). A review of sex differences in peer
relationship processes: Potential trade-offs for the emotional and
behavioral development of girls and boys. Psychological Bulletin, 132,
98–131.
Rosenbaum, J. L. & Chesney-Lind, M. (1994). Appearance and delinquency:
A research note. Crime and Delinquency, 40, 250-261.
Rouwette, E., van Hooff, P., Vennix, J., & Jongebreur, W. (2007). Modeling
crime control in the Netherlands: insights on process. Paper presented
at the 25th International Conference of the System Dynamics Society.
166
Rowe, D. C., Vazsonyi, A. T., & Flannery, D. J. (1995). Sex differences in
crime: Do means and within-sex variation have similar causes? Journal
of Research in Crime and Delinquency, 32, 84.
Rutenfrans, C. J. C. (1989). Criminaliteit en sexe.[Criminal behaviour and
sex] (Doctoral dissertation, University of Nijmegen).
Rutter, M. (1979). Protective factors in children’s responses to stress and
disadvantage. In M. W. Kent & J. E. Rolf (Eds.), Primary prevention of
psychopathology, Vol. 3: Social competence in children (pp. 49-74).
Hanover, NH: University Press of New England.
Sameroff, A. J., Bartko, W. T., Baldwin, A., Baldwin, C., & Seifer,
R.(1998). Family and social influences on the development of child
competence. In M. Wei & C. Fairing (Eds.), Families, risk, and
competence. Mahwah, NJ: Erlbau
Sameroff, A. J., Seifer, R., Barocas, R., Zax, M., & Greenspan, S. (1987).
Intelligence quotient scores of 4-yearold children: Social-environmental
risk factors. Pediatrics, 79, 343-350.
Scholte, R. H. J., Engels, R. C. M., De Kemp, R. A. T., Harakeh, Z., &
Overbeek, G. (2007). Differential parental treatment, sibling
relationships and delinquency in adolescence. Journal of Youth and
Adolescence, 36, 661–671.
Schwartz, S., Klimstra, T., Luyckx, K., Hale, W., Frijns, T., Oosterwegel,
A., van Lier, P., Koot, J., & Meeus, W. (in press). Daily dynamics of
personal identity and self-concept clarity. European Journal of
Personality.
Sealock, M. D. & Simpson, S. S. (1998). Unraveling bias in arrest decisions:
The role of juvenile offender type-scripts. Justice Quarterly, 15, 427.
Simourd, L. & Andrews, D. A. (1994). Correlates of delinquency: A look at
gender differences. Forum on Corrections Research, 6, 26-31.
Singer, J. D. & Willett, J. B. (2003). Applied longitudinal data analysis.
Modeling change and event occurrence. Oxford: University Press.
Slotboom, A., Weerman, F. M., & Bijleveld, C. C. J. H. (2005). Misconduct
and delinquency of schoolfemales. Presentation on ESC, August 2005.
Slotboom, A., Wong, T. M. L., Swier, C., & Van der Broek, T. C. (2011).
Delinquente meisjes: Achtergronden, risicofactoren en interventies.
[Delinquent girls: Background characteristics, risk factors and
interventions]. Den Haag: Boom Juridische uitgevers.
Smith, D. J. & McAra, L. (2004). Gender and youth offending. The
Edinburgh Study of Youth Transitions and Crime, number 2, Centre for
Law and Society, The University of Edinburgh.
Spohn, C. & Beichner, D. (2000). Is preferential treatment of female
offenders a thing of the past? A multisite study of gender, race, and
imprisonment. Criminal Justice Policy Review, 11, 149.
Stattin, H. & Kerr, M. (2000). Parental monitoring: A reinterpretation. Child
Development, 71, 1072-1085.
167
Steffensmeier, D. & Allan, E. (1996). Gender and crime: Toward a gendered
theory of female offending. Annual Review of Sociology, 22, 459-487.
Steffensmeier, D. & Haynie, D. (2000). Gender, structural disadvantage, and
urban crime: Do macrosocial variables also explain female offending
rates? Criminology, 38, 400-438.
Steffensmeier, D., & Schwartz, J. (2009). Trends in girls‘ delinquency and
the gender gap: Statistical assessment of diverse sources. In M. A. Zahn
(Ed.), The delinquent girls (pp. 50-83). Philadelphia: Temple University
Press.
Stouthamer-Loeber, M., Loeber, R., Wei, E., Farrington, D., & Wikström, P.
(2002). Risk and promotive effects in the explanation of persistent
serious delinquency in boys. Journal of Consulting and Clinical
Stouthamer-Loeber, M., Loeber, R., Farrington, D. P., Zhang, Q., Van
Kammen, W., & Maguin, E. (1993). The double edge of protective and
risk factors for delinquency: Interrelations and developmental patterns.
Development and Psychopathology, 5, 683-701.
Sudman, S., Bradburn, N. M., & Schwarz, N. (1996) Thinking about
questions: The application of cognitive processes to survey
methodology. San Francisco: JosseyBass.
Sutherland, E. H. (1947). Principles of Criminology: Fourth Edition.
Philadelphia: Lippincott.
Svensson, R. (2004). Shame as a consequence of the parent-child
relationship: A study of gender differences in juvenile delinquency.
European Journal of Criminology, 1, 477-504.
Swets, J. A. (1964) Signal detection and recognition by human observers.
Contemporary readings. Wiley, New York.
Tankersley, V. L. & Rycraft, J. R. (2007). A qualitative examination of risk
factors influencing female adolescent delinquency in Dallas county,
Texas. University of Texas.
Thornberry, T. P. & Krohn, M. D. (2000). The self-report method for
measuring delinquency and crime. Measurement and Analysis of Crime
and Justice, 4, 33-83.
Thornberry, T. P. & Krohn, M. D. (2003). Taking stock of delinquency: An
overview of findings from contemporary longitudinal studies. New
York: Plenum Publishers.
TK. (2002/2003b). Jeugd terecht; Actieprogramma Aanpak
Jeugdcriminaliteit 2003-2006. 28 741, nr. 1
Tonry, M. (1999). Why are US incarceration rates so high? Crime and
Delinquency, 45, 419-437.
Torstensson, M. (1990). Female delinquents in a birth cohort: Tests of some
aspects of control theory. Journal of Quantitative Criminology, 6, 101115.
168
Tourangeau, R., Rips, L. J., & Rasinski, K.A. (2000) The Psychology of
Survey Response. Cambridge: Cambridge University Press
Van der Geest, V., & Bijleveld, C. C. J. H. (2008). Personal, background and
treatment characteristics associated with offending after residential
treatment: A 13-year follow up in adolescent males. Psychology, Crime
& Law, 14, 159-176.
Van der Heijden, P. G. M., Sijtsma, K., & 't Hart, H. (1995). Self-report
delinquentie-schalen zijn nog steeds betrouwbaar. Een reactie op de
studies van Bruinsma. [Self reported delinquency scales are still
reliable. A response to the studies of Bruinsma]. Tijdschrift voor
Van der Laan, A. M. & Blom, M. (2006). Jeugddelinquentie: Risico’s en
bescherming. Bevindingen uit de WODC Monitor Zelfgerapporteerde
Jeugdcriminaliteit 2005. [Juvenile delinquency: Risks and protective
factors. Findings of the WODC
Youth Delinquency Survey, 2005]. Den Haag: Boom Juridische
uitgevers.
Van der Laan, A. M., & Van der Schans, C. (2010). Delinquente meisjes:
Zijn ze anders dan jongens? Risico- en beschermende factoren bij
jongeren die een basisraadsonderzoek ondergaan. [Delinquent girls: Do
they differ from boys? Risk and promotive factors of juveniles getting a
prescreen by the Child Protection Board]. Tijdschrift voor
Orthopedagogiek, 49, 149-162.
Van der Rakt, M., Weerman, F., & Need, A. (2005). Delinquent gedrag van
jongens en meisjes: Het (anti)sociale kapitaal van vriendschapsrelaties.
[Delinquent behaviour of boys and girls: The (anti)social value of peer
relationships]. Mens en Maatschappij, 80, 329-354.
Van de Schoot, R., & Wong, T. M. L. (in press). Do delinquent young adults
have a high or a low level of self-concept? Self & Identity.
Van Lier, P. A. C., Frijns, T., Neumann, A., Den Exter Blokland, E., Koot, J.
M., & Meeus, W. (submitted). The RADAR study: Design, description
of sample and validation of cohort assignment.
Warr, M. (2002). Companions in crime: The social aspects of criminal
conduct. Cambridge: Cambridge University Press.
Weerman, F. (2007). Juvenile offending. In M. H. Tonry & C. C. J. H.
Bijleveld (Eds.), Crime and Justice in the Netherlands, Vol. 35.
Chicago: University of Chicago Press Journals.
Williams, M. R. (1999). Gender and sentencing: An analysis of indicators.
Criminal Justice Policy Review, 10, 471.
Wong, T. M. L., Blom, M., & Van der Laan, A. (in press). De inhaalslag van
vrouwen? Omvang, aard en trends in criminaliteit onder meisjes en
vrouwen. In A. Slotboom, M. Hoeve, M. Ezinga, & P. van der Helm
(Eds.), Criminele meisjes en vrouwen: Achtergronden en aanpak. Den
Haag: Boom Juridische Uitgevers.
169
Wong, T. M. L., Loeber, R. Slotboom, A., & Bijleveld, C. C. J. H.
(submitted). Identifying the threshold: The sex difference in the risk
threshold for delinquency at different age periods.
Wong, T. M. L., Slotboom, A., & Bijleveld, C. C. J. H. (2010). Risk factors
for delinquency in adolescent and young adult females: A European
review. European Journal of Criminology, 7, 266-284.
Wong, T. M. L., Slotboom, A., Bijleveld, C. C. J. H., Van Lier, P. A. C.,
Meeus, W. H. J. & Koot, J. M. (submitted). The sex differences in
delinquency: Do girls need a bigger push?
Wong, T. M. L., Slotboom, A., Zahn, M. Bijleveld, C. C .J. H., Van Lier, P.
A. C., Frijns, T., Meeus, W. H. J., & Koot, J.M. (submitted). The sex
difference in delinquency explained? Differential exposure and
vulnerability in the parent and peer domain.
Youden, W. J. (1950). And index for rating diagnostic tests. Cancer, 3, 32,
35.
Yu, R., Branje, S., Keijsers, L., & Meeus, W. (2011). Psychometric
characteristics of Carver and White's BIS/BAS Scales in Dutch
adolescents and their mothers. Journal of Personality Assessment, 93,
500-507.
Zahn, M.A. (2009). The delinquent girl. Philedelphia: Temple University
Press.
Zahn-Waxler, C. & Polanichka, N. (2004). All things interpersonal.
Socialization and female aggression. In: M. Putallaz & K. L. Bierman
(Eds.), Aggression, antisocial behavior, and violence among girls: A
developmental perspective (pp. 48-68). New York: The Guilford Press.
170
Publications
Samenvatting (summary)
Dankwoord (acknowledgement)
Curriculum Vitae
171
172
PUBLICATIONS
This dissertation
Wong, T. M. L., Slotboom, A., & Bijleveld, C. C. J. H. (2010). Risk factors
for delinquency in adolescent and youg adult females: A European
review. European Journal of Criminology, 7,266-284.
Wong, T. M. L., Blom, M., & Van der Laan, A. (in press). De inhaalslag van
vrouwen? Omvang, aard en trends in criminaliteit onder meisjes en
vrouwen. In A. Slotboom, M. Hoeve, M. Ezinga, & P. Van der Helm
(Eds.), Criminele meisjes en vrouwen: Achtergronden en aanpak. Den
Haag: Boom Juridische Uitgevers.
Wong, T. M. L., Slotboom, A., Zahn, M., Bijleveld, C. C. J. H., Van Lier, P.
A. C., Frijns, T., Meeus, W. H. J. & Koot, J. M. (submitted). The
relationship between the sex difference in delinquency and differential
exposure and vulnerability in the parent and peer domain.
Wong, T. M. L., Slotboom, A., Bijleveld, C. C. J. H., Van Lier, P. A. C.,
Meeus, W. H. J. & Koot, J. M. (submitted). The sex difference in
delinquency: Do girls need a bigger push?
Wong, T. M. L., Loeber, R., Slotboom, A., Bijleveld, C. C. J. H., & Koot,
J. M. (submitted). Sex and age differences in the risk threshold for
delinquency.
Other publications
Wong, T. M. L., Branje, S. J. T., VanderValk, I. E., Hawk, S. T., & Meeus,
W. H. J. (2010). The role of siblings in identity development in
adolescence and emerging adulthood. Journal of Adolescence, 33, 673682.
Wong, T. M. L., Bijleveld, C. C. J. H., & Slotboom, A. (2009). Van
probleemmeisje naar delinquente vrouw? Criminele carrières van
residentieel behandelde meisjes, van 12 tot 32 jaar. Tijdschrift voor
Slotboom, A., Wong, T. M. L., Swier, C., & Van der Broek, T., (2011).
Delinquente meisjes: Achtergronden, risicofactoren en interventies.
Den Haag: Boom Juridische Uitgevers.
Van de Schoot, R. & Wong, T. M. L. (in press). Do delinquent young adults
have a high or a low level of self-concept? Self and Identity.
Wong, T. M. L. & Van de Schoot, R. (in press). The Effect of Offenders' Sex
on Reporting Crimes to the Police. Journal of Interpersonal Violence.
Hoeve, M., Vogelvang, L., Wong, T. M. L. & Kruithof, B. (In press). Het
mysterie van de criminele vrouw: Theorieën over criminaliteit door
meisjes en vrouwen. In A. Slotboom, M. Hoeve, M. Ezinga, & P. Van
der Helm (Eds.), Criminele meisjes en vrouwen: Achtergronden en
aanpak. Den Haag: Boom Juridische Uitgevers.
173
Van de Schoot, R., Wong, T. M. L., Lugtig, P., Denissen, J. J. A., Oberski,
D. L., Koot, J. M., Van Lier, P. A. C., & Meeus, W. (submitted). How
to handle missing data in Mplus: A comparison of established and novel
approaches.
Van de Weijer, S., Wong, T., Verbruggen, J., Tolsma, J. & Bijleveld, C.
(submitted). The criminal development of high-risk girls in the
Netherlands.
174
SAMENVATTING (summary)
Lange tijd is er weinig aandacht geweest voor meisjescriminaliteit in de
wetenschap, zeker als we dat vergelijken met de aandacht die er voor
criminele jongens was. Nu echter steeds meer meisjes in contact komen met
de politie, neemt de noodzaak toe om meer te weten over hoe hiermee om te
gaan. Het is van belang om te weten welke factoren meisjesdelinquentie
beïnvloeden aangezien er geïntervenieerd zal moeten worden op
risicofactoren voor meisjesdelinquentie om het delinquente gedrag terug te
dringen. Daarom richt dit proefschrift zich in de eerste plaats op het
vaststellen van welke factoren gerelateerd zijn aan meisjesdelinquentie.
Hierbij wordt ook nagegaan of deze factoren gelijk zijn aan de factoren die
zijn gerelateerd aan jongensdelinquentie. Een ander doel van dit proefschrift
is verklaringen zoeken voor het structurele sekseverschil in de prevalentie
van delinquentie, een verschil waarover nog weinig bekend is.
Om eerst een compleet beeld te krijgen van de omvang, aard en trends
van crimineel gedrag van meisjes en jongens, zijn zowel zelfrapportages als
politieregistraties bekeken. Minder meisjes dan jongens blijken delicten te
plegen, ongeacht de bron van gegevens en ongeacht het type delict.
Desalniettemin blijken de sekseverschillen het grootst binnen de
geregistreerde criminaliteit, wat deels lijkt te duiden op een verschil in
behandeling door politie en justitie tussen meisjes en jongens en deels op
selectieve zelfrapportage. Volgens beide bronnen piekt meisjescriminaliteit
op 15-jarige leeftijd. Jongensdelinquentie piekt echter op 16-jarige leeftijd
volgens zelfrapportages en op 19-jarige leeftijd volgens politiecijfers. Op
grond van bestudering van zelfrapportage- en politiegegevens kan
geconcludeerd worden dat meisjescriminaliteit de afgelopen jaren relatief
harder is gestegen dan jongenscriminaliteit.
Wat betreft het eerste doel van dit proefschrift, blijken meisjes en
jongens veel gemeenschappelijke risicofactoren voor delinquentie te hebben.
Vooral op het gebied van persoonlijkheid, bepaalde opvoedingskenmerken
(meer specifiek: monitoring, harde disciplinering, het aantal regels thuis, de
mate waarin ouders informeren naar hun kinderen en de mate waarin ouders
zich bemoeien met school) en leeftijdsgenoten. Hiernaast zijn er een aantal
unieke factoren voor meisjes- en jongensdelinquentie. In het individuele
domein blijken vooral geboortecomplicaties uniek gerelateerd aan
delinquentie van jongens, en mentale gezondheidsproblemen aan dat van
meisjes. In het gezinsdomein zijn er vooral unieke factoren met betrekking
tot moeder. Verschillende unieke moederfactoren zijn gevonden voor
jongens en voor meisjes, terwijl factoren met betrekking tot vader niet
seksespecifiek zijn. Hiernaast is het bijzonder dat factoren die uniek zijn
voor meisjesdelinquentie vaak te maken hebben met de ernst van risico’s,
zoals het aantal negatieve levensgebeurtenissen, de mate van delinquentie
van vrienden en dat van de beste vriend.
175
Voor het tweede doel van dit proefschrift, zijn drie mogelijke
verklaringen voor het sekseverschil in delinquentie onderzocht: de exposure
hypothese, de vulnerability hypothese en de threshold hypothese. De
exposure hypothese beweert dat meisjes minder worden blootgesteld
(exposed) aan risicofactoren en meer aan beschermende factoren. Hierdoor
lopen zij minder risico om delinquent te zijn. Deze hypothese is grotendeels
bevestigd in dit proefschrift. In vergelijking met jongens, zijn meisjes meer
open over zichzelf naar hun moeder (disclosure) en vergaren moeders meer
informatie over hun dochter (solicitation). Beide factoren hebben een
beschermend effect op delinquentie. Aan de andere kant worden jongens
juist meer blootgesteld aan delinquente vrienden, een risicofactor voor
delinquentie. Het verschil in blootstelling aan deze drie risicofactoren tussen
jongens en meisjes blijkt gerelateerd aan het sekseverschil in delinquentie.
De vulnerability-hypothese is niet bevestigd in dit proefschrift. Volgens
deze hypothese zijn meisjes meer sensitief (vulnerable) voor beschermende
factoren dan jongens en minder sensitief voor risicofactoren. In dit
proefschrift worden wel sekseverschillen gevonden in sensitiviteit voor
verschillende factoren, maar niet altijd in de richting die deze hypothese zou
ondersteunen. Zo blijkt steun van moeder, naar verwachting, een sterker
beschermend effect te hebben op meisjes dan op jongens, maar blijkt
controle door moeder een beschermend effect te hebben op jongens en juist
een risico-effect op meisjes. Daarnaast blijkt de mate van delinquentie van
de beste vriend een risico voor delinquentie van meisjes, maar niet voor dat
van jongens. Ook blijkt dat meisjes in algemene zin meer risicofactoren
hebben dan jongens, wat men ook niet zou verwachten op basis van de
hypothese. Meisjes hebben echter ook meer beschermende factoren dan
jongens.
Verschillen in blootstelling en sensitiviteit blijken soms ook gelijktijdig
plaats te kunnen vinden, en wel als het gaat om opvoedingsfactoren van
moeder. Het samengaan van de verschillen in blootstelling en kwetsbaarheid
zijn echter niet gerelateerd aan het sekseverschil in delinquentie.
De threshold-hypothese, die ervan uit gaat dat meisjes een hoger aantal
risico’s hebben dan jongens voordat zij delinquent worden, is ook niet
bevestigd. Delinquente meisjes blijken geen hoger risiconiveau te hebben
dan delinquente jongens. Ook het onderzoek naar de locatie van de threshold
levert geen bewijs dat de threshold hoger ligt voor jongens dan voor meisjes,
voor zowel delinquentie tijdens de late kindertijd (10 tot 12 jaar) als tijdens
de adolescentie (13 tot 16 jaar). Wel blijkt de relatie tussen het aantal risico’s
en de kans op delinquentie kwadratisch te zijn voor meisjes, in tegenstelling
tot de lineaire relatie bij jongens. Dit wijst erop dat de kans op delinquent
gedrag relatief steeds groter wordt bij meisjes, naarmate de risico’s
toenemen.
In het kort heeft dit proefschrift laten zien dat de relatie met moeder en
het aantal delinquente vrienden belangrijke ingrediënten zijn voor het
176
sekseverschil in de prevalentie van delinquentie. Hiernaast zijn bepaalde
moederfactoren en mentale gezondheidsproblemen unieke risicofactoren
voor meisjes. Verschillen in sensitiviteit aan factoren kunnen geen
verklaring bieden voor het sekseverschil in delinquentie. Bovendien blijken
meisjes en jongens een gelijke threshold voor delinquentie te hebben.
177
178
DANKWOORD (acknowledgement)
Na vele uren lezen en schrijven, plannen en bijstellen, denken en staren, en
analyseren en structureren, is mijn proefschrift dan eindelijk af! Hoewel
promoveren grotendeels solistisch werk is, heb ik in die vier jaar erg fijne
mensen om mee heen gehad. Een aantal van hen wil ik in het bijzonder
bedanken.
Allereerst, mijn promotoren. Catrien, ik heb erg veel geleerd van jouw
kijk op paper schrijven. Daarnaast heb ik grote bewondering voor het tempo
waarin jij denkt en handelt (zie bijvoorbeeld jouw reactiesnelheid op
mailtjes). Ik vond het altijd ongelooflijk dat wanneer ik het gevoel had vast
te zitten, jij altijd met veel gemak een oplossing wist aan te dragen en me
weer vertrouwen wist te geven. Bedankt voor je goede begeleiding, je
positieve instelling, je bevlogenheid en scherpe pen.
Anne-Marie, veel dank voor de prettige samenwerking. Zowel bij dit
proefschrift als het WODC-rapport. Je wist me altijd enorm te motiveren,
maar ook vragen bij me op te roepen waar ik nog niet aan had gedacht of
waar ik (stiekem) niet aan wilde denken. Dank ook dat je me altijd betrok bij
alles wat voorbijkwam rondom het thema criminele meisjes. En bovenal
bedankt voor je betrokkenheid, gedrevenheid, toegankelijkheid en
enthousiasme!
Hans, het was erg waardevol om ook door iemand begeleid te worden
die van ‘buitenaf’ kritisch naar mijn stukken keek. Dank hiervoor.
Ik heb mogen werken op een prettige afdeling met zeer diverse mensen.
Ik vond het dan ook erg fijn dat er altijd wel een expert te vinden was op
welk criminologisch gebied dan ook. Hiernaast heb ik ook een hele leuke tijd
gehad op de VU. Vooral de congresbezoekjes waren – naast het feit dat ze
natúúrlijk erg nuttig waren – heel plezierig. Ik wil hiervoor mijn collega’s
van Criminologie erg bedanken. Ik hoop dat ik nog vaak bij jullie mag
komen borrelen. De afdeling was echter niet compleet zonder de mensen van
Strafrecht, waar ik ook altijd leuk contact mee had. Mijn tijd op de VU was
nog leuker doordat ik mocht zingen bij de Faculty All Stars. Geweldig om
soms even mijn proefschrift te laten voor wat het was, om koortjes te
oefenen in de kerkzaal of te repeteren met de hele band. You rock!
De mensen van RADAR wil ik graag bedanken voor het opzetten van
zo’n rijk onderzoek. Het was erg interessant om te mogen werken met zo’n
veelzijdige dataset. Dank aan Endy voor de coördinatie van de
dataverzameling; aan Pieter Bas voor het structureren en verspreiden van de
data; en aan Wim, Pol en Tom voor jullie waardevolle commentaar op de
studies die ik met de RADAR-data uitgevoerd heb.
Hiernaast veel dank aan coauteurs van mijn proefschriftartikelen die ik
nog niet genoemd heb. Margaret, I would like to thank you for the inspiring
meetings we had. Your positive attitude is admirable and contagious. Rolf,
hartelijk dank dat ik mocht werken met de mooie datasets uit Pittsburgh. Het
179
was niet alleen zeer uitdagend en interessant om hiermee aan de slag te gaan,
maar ook erg leerzaam om mijn ideeën over de threshold met jou te kunnen
delen. Dank voor je kritische blik. Daarnaast vind ik het ontzettend
sympathiek hoe jij en Magda mij hebben ontvangen in Pittsburgh.
Dan natuurlijk mijn paranimfen. Barbora, promoveren zonder jou was
onmogelijk geweest. In elk geval zo veel minder leuk! Dat we al in het eerste
jaar hadden besloten elkaars paranimfen te worden, zegt genoeg. Veel
tegengestelde eigenschappen (en smaken!), maar toch zo’n goede match. In
het Engels, but also in Dutch. Sharing frustrations, but also many delights.
We were experts on anticipating on each other’s moods, but we especially
had many many laughs. Thanks for being my roomy! Rens, dank dat ik je
altijd kon bellen met mijn statistische vragen en frustraties. Fijn om soms
ook met iemand van buiten de VU te kunnen sparren. Maar het was vooral
leuk om samen onderzoek te doen en te publiceren.
Goed, dan een ander onderdeel van dit boek. Floor, ontzettend bedankt
dat je zulke mooie foto’s en zo’n sprekende omslag voor me hebt gemaakt!
Je leek gelijk te snappen wat ik wilde, terwijl ik dat eigenlijk zelf nog niet
eens zo goed wist. En Lotte, veel dank dat je het aandurfde om te schitteren
op de voorkant van een proefschrift over criminele meisjes. Ik ben er van
overtuigd dat dit jouw onderzoekscarrière alleen maar ten goede zal komen!
Lieve vriendinnen, dankzij jullie had ik, naast promoveren, vooral
gewoon een leven! Tegelijk kon ik wel altijd bij jullie proefschrift-spuien als
dit nodig was. Humor en uitlaatklep.
Lieve pap, mam, en ook Ilbert. Veel dank voor jullie onvoorwaardelijke
vertrouwen in mij en jullie steun. Heerlijk om het gevoel te hebben dat jullie
altijd achter me staan.
Ook Peter, Bea en Ruben, bedankt voor jullie immer aanwezige
betrokkenheid en interesse.
Maar vooral veel dank aan Daan. Met jou erbij is alles zoveel
makkelijker én leuker! Zo ook het schrijven van mijn proefschrift. Fijn dat jij
goed aanvoelt wanneer je me met beide benen op de grond moet zetten of
juist even moet laten ‘fladderen’. Dank je wel.
180
CURRICULUM VITAE
Thessa Wong (16-01-1984) attended the Utrecht University from 2002 to
2007 where she earned a bachelor’s degree in pedagogics and graduated cum
laude for the research master ‘Development and Socialization in Childhood
and Adolescence’. In June 2007 she started a PhD-project at the Department
of Criminal Law and Criminology at the VU University Amsterdam. During
her PhD she also coordinated and co-authored a research project on behalf of
the Research and Documentation Centre of the Ministry of Security and
Justice, titled ‘Delinquent girls: Background characteristics, risk factors and
interventions (orginal title: ‘Delinquente meisjes: Achtergronden,
risicofactoren en interventies’). In November 2011 she started working in
the police force Amsterdam-Amstelland as a criminal investigator in
training.
181

Girl delinquency A study on sex differences in - VU-dare

Transcription

Similar documents

Workshop 1C

Georgy-Porgy

McDade Park, Scranton, PA

McDade Park, Scranton, PA

Invitation - Girls Write Now

Athletic Newsletter February 19, 2015

7:07 Attending: Matt Biscan, Matt Fisher, Melissa Nilles, Suzanne

The Rampage - Worth County Schools

FIFTH DIMENSION - University of California, San Diego

Our Group Homes - Heart For Orphans