AGGRESSIVE BEHAVIOR
Volume 31, pages 24–39 (2005)
Correlates of Aggressive and Delinquent
Conduct Problems in Adolescence
Sven Barnown, Michael Lucht, and Harald-J. Freyberger
Department of Psychiatry, Ernst Moritz Arndt University of Greifswald, Greifswald, Germany
: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
The primary purpose of the study was to test a hypothetical model for aggressive and delinquent
conduct problems in adolescence. A total of 168 adolescents from 121 families were studied using
several questionnaires and a semi-structured interview. The following factors were considered: obstetric
complications, temperament (novelty-seeking), self-esteem, family influences (perceived parenting,
alcohol abuse/dependence of parents, antisocial personality disorder of the father) and peer-group
characteristics (peer rejection and membership in a deviant peer group). The evaluation methods
applied included correlation analyses and testing of two hypothetical models using structural equation
modeling. The correlation analyses revealed significant relationships between adolescent aggressive and
delinquent behavioral problems and parental antisocial behavior; perceived parental rejection and low
emotional warmth; adolescent novelty-seeking, self-esteem, peer rejection and peer deviance. The two
empirical models, separately for aggressive and delinquent behavior problems, revealed direct
relationships between paternal antisocial behavior, parental rejection, adolescent novelty seeking, peer
deviance, peer rejection, and offspring aggression and delinquency. There were, however, two
differences with respect to the relationships between peer rejection, peer deviance, aggression and
delinquency. First, peer rejection was more strongly associated with aggressive behavior and only
moderately linked to delinquency. Second, deviance in the peer-group was found to be closely related
with delinquency but only moderately with aggression. Our findings suggest that several pathways for
aggressive and delinquent conduct problems are comparable while others were not. Regarding the
findings of the empirical models, we conclude that only intervention measures that include parents,
peers and individual adolescents may help decrease the incidence of aggressive and delinquent conduct
problems. Aggr. Behav. 31:24–39, 2005. r 2005 Wiley-Liss, Inc.
: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
Keywords: aggression; delinquency; adolescence; SEM; behavioral problems; parenting; peers; ASPD
INTRODUCTION
Externalizing symptoms are of great interest to mental health professionals because they are
predictive of criminality, substance misuse and personality disorders in adulthood, and can
cause disruptions in the family, school, and peer relations [Coie et al., 1982; Elliott, 1994].
Research has identified two general types of externalizing symptoms: overtly aggressive
n
Correspondence to: Sven Barnow, Department of Psychiatry, Ernst Moritz Arndt University of Greifswald,
Ellernholzstr. 1-2, 17487 Greifswald, Germany. E-mail: [email protected]
Received 7 April 2003; amended version accepted 8 August 2003
Published online in Wiley Interscience (www.interscience.wiley.com). DOI: 10.1002/ab.20033
r 2005 Wiley-Liss, Inc.
Behavioral Problems in Adolescence
25
behavior (e.g., fighting, temper tantrums, and cruelty) and more covert delinquent behavior
(e.g., lying, stealing, and substance abuse) [Fergusson et al., 1994; Lahey and Loeber, 1994].
Together, these behaviors comprise a higher order externalizing factor [Achenbach, 1991].
Although aggressive and delinquent behavioral problems are highly associated with each
other, findings of many studies suggest that these conduct problems represent two distinct
clinical entities, and should not be considered unidimensional [Achenbach et al., 1995;
Stanger et al., 1996, 1997]. Thus, we use aggressive and delinquent conduct problems as two
separate dependent variables in our hypothesized models.
Factors that contribute to delinquent and/or aggressive behavior are likely to be
heterogeneous. They include biological characteristics [Cloninger, 1994] and shared or
unshared environmental conditions affecting the individual [Garnefski and Diekstra, 1996].
In light of these complexities, an optimal understanding of the process through which normal
behavior develops into aggressive or delinquent behavior requires one to consider multiple
domains of influence. The purpose of this study was to test a hypothetical model for
aggressive (model 1) and delinquent behavioral problems (model 2) in adolescence using
structural equation modeling (SEM), taking into account multiple risk domains.
Obstetric complications (OCs) are among the earliest factors that can influence behavioral
development. Recent reports from two large scale general population studies have revealed
that obstetric complications were associated with the risk of externalizing disorders
[Allen et al., 1998; Raine et al., 1997]. Moffitt [1993] theorizes that OCs can produce
neuropsychological deficits in the infant’s nervous system that may lead to cognitive deficits.
These deficits have been found to be correlated with difficult temperament and subsequent
behavioral problems [Moffitt, 1990].
A second domain of interest relates to parenting characteristics. Several studies have
emphasized the potential importance of negative parenting characteristics on the development of externalizing symptoms in children [Barnow et al., 2002b; Barnow and Freyberger,
2003; Dodge et al., 1994; Loeber and Farrington, 1998; Patterson et al., 1998]. Harsh,
rejecting parenting and antisocial traits in parents are described as especially influential on
adolescents’ behavioral problems [Bank et al., 1993; Barnow, 2001; Barnow et al., 2001;
Henry et al., 1993; Patterson, 1999]. For example, subjects who perceived their parents as
rejecting and low in emotional support had more behavioral problems such as aggression and
delinquency than those who viewed their parents as supportive [Asher et al., 1990; Kumpfer
and Turner, 1990].
Another domain of interest is parental drinking behavior [Barnow et al., 2002a,b]. Many
studies have reported that adolescents who were raised in alcoholic households had higher
scores on measures of aggression and delinquency [Alterman et al., 1998; Giancola et al.,
1996; Greenfield et al., 1993]. The high incidence of co-morbidity of antisocial personality
disorders (ASPD) in alcoholic fathers, however, appears to be one variable that might
mediate parental problem drinking and the development of behavioral problems among
children of alcoholics (COAs). It is likely that the association between childhood behavioral
problems and parental ASPD is at least partially mediated through negative parenting styles
(e.g., rejection, harsh punishment, and coercive parenting) of the antisocial parents [Rutter
et al., 1998].
Several longitudinal studies have supported the role of temperamental factors, especially
early child impulsivity, in the development of externalizing problems [Bates et al., 1998; Caspi
et al., 1995]. Cloninger [1987] discussed the fact that factors of personality-related
temperament such as marked ‘‘sensation-seeking,’’ can predispose one to delinquency, and
26
Barnow et al.
rejection by parents and peers, as well as behavioral problems (e.g., substance abuse). Raine
and colleagues [1998] have reported an association between stimulation-seeking at age three
and aggression at age eleven. Predictive associations have been modest, and most studies
have been based solely on maternal reports of infant/toddler temperament.
A fourth characteristic refers to how a child perceives his or her self-worth. The level of
perceived social support and rejection by parents and peers has been reported to relate to a
teenager’s self esteem [Brutseart, 1990; Sandler et al., 1989]. Prospective studies that
examined the associations between rejection by peers and parents, self-esteem, and behavioral
problems reported that rejection by peers and parents appeared to be an antecedent of
aggressive and delinquent conduct problems, while low self-esteem was indirectly related,
through a higher probability of rejection by others, to childhood behavioral problems
[Sandler et al., 1989].
Finally, the quality of relationships with peers is often cited as a correlate of problematic
behavior of children. Specifically, instability in peer relationships has been implicated in the
etiology of aggressive behavior [Hymel et al., 1990; Coie and Cillessen, 1993]. Children who
are unable to form peer relationships are not only at risk for later delinquency and substance
use, but are also more likely to associate with other deviant peers [Dishion, 2000].
In this study, we examined the interrelationships of individual risk factors (children’s
temperament and self-esteem), family environment (parenting styles, parents’ mental
disorders) and peer relationships (peer rejection and membership in a deviant peer group)
with aggressive and delinquent behavior problems among 168 adolescent in Germany. The
model is based upon the theoretical supposition that a positive family history of alcoholism
(FH+) is related to an additional psychiatric diagnosis of antisocial personality disorder
(ASPD) in the father and vice-versa. Furthermore, we hypothesized a potential direct
influence of parental alcoholism and/or presence of ASPD of the father on both child’s
temperament, as indexed by the extent of novelty-seeking, and the parenting styles of
rejection and emotional warmth. Furthermore, paternal ASPD is thought to be directly
associated with aggression, as well as delinquency, in offspring. Obstetric complications are
hypothesized to be related to the child’s rating on the subscale of novelty seeking (NS), while
NS is assumed to be associated with perceived parenting styles, peer rejection and with
aggression or delinquency. Parental characteristics such as perceived rejection and emotional
warmth are viewed as associated with both self-esteem and aggressive/delinquent behavior.
Perceived self-worth is considered to be related to a higher probability of peer rejection, while
rejection by peers is hypothesized as likely to increase the probability of association with
deviant peers, and to heighten the risk of adolescent aggression/delinquency. Finally, we
assume that membership in a deviant peer group is related to aggressive, as well as,
delinquent behavior. The hypothetical model is shown in Figure 1. The model was then tested
separately for aggressive and delinquent conduct problems.
METHODS
Subjects
The data were derived from the population-based Study of Health in Pomerania, Germany
[SHIP; John et al., 2001]. In the SHIP, 3748 subjects aged 20 to 79 were drawn at random
between March 1997 and May 2000, proportional to the population size of each community,
and stratified by age and gender. From this population, 527 participants were chosen
Behavioral Problems in Adolescence
27
reflecting the presence of one or more biological offspring between 12 and 18 years. Among
these, 141 persons could not be located or did not respond to at least four phone calls and
two letters, and at least one member of 75 families refused to participate, resulting in 311
(69%) family groups who gave informed consent. By October, 31, 2001, 122 of these families
had been investigated. Because of missing data on at least two measures, 21 cases were
excluded from the current study, leaving 104 families with 145 adolescents. There were no
statistically significant differences between the 104 families included in the current study and
the 423 nonparticipants in terms of the parents’ education (p = 0.64), monthly income
(p = 0.32), and the number of children in the household (p = 0.27). Regarding marital status,
a greater proportion of the parents in the investigated sample were married (87.5% vs.
75.3%, C2 = 13.65, df = 3, po0.05).
In an effort to increase the number of FH+ adolescents, seventeen additional families with
23 adolescents were recruited from outpatient addiction facilities in Pomerania. For this
subset of the sample, several outpatient facilities in Pomerania were invited to refer families in
which there was at least one biological parent with a clinical diagnosis of an AUD, where at
least one biological offspring between the ages of 12 and 18 lived in the household, and who
gave informed consent to take part in the study. A comparison of these 23 adolescents with
the 145 who were recruited from the SHIP revealed no differences in terms of age (14.4 vs.
14.8, t = 0.74, df = 166, p = 0.46), while there was a trend for a higher proportion of those
23 adolescents to be a female (males: 51.0% vs. 30.4%, w2 = 3.37, df = 1, p = 0.066). The
average age of the entire sample (N = 168 adolescents who came from 121 families) was 14.5
(SD = 2.10) years, with similar proportions of males (48.2%) and females. All participants
were Caucasians who lived in Pomerania (Germany).
Measurements
The semi-structured interview, the SSAGA [Bucholz et al., 1994], was used to determine
the adolescents’ family history status of alcoholism. Cross-center studies of the SSAGA
indicate good reliability regarding alcohol-use disorders, with test-retest agreement (Kappas)
for the DSM-III-R diagnoses of alcohol dependence between 0.87–0.89, and between 0.57–
0.74 for abuse (the range reflects the results of different studies). Using the SSAGA, 64
adolescents (FH+ sample) with at least one parent with alcohol abuse or dependence were
identified from the total of 168 offspring.
Antisocial personality disorder of parents and adolescents older than 15 years was assessed
with the Structured Clinical Interview SCID-II [First et al., 1997]. A value of more than three
in the self-rating section of the interview indicates a tendency toward antisocial personality
disorder. Because no mother had a value higher/equal 3 in the self rating section, only
paternal antisocial personality disorder was considered in this study. The SCID-II has
Kappas between 0.51 and 0.68 for lifetime diagnoses [Williams et al., 1992]. For children age
fifteen and younger, the Diagnostic Interview for Psychiatric Disorders [Children-DIPS;
Unnewehr, 1995] was used to determine the ICD-10 [World Health Organization, WHO,
1991] diagnosis (F91.x) of oppositional defiant disorder (ODD) or conduct disorder (CD).
Obstetric complications (OCs) were assessed using three items gathered from an interview
[Raine et al., 1997] with the mother: a gestational period of less than 37 weeks, placement of
the infant in an incubator after birth, and low birth weight (o2500g). These items were
selected due to their importance for later brain development [Cohen et al., 1989].
28
Barnow et al.
The OCs score was determined from the z score transformation of these three items
(a = 0.76).
Novelty-seeking was assessed with the 9th version of the Temperament and Character
Inventory [TCI; Cloninger et al., 1998]. This inventory measures seven dimensions of
personality: four temperament and three character dimensions. The temperament scale of
novelty-seeking (NS; 40 items) includes 4 lower-order subscales (e.g., excitability,
impulsiveness vs. thoughtfulness, extravagance vs. simplicity). The validity of the
questionnaire was confirmed through various studies. In the analyses, the respective raw
values of the entire scale of NS were used. In the age group of 12–15 years, a previously
unpublished, abbreviated version of the TCI (jTCI) was used. Z-score transformation of the
scales of NS in the jTCI and TCI were combined to obtain one value for the different age
groups.
Self-esteem of adolescents was determined using the Rosenberg Self Worth Scale
[Rosenberg, 1985]. This scale contains 10 true/false items for evaluating global self-esteem.
Test-retest reliability and internal consistency of the scale have been reported as satisfactory,
with r = 0.85 and a = 0.88 (in our study a = 0.79).
Parenting styles were examined with the short EMBU [Swedish acronym for ‘‘own
memories concerning upbringing’’; Arrindell et al., 2001], which was completed by all
offspring. All questions were evaluated separately regarding the father and mother using a
four-point Likert-type scale (0 = does not apply, 3 = very often). For the purpose of this
study, rejection (7 items) and emotional warmth (6 items) were used in the analysis as
reflecting abusing and neglectful parental rearing patterns. The Cronbachs alphas are 0.70 for
rejection and 0.84 for emotional warmth.
The Youth Self Report [YSR; Achenbach, 1994], is a 119-item checklist in which
adolescents use a three-point scale to rate how well a series of behavioral problems
describe their own behavior. Second-order principal factor analyses have revealed two
broadband groupings of the syndromes, labeled emotional problems and behavioral
problems. To determine the extent of behavioral problems, we used the subscales of
aggression, which consists of 19 items (e.g., destroys things, temper tantrums, physically
attacks people) (alpha = 0.82 and 0.84 for boys and girls, respectively); and delinquency,
which consist of 11 items (e.g., lying or cheating, running away from home, stealing at home
or outside, using alcohol or drugs) (alpha = 0.70 and 0.77), Additionally, the similar Child
Behavior Checklist [CBCL; Achenbach, 1991] was filled out by the mother. The correlations
between mothers’ (CBCL) and children’s (YSR) ratings in the subscales of aggression and
delinquency were only moderate, with a Pearson correlation coefficient (Pearson) of
r = 0.293 (po0.001) for aggressive conduct problem and r = 0.399 (po0.001) for
delinquency.
Peer deviance was assessed by a questionnaire developed by our research group. To
determine the degree of deviant behaviors within the peer group, adolescents were asked
three questions in true/false format: 1) the use of alcohol/drug at least once per week among
friends; 2) the use of violence (fighting) to solve problems in the group; and 3) the presence of
members who committed crime and/or had contact with the police because of deviant
behavior (a = 0.50).
To determine the extent of peer rejection, we selected three items from the CBCL that dealt
specifically with whether the parents thought their child had problems with peers (doesn’t
get along with others, gets teased a lot, not liked by other children). Additionally, the similar
items in the YSR (here, the answers were given by the child) were used. The construct of peer
Behavioral Problems in Adolescence
29
rejection was measured by the combination of z-score transformed values from the three
CBCL and three YSR items (a = 0.70).
Data Analysis
In the first step, frequencies of categorical data and mean, standard deviations, skewing,
and probabilities for Kolmogorov-Smirnov tests of numeric variables were calculated. In an
effort to understand how the individual risk factors are related to each other, in the second
step, an intercorrelation matrix was computed for all the variables.
The hypothetical model, as described in the introduction and shown in Figure 1, was
developed, in the third step. Measures on aggression and delinquency according to mothers’
and children’s ratings in the CBCL and YSR were used as the central dependent variables. To
include information from both informants (mother and adolescent), a measurement model
was used. This was necessary because the correlation between mothers’ (CBCL) and
children’s (YSR) ratings in the scales of aggression and delinquency were only moderate (see
above). Thus, aggressive, as well as delinquent conduct problems, were treated as unobserved
variables that are indirectly measured by the two corresponding subscales in the CBCL and
the YSR. The set of connections between the observed and unobserved variables (the latter
are displayed as ellipses) is called the measurement model.
The hypothetical model was then tested using the Amos Structural Equation Program
[Amos 4.01; Arbuckle, 1999], which is based on the Maximum Likelihood (ML) method for
analysis of the variance/covariance matrix, using x2 for the goodness of fit statistic. While the
traditional power tables are not directly applicable to structural modeling, Ding and
colleagues [1995] and Boomsma [1987] showed that as few as 100 participants can be
acceptable, and the recommended ratio of sample size to free parameters is generally accepted
as between 5:1 to 10:1 [Bentler and Chou, 1987]. Due to the small sample size, all indicators
except aggressive and delinquent behavioral problems were treated as manifest. To determine
Fig. 1. Hypothetical model of aggressive and delinquent conduct problems in adolescence.
30
Barnow et al.
the regression model, the individual pathway coefficients of the error variables were
set at 1, while the significance level of the individual pathways was set at po0.05. The
overall model fit was assessed by examining the Nonnormed Fit Index [NNFI; Bentler and
Chou, 1995] and the Comparative Fit Index [CFI; Bentler, 1990], as well as the
Residual Mean Squared Error Approximation (RMSEA) Index [Hu and Bentler,
1998]. The NNFI and the CFI have typical values between 0 and 1, with a value close to
1 indicating a very good model fit. However, scores close to zero in the RMSEA denote a
better fit.
RESULTS
Psychosocial characteristics, percentages of ASPD in children and fathers and percentages
of parental AUDs are shown in Table I.
Regarding numeric variables, the distributions of scores for peer rejection, peer deviance,
and self-esteem were skewed, while other factors followed the normal distribution. To
evaluate this further, we calculated several Kolmogorov-Smirnov tests revealing that all but
the above-mentioned variables were normal distributed. Our efforts to normalize these
variables by submitting them to a log transformation were not successful. In order to examine
the empirical impact of the observed degree of non-normality, nonparametric relationships
(Spearman rank ordered correlations) were examined and compared with parametric
correlations as displayed in Table II. These analyses revealed very similar results such that the
use of parametric tests could be carried out. It should be kept in mind, however, that the
results of the empirical path models (see below) must be discussed with caution and cannot
automatically lead to asymptotic conclusions (conclusions that are approximately true for
large samples) [Arbuckle, 1999].
Table II describes the zero-order correlations among the twelve risk factors with aggression
and delinquency. Seven factors correlated significantly with delinquency as well as with
aggression, including parents’ and child’s ASPD, peer rejection, peer deviance, parental
rejection and emotional warmth, novelty-seeking, and self-esteem. Age and gender were only
TABLE I. Characteristics of the Sample
factors
age (mean/SD)
gender: % of males
Children per household (mean/SD)
% of living with:
both parents
biological mother only
% of Family History of alcoholism
% of child’s ASPD (age415), OD or CD (age 11–15)
% of paternal ASPDn
obstetric complications
% of low birth weight (o1500g)
% of need incubator
% of short pregnancy
n
scores 3 and higher in the SCID self rating section.
% (mean/SD for continoues variables)
14.48/2.10
48.2
1.38 (range 1–3)
87.5
12.5
38.1
7.7
10.7
7.7
1.8
3.0
age
1.000
gender
FH+
ASPD of parents
obstetric complications
child’s ASPD (age 415),
OD or CD (age 11–15)
peer characteristics
peer rejection (z-score)
peer deviance
parenting styles
rejection
emotional warmth
individual
characteristics
novelty seeking (z-score)
self-esteem
behavior problems YSR,
CBCL combined (z-score)
delinquency
aggression
age
.052
1.000
.071
.193n
1.000
.089
.166n
.243nn
1.000
.021nn
.092
.060
.105
1.000
.221nn
.077
.044
.044
.096
1.000
gender F+++ ASPD
obstetric
child’s ASPD
of father complications (age 415),
OD or CD
(age 11–15)
TABLE II. Intercorrelation Matrix of Included Variables
parenting
styles
individual
characteristics
behavior
problems
1.000
.108
.050
.017
.033
.021
.025
.053
1.000
.379nn
.165n
.062
.093
.047
.236nn
.140
.110
.277nn
1.000
1.000
.131
.015
.120
.135
.068
.199nn
.145
.089
.131
.063
.057
.011
.050
.168n
1.000
.130
.018
.034
.003
.049
.015
.008
.111
.198n
.106
.103
1.000
.211nn
.248nn
.223nn
.032
.074
106
.073
.007
.060
.039
.175n
.242
.616nn
1.000
.236nn
.209nn
1.000
.616nn
.372nn
.168n
.359nn
.110
.179n
.309nn
.459nn
.246nn
.037
.014
.075
.178n
.037
.340nn
.219nn
.207nn
.027
.188n
.019
.428nn
peer
peer
rejection Emotional novelty self-esteem delinquency aggression
rejection deviance
warmth
seeking
(CBCL,
(CBCL,
z-score)
(z-score)
YSR)
YSR)
peer
characteristics
Behavioral Problems in Adolescence
31
32
Barnow et al.
Fig. 2. Empirical path model 1; pathways toward aggression in adolescents (N = 168). Only paths significant at po0.05
are shown. npo0.05; nnpo0.01; CBCL: Child Behaviour Checklist; YSR: Youth Self Report.
significantly related to delinquent conduct problems. The directions of the relationships
among variables were consistent with what we hypothesized in Figure 1.
The test of the hypothesized model for aggression (model 1) resulted in the empirical values
presented in Figure 2. Here, only vectors with standardized regression coefficients that were
significant (po0.05) are shown. Overall, the goodness of fit of this model was good (x2
= 46.993, df = 30, p = 0.056, with a x2 df ratio of 1.566; NNFI = 0.978; CFI = 0.990; and
RMSEA = 0.058). The squared multiple correlations of the endogenous variable ‘‘aggression’’ was R2 = 0.62. Results indicated that parental ASPD, NS, parental rejection, peer
rejection and peer deviance were each significantly related to increased aggression.
Furthermore, OCs were associated with NS. Perceived feelings of emotional warmth of
parents were positively related to self-esteem, while parental rejection was negatively
associated with self-esteem. Children’s reports of lower self-esteem were related to elevated
peer rejection.
The model using delinquency as the central dependent variable (model 2) also fits the data
quite well (x2 = 40.399, df = 30, p = 0.097, NNFI = 0.994, CFI = 0.994, RMSEA = 0.046)
and the squared multiple correlations of the endogenous variable was R2 = 0.62. These
values did not differ greatly from model 1. Considering the magnitude of standardized
regression coefficients, the relation between peer rejection and delinquency decreased from
beta = 0.48 in model 1 to beta = .17 in model 2. Second, the path coefficient from peer
Behavioral Problems in Adolescence
33
Fig. 3. Empirical path model 2; pathways toward aggression in adolescents (N = 168). Only paths significant at po0.05
are shown. npo0.05; nnpo0.01; CBCL: Child Behaviour Checklist; YSR: Youth Self Report.
deviance to delinquency in model 2 was markedly higher compared to the path coefficient of
the relationship between peer deviance and aggression (.16 vs. .38). The empirical values of
model 2 are shown in Figure 3.
Additional iterations of the empirical models were tested using age and gender as
covariates. However, neither age nor gender were significantly associated with aggression and
delinquency, and the inclusion of both covariates in the two models yielded only small
changes for the estimation of the magnitudes of paths. Specifically, when age and gender were
added to model 1, the R2 only increased to 0.65, while the inclusion of age and gender in
model 2 yielded an R2 of 0.62, which was similar to that what we found in model 1.
In order to address the concern that some subjects came from addiction outpatient
facilities, both model were re-evaluated after excluding these 23 youths. Using the 145 SHIP
adolescents, all coefficients remained stable and the results were almost identical (not shown)
to those reported in Figures 2 and 3, suggesting that the overrepresentation of children with
FH+ did not have an impact on the empirical models described above.
DISCUSSION
The goal of this study was to identify complex interrelationships between various different
risk domains and aggressive, as well as, delinquent conduct problems in adolescence. One
34
Barnow et al.
unanticipated finding was that gender and age were significantly correlated with delinquency,
but not with aggression. Although many studies reported more behavioral problems in boys,
the evidence about gender differences is conflicting. Behavioral problems have generally been
found to be more common in boys, other evidence suggests, however, that girls are simply
less likely than boys to exhibit serious conduct problems, and that the predictors of conduct
problems are similar for boys and girls [Ackermann et al., 2003; Ferguson and Horwood,
2002; Lahey et al., 1999; Moffitt et al., 2000]. Regarding age, there is a growing evidence that
after approximately age 10, aggressive syndrome scores decline, whereas delinquent
syndrome scores increase until about age 17 [Stanger et al., 1997] which support our finding
in that only delinquent conduct problems were significantly related to age.
One strength of our study was that we distinguished between aggressive and delinquent
behavioral problems using both kinds of behavior as central dependent variables in two
separate models. Although, a number of the relationships predicted in the two models were
confirmed for aggressive, as well as, delinquent behavior, several assumptions proved to be
incorrect. In the first step, we hypothesized that the presence of a paternal antisocial
personality disorder (ASPD) and/or alcohol-abuse disorder (AUD) was significantly related
to parental behavior perceived by the adolescent as negative (rejection or low emotional
warmth) (see Figure 1). This assumption was not confirmed by the empirical analysis of both
models. The data available on this matter is a focus of controversy, however [Johnson et al.,
2001]. For example, several studies found that the relationship between parental ASPD and
behavioral problems, such as aggressive and delinquent conduct problems in offspring, are
mediated primarily by genetic mechanisms and that parenting behavior appears to be
relatively insignificant as a mediator [Slutske et al., 1998]. It should also be noted that there
was insufficient power due to the small size of the sample, and, thus, only moderate effects
could be shown. Thus, the standardized regression coefficient for the path between the factors
paternal ASPD and emotional warmth as perceived by children was beta =
0.12,
confirming the direction indicated by the prediction, but failing to achieve significance.
Our second hypothesis that diagnosed paternal ASPD is directly associated with the degree
of aggression and delinquency in these adolescents was confirmed both in the correlation
studies and in empirical testing of the models. This supports our above-mentioned
assumption that paternal ASPD tends to have a direct influence on subsequent aggressive
and delinquent behavior in offspring [Barnow et al., unpublished data], although we cannot
make reliable statements about the mediation process.
The third hypothesis in the model assumed a causal relationship between OCs and
adolescent impulsivity as indexed with the extent of novelty seeking. This hypothesis was
based on the findings of Moffitt [1990, 1993], which suggest that OCs leading to brain
damage can result in a decreased capacity to adapt to social situations and lower self-control
(impulsivity). The empirical test of the hypothetical model supported these findings to the
extent that a relationship—although only a moderate one—was found between Ocs and the
degree of NS in the model.
For our fourth hypothesis, we assumed in the hypothetical models that a temperament
inclined toward impulsiveness was associated with stronger feelings of parental and peer
rejection, and directly related to aggressive or delinquent behavior problems. While the
hypothesis of a direct relationship between NS and adolescent aggressive and delinquent
behavior, respectively, was confirmed, no significant association between NS and parental-,
and peer rejection was found. Regarding the latter, our results are consistent with findings,
however, showing that child behavior tends to have only a moderate influence on parenting
Behavioral Problems in Adolescence
35
behavior [Campbell et al., 1996]. The confirmed direct relationship between NS and
aggression and delinquency agrees with a number of studies that showed that impulsive
children were less capable of controlling their behavior, which lead to a stronger propensity
to develop aggressive and delinquent conduct problems later on [Chassin et al., 1993;
Cloninger et al., 1988; Posner and Rothbart, 1998].
The fifth hypothesis in our model assumed a relationship between parenting styles, and
aggressive and delinquent conduct problems, respectively. For example, there is strong
evidence for relations between coercive parenting and the externalizing problems of boys
[Patterson et al., 1998]. Our findings showed, however, that perceived parental rejection was
associated with aggression and delinquency in offspring only, whereas neither aggression nor
delinquency was significantly correlated with perceived parental emotional warmth. Due to
the design of the study (use of cross-sectional data), however, it was not possible to
demonstrate causal relationships. Thus, the alternative hypothesis that children with
behavioral problems tend to feel rejected by their parents cannot be ruled out. It is also
important to consider that the expected protective effect of parenting styles characterized by
emotional warmth was confirmed in the correlation analyses (see Table II). This relationship
was, however, only moderate. It should be noted that these same parental characteristics have
also been found to be associated with other forms of psychopathology in children
(e.g., phobias and social anxiety, obsessive-compulsive disorders and emotional problems).
Thus, it is likely that such negative parental child-rearing practices may make children
vulnerable to psychopathology in general, rather than to one or a few specific forms of
psychopathology.
Our assumptions regarding parenting behavior and feelings of self-worth in adolescents
were confirmed in both the correlation analyses and in the empirical models. While the
perception of a parenting style characterized by rejection was negatively associated with the
assessment of self-worth, perceived parental emotional support strengthened feelings of selfworth. Buikhuisen [1988] argued that it is likely that parental rejection impedes the
development of a conscience, decreases feelings of empathy, and lowers the child’s selfesteem; while perceived emotional support increases feelings of self-worth due to positive
feedback from parents [Cohen et al., 1994]. One finding also showed that adolescents,
who had a low self-esteem, were more frequently rejected by their peers. This is consistent
with that insecure children who tend to be more anxious are more likely to be rejected by
members of the peer group and are generally less popular [Cloninger, 1987]. This illustrates
the necessity of considering complex patterns of interaction in order to develop an
understanding of how behavioral problems develop. The absence of a direct relationship
between feelings of self-worth and aggressive or delinquent conduct problems, however, must
not lead to the conclusion that low self-esteem plays no role in the development of conduct
problems.
Our last hypothesis was based on findings that confirm the influence of the peer group on
adolescent personality and behavior [Dishion et al., 1995; Fergusson et al., 1995]. We
predicted that peer rejection may lead directly to both aggressive and delinquent behavior by
prompting rejected individuals to seek out deviant, like-minded peers. This assumption was,
however, not supported by our data. Moreover, the separate testing of the theoretical models
for aggression and delinquency revealed some differences in the magnitudes of the individual
path coefficients with respect to the relationships between peer deviance, peer rejection, and
aggression and delinquency. Although peer rejection and membership in a deviant peer group
correlated positively with the extent of aggressive and delinquent behavior problems, peer
36
Barnow et al.
deviance was more strongly associated with delinquency, while peer rejection was more
tightly linked to aggression. Regarding the association between peer rejection and aggression,
there has been a growing recognition of the contribution made by peer relations in
forecasting subsequent aggressive conduct problems. For example, Coie and colleagues
[1992] showed that both childhood aggression and peer rejection independently predicted
subsequent aggressive behavior. Bierman and Wargo [1995] found in their longitudinal study
that boys who were both rejected and aggressive continued to exhibit higher levels of
aggression than did boys who were aggressive but not rejected [see also Miller-Johnson et al.,
2002]. Furthermore, rejected children can be subjected to negative experiences that might
shape their attitudes toward other peers in a way that amplifies aggressive behavior. On the
other hand, the influence of group membership in a deviant peer group on adolescent
delinquent behavior problems identified in the study may be explained by considering such
processes as overt group pressure, peer modeling, and adaptation to group attitudes. These
factors can lead to increased delinquent conduct problems in each individual group member
in groups in which the behaviors are regarded as positive, which is often the case in deviant
peer groups.
In summary, the results of our study point to the significance of child personality traits,
family environment and group influences on aggressive and delinquent adolescent behavior.
It is possible to derive approaches to prevention and intervention from these findings. First, it
is likely that only intervention and prevention strategies that integrate multiple areas of life
(shared as well as unshared environment) can effectively reduce behavioral problems such as
aggression and delinquency. Secondly, failure to distinguish between aggressive and
delinquent conduct problems could limit the generalizablity of results by neglecting different
developmental courses of aggression and delinquency.
Finally, several limitations of the study should be mentioned. First, cause-effect
relationships could not be determined from these cross-sectional data. Prospective
examinations, however, are underway and might be more helpful in evaluating the proposed
model. Second, relatively few of the total of 317 potential family groups were available for
the current analyses. Thus, the sample of this study was relatively small considering the size
of the model being evaluated. Our results should, therefore, be considered as preliminary. It is
also likely that only medium or higher effects could be identified. Third, because there were
multiple children in some families, the observations were not fully independent. Fourth, some
data were collected using self-reporting questionnaires. Responses to questions about
problems and social support may be biased by individuals’ willingness to self-disclose their
feelings, selective recall, and their desire to present themselves in a socially desirable way.
Fifth, the extent to which the findings can be generalized beyond the scope of our sample
remains unclear.
ACKNOWLEDGMENT
The work is part of the Community Medicine Research net (CMR) of the University of
Greifswald, Germany, which is funded by the Federal Ministry of Education and Research
(grant no. ZZ9603), the Ministry of Cultural Affairs as well as the Social Ministry of the
Federal State of Mecklenburg-West Pomerania. The CMR gathers several research projects
which are sharing data of the population-based Study of Health in Pomerania (SHIP; http://
www.medizin.uni-greifswald.de/cm).
Behavioral Problems in Adolescence
37
REFERENCES
Achenbach TM. 1991. Integrative guide for the 1991
CBCL/4-18, YSR, and TRF. Burlington: University
of Vermont, Department of Psychiatry.
Achenbach TM. 1994. Child Behavior Checklist and
related instruments. In: Maruish M, editor. The use
of psychological testing for treatment planning and
outcome assessmen. Hillsdale: Lawrence Erlbaum
Associates. p. 517–549.
Achenbach TM, Howell CT, McConaughy SH, Stanger
C. 1995. Six-year course of behavioral/emotional
problems in a national sample of children and youth:
I. Predictors of syndromes. J Am Acad Child
Adolesc Psychtr 34:336–347.
Ackermann BP, Brown E, Izard CE. 2003. Continuity
and change in levels of externalization behavior in
school of children from economically disadvantaged
families. Child Dev 74: 694–716.
Allen NB, Lewinsohn PM, Seeley JR. 1998. Prenatal and
perinatal influences on risk for psychopathology in
childhood and adolescence. Devl Psychopathol
10:513–529.
Alterman AI, Bedrick J, Cacciola JS, Rutherford MJ,
Searles JS, McKay JR, Cook TG. 1998. Personality
pathology and drinking in young men at high and
low familial risk for alcoholism. J Stud Alcohol
59:495–502.
Arbuckle JL, Wothke W. 1999. AMOS 4.1 Users Guide
(Handbook). Chicago: Small Waters Corporation.
Arrindell WA, Richter J, Eisemann M, Garling T,
Ryden O, Hansson SB, Kasielke E, Frindte W,
Gillholm R, Gustafsson M. 2001. The short-EMBU
in East-Germany and Sweden: a cross-national
factorial validity extension. Scand J Psychol
42:157–160.
Asher SR, Parkhurst JT, Hymel S, Williams GA. 1990.
Peer rejection and loneliness in childhood. In:
Asher S, editor. Peer rejection in childhood.
Cambridge studies in social and emotional development. New York: Cambridge University Press.
p. 253–273.
Bank L, Forgatch M, Patterson G, Fetrow R. 1993.
Parenting practices of single mothers. J Marriage
Family 55:371–384.
Barnow S. 2001. Aggressives Verhalten im Jugendalter:
Empirische Befunde zu familialen Einflussfaktoren.
In: Mauthe J, editor. Prävention in Psychiatrie
und Psychotherapie. Königslutter: Axept Verlag.
p. 51–81.
Barnow S, Freyberger HJ. 2003. The family environment in early life and aggressive behavior in
adolescents and young adults. In: Mattson MP
editor. Neurobiology of Aggression: Understanding
and Preventing Violence. Totowa, NJ: Humana
Press Inc.p. 213–230.
Barnow S, Lucht M, Freyberger HJ. 2001. Influence of
punishment, emotional rejection, child abuse, and
broken home on aggression in adolescence: an
examination of aggressive adolescents in Germany.
Psychopathol 34:167–173.
Barnow S, Schuckit MA, Lucht M, John U, Freyberger
HJ. 2002a. The importance of a positive family
history of alcoholism, parental rejection and emotional warmth, behavioral problems and peer substance use for alcohol problems in teenagers: a path
analysis. J Stud Alcohol 63:305–315.
Barnow S, Schuckit M, Smith TL, Preuss U, Danko G.
2002b. The relationship between the family density
of alcoholism and externalizing symptoms among
146 children. Alcohol Alcohol 37:383–387.
Bates JE, Pettit GS, Dodge KA, Ridge B. 1998.
Interaction of temperamental resistance to control
and restrictive parenting in the development of
externalizing behavior. Dev Psychol 34:982–995.
Bentler PM. 1990. Comparative fit indexes in structural
models. Psychol Bul 107:238–246.
Bentler PM, Chou C. 1987. Practical issues in structural
modeling. Sociol Methods Res 16:78–117.
Bentler PM, Chou CP. 1995. Estimates and tests in
structural equation modeling. SEM. Conc Issues
Appl 1:37–55.
Bierman KL, Wargo JB. 1995. Predicting the longitudinal course of associated with aggressive—
rejected, aggressive (nonrejected) and rejected (nonaggressive) status. Dev Psychopathol 7:669–682.
Boomsma A. 1987. The robustness of maximum
likelihood estimation in structural equation
models. In: Cuttance P, Ecob R, editors. Structural
Modeling by Example: Applications in Educational,
Sociol Behav Res. Cambridge: University Press.
p. 160–198.
Brutseart H. 1990. Changing sources of self esteem
among girls and boys in secondary schools. Urban
Education 24:432–438.
Bucholz KK, Cadoret R, Cloninger CR, Dinwiddie SH,
Hesselbrock VM, Nurnberger JI, Reich T, Schmidt I,
Schuckit MA. 1994. A new, semi-structured psychiatric interview for use in genetic linkage studies: a
report on the reliability of the SSAGA. J Stud
Alcohol 55:149–158.
Buikhuisen W. 1988. Chronic juvenile delinquency: a
theory. In: Buikhuisen W, Mednick SA, editors.
Explaining criminal behavior. Leiden: Netherlands:
E. J. Brill.
Campbell SB, Pierce EW, Moore G, Marakovitz S,
Newby K. 1996. Boys’ externalizing problems at
elementary school age: pathways from early behavior
problems, maternal control, and family stress. Dev
Psychopathol 8:701–719.
38
Barnow et al.
Caspi A, Henry B, McGee RO, Moffitt TE, Silva PA.
1995. Temperamental origins of child and adolescent
behavior problems: from age three to age fifteen.
Child Develop 66:55–68.
Chassin L, Pillow DR, Curran PJ, Molina BS, Barrera
M Jr. 1993. Relation of parental alcoholism to early
adolescent substance use: A test of three mediating
mechanisms. J Abnorm Psychol 102:3–19.
Cloninger C. 1987. Neurogenetic adaptive mechanisms.
Science 236:410–416.
Cloninger CR. 1994. Temperament and personality.
Curr Opinion Neurobiol 4:266–273.
Cloninger CR, Sigvardsson S, Bohman M. 1988. Childhood personality predicts alcohol abuse in young
adults. Alcohol Clin Exp Res 12:494–505.
Cloninger CR, Bayon C, Svrakic DM. 1998. Measurement of temperament and character in mood
disorders: a model of fundamental states as personality types. J Affect Dis 51:21–32.
Cohen DA, Richardson J, LaBree L. 1994. Parenting
behaviors and the onset of smoking and alcohol use:
a longitudinal study. Pediatrics 94:368–375.
Cohen P, Velez CN, Brook J, Smith J. 1989. Mechanisms of the relation between perinatal problems,
early childhood illness, and psychopathology in late
childhood and adolescence. Child Develop 60:701–
709.
Coie J, Cillessen A. 1993. Peer rejection: Origins and
effects on children’s development. Curr Direct
Psychol Science:89–92.
Coie J, Dodge K, Coppotelli H. 1982. Dimensions and
types of social status: A cross-age perspective. Dev
Psychol 18:557–570.
Coie JD, Lochmann JE, Tery E, Hyman C. 1992.
Predicting early adolescent disorder from childhood
aggression and peer rejection. J Consult Clin Psychol
60:783–792.
Ding L, Velicer WF, Harlow LL. 1995. Effects of
estimation methods, number of indicators per factor,
and improper solutions on structural equation
modeling fit indices. Struct Equation Mod 2:119–
143.
Dishion TJ. 2000. Cross-setting consistency in early
adolescent psychopathology: deviant friendships and
problem behavior sequelae. J Pers 68:1109–1126.
Dishion TJ, Andrews DW, Crosby L. 1995. Antisocial
boys and their friends in early adolescence: relationship characteristics, quality, and interactional process. Child Develop 66:139–151.
Dodge KA, Pettit GS, Bates JE. 1994. Socialization
mediators of the relation between socioeconomic
status and child conduct problems. Child Develop
65:649–665.
Elliott D. 1994. Serious violent offenders: onset,
developmental course and termination. Criminol
32:1–21.
Fergusson DM, Horwood LJ. 2002. Male and female
offending trajectories. Dev Psychopathol 14:
159–178.
Fergusson DM, Horwood LJ, Lynskey MT. 1994.
Structure of DSM-R criteria for disruptive childhood
behaviors: Confirmatory factor models. J Am Child
Adolesc Psychiatr 33:1145–1155.
Fergusson DM, Lynskey MT, Horwood LJ. 1995.
The role of peer affiliations, social, family and
individual factors in continuities in cigarette smoking
between childhood and adolescence. Addict 90:
647–659.
First M, Spitzer RL, Gibbon M, Williams JBW. 1997.
Structured Clinical Interview for DSM-IV Axis I
Disorders, Research Version, Patient/ Non-patient
Edition (SCID-I/P). New York: Biometrics Research; New York State Psychiatric Institute.
Garnefski N, Diekstra RF. 1996. Perceived social
support from family, school, and peers: relationship
with emotional and behavioral problems among
adolescents. J Am Acad Child Adol Psychiatr
35:1657–1664.
Giancola P, Moss H, Martin C, Kirisci L, Tarter R.
1996. Executive cognitive functioning predicts reactive aggression in boys at high risk for substance
abuse: A prospective study. Alcohol Clin Exp Res
20:740–744.
Greenfield SF, Swartz MS, Landerman LR, George LK.
1993. Long-term psychosocial effects of childhood
exposure to parental problem drinking. Am
J Psychiatr 150:608–613.
Henry B, Moffitt T, Robins L, Earls F, Sylvia P. 1993.
Early family predictors of child and adolescent
antisocial behavior: who are the mothers of delinquents? Crim Behav Ment Health 3:97–118.
Hu LT, Bentler PM. 1998. Fit indices in covariance
structure analysis: Sensitivity to underparametersized model misspecification. Psychol Methods
4:424–453.
Hymel S, Rubin KH, Rowden L, Lemare L. 1990.
Children’s peer relationships: Longitudinal prediction of internalizing and externalizing problems from
middle to late childhoot. Child Develop 61:2004–
2021.
John U, Greiner B, Hensel E, Lüdemann J, Piek M,
Sauer S, Adam C, Born G, Alte D, Greiser E, Härtel
U, Hense HW, Haerting J, Willich S, Kessler C.
2001. Study of Health in Pomerania (SHIP): A
health examination survey in an east German region.
Objectives and design. Sozial und Präventivmed
46:186–194.
Johnson JG, Cohen P, Kasen S, Smailes E, Brook JS.
2001. Association of maladaptive parental behavior
with psychiatric disorder among parents and their
offspring. Archives of General Psychiatry 58:
453–460.
Behavioral Problems in Adolescence
Kumpfer KL, Turner CW. 1990. The social ecology
model of adolescent substance abuse: implications
for prevention. Int J Addict 25:435–463.
Lahey BB, Loeber R. 1994. Framework for developmental model of Oppositional Defiant Disorder and
Conduct Disorder. In: Routh D editor. Disruptive
behaviors in childhood. New York: Plenum
213–230.
Lahey BB, Goodman SH, Waldman ID, Bird H, Canino
G, Jensen P. 1999. Relation of age of onset to the
type on and severity of child and adolescent conduct
problems. Journal of Abnormal Psychology 27:
247–260.
Loeber R, Farrington DP. 1998. Serious and violent
juvenile offenders. Thousands Oaks, CA: Sage.
Moffitt TE. 1990. The neuropsychology of juvenile
delinquency. In: Tonry N, Morris N, editors. Crime
and justice. A review of research). Chicago: University of Chicago Press. p. 99–169.
Moffitt TE. 1993. The neuropsychology of conduct
disorder. Develop Psychopathol 5:135–151.
Moffitt TE, Caspi A, Rutter M, Silva PA. 2000. Sex
differences in antisocial behavior. Cambridge, England: Cambridge University Press.
Miller-Johnson S, Coie JD, Maumary-Gremaud A,
Bierman K. 2002. Peer rejection and early starter
models of conduct disorder. J Abnorm Child Psychol
30:217–230.
Patterson GR. 1999. A proposal relating a theory of
delinquency to societal rates of juvenile crime:
Putting Humpty Dumpty together again. In: Cox
MJ, Brooks-Gunn J editors. Conflict and cohesion in
families. Mahwah, NJ: Erlbaum p. 11–35.
Patterson GR, Forgatch MS, Yoerger KL, Stoolmiller
M. 1998. Variables that initiate and maintain an
early-onset trajectory for juvenile offending. Dev
Psychopathol 10: 531–547.
Posner ML, Rothbart MK. 1998. Attention, self
regulation, and consciousness. Philosop Transact
Royal Soc London 353:1915–1927.
Raine A, Brennan P, Mednick SA. 1997. Interaction
between birth complications and early maternal
39
rejection in predisposing individuals to adult violence: specificity to serious, early-onset violence. Am
J Psychiatr 154:1265–1271.
Raine A, Reynolds C, Venables P, Mednick S,
Farrington D. 1998. Fearless, stimulation-seeking,
and large body size at age 3 years as early
predispositions to childhood aggression at age 11
years. Arch Gen Psychiatr 55:745–751.
Rosenberg M. 1985. Self concept and psychological wellbeing in adolescence. In: Leahy RL, editor. The
development of the self. Orlando: Academic Press. p.
205–246.
Rutter M, Giller H, Hagell A. 1998. Antisocial behavior
by young people. Cambridge: University Press.
Sandler I, Miller P, Short J, Wolchik S. 1989. Social
support as a protective factor for children in stress.
In: Belle D, editor. Children’s social network and
social supports. New York: Wiley. p. 277–307.
Slutske W, Heath A, Dinwiddie S, Madden P, Bucholz
K, Dunne M, Statham D, Martin N. 1998. Common
genetic risk factors for conduct disorder and alcohol
dependence. J Abnorm Psychol 107:363–374.
Stanger C, MacDonald V, McConaughy SH, Achenbach TM. 1996. Predictors of cross informant
syndromes among children and youths referred for
mental health service. J Abnorm Child Psychol
24:597–614.
Stanger C, Achenbach TM, Verhulst FC. 1997.
Accelerated longitudinal comparisons of aggressive
versus delinquent syndromes. Dev Psychopathol
9:43–58.
Unnewehr S. 1995. Diagnosis of psychiatric disorders in
children and adolescents using structured interviews.
Z Kinder Jugendpsychiatr 23:121–132.
Williams JB, Gibbon M, First MB, Spitzer RL, Davies
M, Borus J, Howes MJ, Kane J, Pope HG. Jr,
Rounsaville B, et al. 1992. The Structured Clinical
Interview for DSM-III-R (SCID). II. Multisite testretest reliability. Arch Gen Psychiatr 49:630–636.
World-Health-Organisation. 1991. International Classifikation of Disorders, 10 th revision, ICD-10. Bern:
Huber.