Sei sulla pagina 1di 17

AGGRESSIVE BEHAVIOR

Volume 37, pages 145160 (2011)


Criminal Thinking Patterns, Aggression Styles,
and the Psychopathic Traits of Late High School
Bullies and Bully-Victims
Laurie L. Ragatz

, Ryan J. Anderson, William Fremouw, and Rebecca Schwartz


West Virginia University, Morgantown, West Virginia
: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
This study explored the current psychological characteristics and criminal behavior history of individuals who retrospectively
reported being bullies, bully-victims, victims, or controls (i.e. neither victims nor bullies) during their last 2 years of high school.
College students (n 5960) completed measures of criminal thinking, aggression, psychopathy, and criminal behavior online. We
predicted bullies and bully-victims would demonstrate the highest scores for criminal thinking, proactive aggression, psychopathy,
and have the most criminal infractions. Bullies and bully-victims had signicantly higher scores on criminal thinking, aggression,
psychopathy, and criminal behaviors than victims or controls. Additionally, men were signicantly higher in criminal thinking,
aggression, psychopathy, and had more criminal acts than women. There were no gender by bully group interactions. Logistic
regression analyses differentiated bully-victims from bullies. Bully-victims tended to be male, higher in criminal thinking, and
higher in reactive aggression. In addition, bully-victims were distinct from victims, showing higher criminal thinking and higher
proactive aggression. Aggr. Behav. 37:145160, 2011.
r
2010 Wiley-Liss, Inc.
: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
Keywords: bully-victim; bullying; criminal thinking; psychopathy
INTRODUCTION
The dramatic episodes of school violence at
Columbine [Kass, 2000] and Virginia Tech [Hernan-
dez, 2007] have furthered research interest with
regard to the role of bullying on adolescent
adjustment and aggressive behavior. Initially, re-
searchers studied bullies and victims, but more
recently researchers have identied a third group
bully-victimsdened as individuals who have been
victims and perpetrators of bullying. A US self-
report study of 15,686 students in 6th through 10th
grade revealed that 13.0% of respondents acknowl-
edged bullying, 10.6% of respondents disclosed
being victims of bullying, and 6.3% of respondents
t the designation of bully-victims [Nansel et al.,
2001].
Psychological and Behavioral Differences
Between Bullies and Bully-Victims
Researchers have begun to look at how bullies
and bully-victims differ from each other on several
psychological and behavioral dimensions. For example,
several scholars have found that bully-victims show
higher levels of anxiety compared with bullies
[Espelage and Holt, 2006; Swearer et al., 2001].
In addition, bully-victims have been found to exhibit
lower levels of self-control and more frequent
displays of temper compared with bullies [Haynie
et al., 2001]. Likewise, Toblin et al. [2005] found that
bully-victims showed greater hyperactivity and
poorer ability to regulate their emotions compared
with bullies. Mixed ndings have been reported with
regard to the experience of depression. On the one
hand, ndings point to signicantly more depressive
symptoms in bully-victims compared with bullies
[Espelage and Holt, 2006; Haynie et al., 2001;
Toblin et al., 2005], whereas, on the other hand,
ndings of no difference have been reported with
both bullies and bully-victims showing high levels of
depression [Swearer et al., 2001]. In keeping with the
general pattern of ndings that bully-victims show
Published online 23 November 2010 in Wiley Online Library (wiley
onlinelibrary.com). DOI: 10.1002/ab.20377
Received 12 September 2009; Accepted 13 October 2010

Correspondence to: Laurie L. Ragatz, Department of Psychology,


West Virginia University, 53 Campus Drive, 1124 Life Sciences
Building, P.O. Box 6040, Morgantown, WV 26506-6040.
E-mail: Laurie.Ragatz@mail.wvu.edu
r 2010 Wiley-Liss, Inc.
more negative affect and poorer self-regulation than
bullies, Stein et al. [2007] documented lower self-
esteem among bully-victims compared with bullies.
The net psychological and behavioral ndings across
these studies indicate that bully-victims are an
emotionally and behaviorally complicated group
prone to depression, anger, anxiety, and impulsivity.
This prole places them at increased risk for a
number of problem outcomes.
Indeed, scholars have demonstrated that bully-
victims tend to engage in more antisocial acts
compared with bullies. For example, bully-victims
have been shown to engage in more illegal or
problematic acts (e.g. carrying a weapon, using
alcohol, using illegal drugs, ghting, lying to
parents, staying out past curfew) than bullies
[Haynie et al., 2001; Stein et al., 2007]. Moreover,
bully-victims have been found to exhibit higher
levels of callousness (i.e. lower levels of remorse
when committing antisocial acts) than bullies [Fanti
et al., 2009]. In addition, bully-victims are more
inclined to hang out with deviant peers who share
similar antisocial attitudes and who engage in
criminal behavior [Haynie et al., 2001; Menesini
et al., 2009; Stein et al., 2007]. Teachers observa-
tions indicate that bully-victims are less likely than
bullies to engage in prosocial behavior such as
sharing with other students or providing assistance
to peers [Toblin et al., 2005]. In summary, bully-
victims exhibit even greater impairments in social
skills and empathy than do bullies. Given their
propensity toward aversive social behavior, bullies
are likely to have fewer friendship options and may
end up forming alliances with peers who show
similar difculties relating to others.
With regard to attitudinal and other cognitive
variables, bully-victims show multiple decits.
For example, bully-victims evidence deciencies in
problem-solving [Cassidy and Taylor, 2005; Haynie
et al., 2001] compared with bullies. Furthermore,
Cassidy and Taylor [2005] found that bully-victims
perceived that they had very little control over the
problems they experienced. In the area of social-
cognitive decits, Georgiou and Stavrinides [2008]
reported that bully-victims were more inclined to
ascribe blame for their actions to external sources
compared with bullies. In some areas, however,
bully-victims and bullies have shown comparable
decits in problem solving and social cognition such
as the propensity to believe in the acceptability of
aggressive retaliatory behavior [OBrennan et al.,
2009]. The cumulative ndings are in accordance
with the notion that bully-victims exhibit decits in
problem-solving, engage in external blaming, and
endorse aggressive actionsattitudes that typify
criminal or antisocial thinking and are contributory
to an antisocial lifestyle.
Psychological and Behavioral Differences
Between Victims and Bully-Victims
Given that bully-victims show a number of
important differences when compared with bullies,
it is not at all surprising to nd that bully-victims also
are distinguishable from victims who have never
engaged in bullying. For example, research shows
that bully-victims endorse more depressive and
anxious symptoms than victims [Espelage and Holt,
2006; Haynie et al., 2001; Swearer et al., 2001].
Compared with victims, bully-victims also have less
favorable attitudes toward school [Stein et al., 2007],
are more accepting of problem behaviors such as
disrupting teachers and cheating [Haynie et al., 2001],
and demonstrate attitudes supportive of retaliatory
behavior [OBrennan et al., 2009]. Again, this prole
of social and cognitive decits places bully-victims in
a position of relative isolation compared with their
more socially competent peers. This leaves bully-
victims vulnerable to associations with antisocial
peers who are likely to occasion or reinforce
involvement in physical ghts, drug use, lying to
parents, or skipping school [Haynie et al., 2001].
Psychological and Behavioral Differences
Between Controls and Bully-Victims
It is perhaps intuitive that individuals who do not
endorse ever being a victim or perpetrator of bullying
(i.e. control participants) report more favorable
psychological adjustment and fewer behavioral pro-
blems than bully-victims. However, several key
ndings are highlighted to demonstrate the ways in
which the behavioral and emotional experiences of
bully-victims deviate from peers who report more
normative social experiences. Not surprisingly, con-
trol participants tend to demonstrate lower levels of
anxiety, depression, and psychosomatic symptoms
(i.e. sleep difculties, tension, and dizziness) com-
pared with bully-victims [Espelage and Holt, 2006;
Gini, 2008; Marini et al., 2006; Swearer et al., 2001].
Bully-victims have also been found to engage in more
frequent aggressive and criminal behaviors than
uninvolved participants [Gini, 2008; Haynie et al.,
2001; Marini et al., 2006] and are more likely to
believe that engaging in such aggressive or antisocial
behaviors is acceptable [Camodeca et al., 2003].
This increases condence in the conclusion that
experiences with being a bully and a victim are
146 Ragatz et al.
Aggr. Behav.
in some way associated with poorer psychosocial
adjustment.
In summary, when looking at various psycho-
logical and behavioral dimensions (e.g. callousness,
depression, anxiety, and self-esteem) bully-victims
can be distinguished from victims, bullies, and
uninvolved individuals. However, several important
psychological characteristics have yet to be investi-
gated (e.g. criminal thinking patterns, psychopathy)
but could prove informative to the development of
bullying prevention and intervention programming.
Criminal Thinking Patterns
Walters [2006] dened criminal thinking as
thought content and process conducive to the
initiation and maintenance of habitual lawbreaking
behavior (p. 88). Criminal thinking has been
associated with both juvenile and adult antisocial
behavior [see Simourd and Andrews, 1994; Walters,
2006]. Moreover, data suggest that criminal thinking
proles can be predictive of treatment adherence
[Staton-Tindall et al., 2007], reconviction [Palmer
and Hollin, 2004a; Walters, 2005], and disciplinary
infractions in prison populations [Walters, 1996;
Walters and Geyer, 2005]. Thus, criminal thinking
measures offer a means of quantifying change in
thinking that are related to meaningful behavior
change [Palmer and Hollin, 2004a; Walters, 1997,
2004, 2006].
Criminal thinking has shown broad application.
For example, unique patterns of criminal thinking
have been found among sex offenders [Hatch-
Maillette et al., 2001], white-collar offenders
[Walters and Geyer, 2004], male and female offenders
[Staton-Tindall et al., 2007; Walters et al., 1998],
juvenile or young offenders [Dembo et al., 2007;
Palmer and Hollin, 2004b], and, most recently,
college students with and without criminal behaviors
[McCoy et al., 2006; Walters et al., 2009].
Research has demonstrated that bullies and bully-
victims may demonstrate attitudes more accepting
of aggression than victims or controls [OBrennan
et al., 2009; Toblin et al., 2005]. However, the
relation between perpetrating bullying behaviors
and criminal thinking has not been studied.
Although bullying is not considered a criminal
infraction, it can lead to crimes such as harassment,
assault, and in rare instances, murder (i.e. Colum-
bine, Virginia Tech). The deviance generalization
hypothesis [Arluke et al., 1999] states that if an
individual engages in one type of antisocial behavior
(e.g. bullying), he/she is more likely to commit other
antisocial acts; for example, drug use, gambling, and
robbery. In fact, some researchers have found that
bully-victims were more likely to commit criminal
acts such as using drugs and carrying weapons to
school than bullies [Haynie et al., 2001; Menesini
et al., 2009; Stein et al., 2007]. Other studies have
shown that both bullies and bully-victims are at
increased risk for committing illegal or problem
behaviors such as poor school performance, alcohol
use, and other drug use [Baldry and Farrington,
2000; Kaltiala-Heino et al., 2000; Nansel et al., 2001;
Perren and Hornung, 2005]. Ybarra and Mitchell
[2004a] found that online bullies also showed
elevations in criminal behaviors and substance use.
Therefore, understanding the different criminal
thinking patterns of bullies, victims, and bully-
victims could serve to inform prevention and
intervention programs that target bullying and other
antisocial acts.
Aggression Style
Aggression style, particularly proactive aggression,
is another factor that is related to the perpetration of
antisocial acts such as property crimes, drug crimes,
and bullying [Miller and Lynam, 2006]. Proactive
aggression is to be distinguished from reactive
aggression, and these two typologies are conceptua-
lized as having distinct motivations. That is, proactive
aggression has been described as instrumental, pre-
arranged, and callous, whereas reactive aggression
has been described as impulsive, thoughtless, and
carried out in response to provocation [Dodge, 1991;
Dodge and Coie, 1987].
Proactivereactive aggression styles are related to
bullying behavior by children and adolescents
[Dodge et al., 1990] as well as adults [Cornell
et al., 1996]. Specically, elementary and middle
school-age bullies and bully-victims have been found
to use proactive and reactive aggression compared
with victims and uninvolved peers [Camodeca et al.,
2002; Salmivalli and Nieminen, 2002]. In fact,
Salmivalli and Nieminen [2002] found that bully-
victims demonstrated the highest levels of both
proactive and reactive aggression compared with all
other groups. Conversely, Unnever [2005] found
that middle school-age bully-victims were higher in
reactive aggression than victims, but lower on
proactive aggression than bullies. Also, elementary
and middle school-age victims showed greater
reactive aggression compared with control group
peers [Camodeca et al., 2002; Salmivalli and
Nieminen, 2002; Unnever, 2005]. Furthermore, high
reactive aggression, but not proactive aggression,
has been linked to the tendency to over interpret the
147 Characteristics of Bullies and Bully-Victims
Aggr. Behav.
existence of hostile motives for others actions
[Dodge and Coie, 1987]. The present study
expanded on these previous ndings by examining
current aggression styles of self-reported bullies,
victims, and bully-victims during the last 2 years of
high school.
Psychopathy
High proactive aggression scores have been found
to correspond with higher incidence of psychopathic
traits in both adults and adolescents [Cornell et al.,
1996; Raine et al., 2006; Woodworth and Porter,
2002]. Attributes of psychopathy include lack of
empathy, callousness, disregard for the rights of
others, impulsiveness, manipulativeness, parasitic
tendencies, and increased propensity to commit
violent criminal acts [Patrick, 2006]. Several studies
have shown that bullies and bully-victims exhibit
some of the characteristics of psychopathy such as
low empathy and participation in illegal behaviors
[Baldry and Farrington, 2000; Kaltiala-Heino et al.,
2000; Nansel et al., 2001; Perren and Hornung,
2005; Rigby and Slee, 1993; Viding et al., 2009;
Ybarra and Mitchell, 2004a]. For example, Fanti
et al. [2009] showed that bully-victims were the
highest in callousness, a key core component of
psychopathy. Moreover, individuals highest in
callousness were found to be likely to engage in
both reactive and proactive aggression. Cima and
Raine [2009] found reactive aggression to be related
to the fearlessness or low empathy component of
psychopathy. They concluded that individuals
exhibiting fearlessness were more likely to respond
aggressively to a perceived threat rather than escape
the situation, indicating that those with psycho-
pathic attributes may be inclined toward reactive
aggression.
Hare [2003] is credited with developing the gold
standard assessment of psychopathy, the Psycho-
pathy Checklist-Revised (PCL-R). Factor analytic
research with the PCL-R has provided convincing
evidence for a two-factor model of psychopathy.
Factor 1 includes the interpersonal and affective
features of psychopathy (e.g. supercial charm,
callousness, remorselessness, and grandiosity).
Factor 2 is composed of the behavioral or antisocial
aspects of psychopathy (e.g. parasitic lifestyle, lack
of responsibility, impulsiveness, and versatility in
criminal acts). Limitations of the PCL-R include the
extensive time required to implement the assessment
and the inappropriateness of the measure for use with
non-criminal samples such as college students and
community persons [Lilienfeld and Fowler, 2006].
Given research interest in the psychopathy
construct and given that use of the PCL-R is not
always feasible, self-report measures of psycho-
pathy, such as the Levenson Primary and Secondary
Psychopathy Scales [LSRP; Levenson et al., 1995],
are receiving greater attention and use. The LSRP
was developed for use with non-forensic popula-
tions; thus, its items do not assess criminal
behaviors. The total LSRP score has been found
to moderately correlate (r 5.35) with the PCL-R
total score. Furthermore, as with the PCL-R, the
LSRP also is described according to a two-factor
model of psychopathy. The Primary Psychopathy
Factor (i.e. carelessness, self-centeredness, tendency
to manipulate others) of the LSRP shows a
moderate correlation with Factor 1 of the PCL-R.
However, both the Primary and Secondary Psycho-
pathy (i.e. irresponsibility, tendency to be self-
defeating) Factors of the LSRP have shown
moderate correlations with Factor 2 of the PCL-R
[Brinkley et al., 2001]. Research has not examined
whether bullies or bully-victims differ on the two
factors of psychopathy.
Present Study
Most research to date has focused on middle
school bullying with little attention to the charac-
teristics of high school bullies, bully-victims, and
victims. This study explored the current psycho-
logical characteristics (i.e. criminal thinking, aggres-
sion style, and psychopathy) and criminal behavior
history of individuals who retrospectively reported
being bullies, bully-victims, victims, or uninvolved
(i.e. control group, neither victims nor bullies)
during their last 2 years of high school.
Previous research shows that criminal thinking
patterns are associated with participation in anti-
social acts for both juveniles and adults [see Simourd
and Andrews, 1994; Walters, 2006]. Moreover, some
scholars demonstrate that bully-victims are more
likely to perpetrate criminal or problematic beha-
viors than bullies or victims [Haynie et al., 2001;
Menesini et al., 2009; Stein et al., 2007]. Conversely,
some researchers have found both bullies and bully-
victims to engage in antisocial acts at equal rates
[Baldry and Farrington, 2000; Heino et al., 2000;
Kaltiala-Nansel et al., 2001; Perren and Hornung,
2005]. In addition, bully-victims as well as bullies
have been found to exhibit attitudes accepting of
aggression and engaging in antisocial behaviors.
Therefore, with regard to criminal thinking, we
hypothesized that bullies and bully-victims would
demonstrate the highest elevations in criminal
148 Ragatz et al.
Aggr. Behav.
thinking compared with victims. Furthermore, we
hypothesized that when compared directly, bully-
victims would show higher elevations in criminal
thinking compared with bullies.
The deviance generalization hypothesis [Arluke
et al., 1999] states that if an individual takes part in
one type of antisocial behavior (e.g. bullying), he or
she is likely to commit other antisocial acts (e.g. drug
use, gambling, and robbery). Furthermore, previous
research shows a link between bullying and partici-
pation in illegal behavior [Baldry and Farrington,
2000; Haynie et al., 2001; Menesini et al., 2009;
Perren and Hornung, 2005; Stein et al., 2007].
On the basis of this research, we hypothesized that
bullies and bully-victims would self-report the
greatest number of criminal infractions compared
with victims and uninvolved individuals. We also
hypothesized that when contrasted, bully-victims
would report more involvement in criminal behavior
compared with bullies.
Salmivalli and Nieminen [2002] found that bully-
victims were the highest in proactive and reactive
aggression compared with bullies and victims,
whereas Unnever [2005] found bully-victims to be
the highest in reactive aggression but lower in
proactive aggression when compared with bullies.
This study investigated how aggression styles differ
among bullies, bully-victims, victims, and uninvolved
individuals. On the basis of previous research, we
hypothesized that bullies and bully-victims would
endorse the highest levels of proactive and reactive
aggression compared with victims and uninvolved
individuals. Furthermore, we hypothesized that when
compared directly, bully-victims would report more
reactive aggression compared with bullies. Given
mixed results from previous investigations, it was not
clear whether or not bullies and bully-victims would
differ on proactive aggression.
Research has demonstrated a relation between high
levels of proactive aggression and psychopathy
[Cornell et al., 1996; Raine et al., 2006; Woodworth
and Porter, 2002]. To our knowledge, no previous
study has investigated the relation between bullying
and psychopathy. Several studies have shown that
bullies and bully-victims exhibit some characteristics
that tend to be associated with psychopathy such as
low empathy and participation in illegal behaviors
[Baldry and Farrington, 2000; Kaltiala-Heino et al.,
2000; Nansel et al., 2001; Perren and Hornung, 2005;
Rigby and Slee, 1993; Viding et al., 2009; Ybarra and
Mitchell, 2004a]. This study explored how psycho-
pathy differs among bullies, bully-victims, victims,
and uninvolved individuals. Many of the traits
(e.g. callousness, low empathy) exhibited by bullies
and bully-victims as identied in previous investiga-
tions share common features with Levensons
denition of primary psychopathy (i.e. carelessness,
self-centeredness, tendency to manipulate others).
Therefore, we hypothesized that both bullies and
bully-victims would show elevations on primary
psychopathy compared with victims and uninvolved
individuals. Furthermore, we hypothesized that
bully-victims would show elevations on secondary
psychopathy as dened by Levenson (i.e. irresponsi-
bility, tendency to be self-defeating) compared
with bullies or victims and uninvolved individuals.
This latter hypothesis stems from previous research
demonstrating that bully-victims show characteristics
that tend to be related to the construct of secondary
psychopathy; for example, participation in diverse
criminal acts and impulsive, reactive aggression
[Rigby and Slee, 1993; Salmivalli and Nieminen,
2002; Stein et al., 2007; Unnever, 2005].
Finally, this study examined which of the
psychological and behavioral dimensions (i.e.
psychopathy, aggression type, criminal thinking,
and frequency of criminal behaviors) predicted
participant status as a bully vs. bully-victim and
bully-victim vs. victim. Since previous research on
bullying has not investigated the role of psycho-
pathy or criminal thinking, these analyses were
exploratory.
METHOD
Participants
A sample of 1,707 undergraduates enrolled in
a large Middle Atlantic University took part in
this online study. Similar to Ragatz et al. [2009],
the infrequency validity scale adopted from the
ZuckermanKuhlman Personality Questionnaire
[Zuckerman, 2002] was utilized to select out parti-
cipants who likely did not attend to item content
when responding to the measures utilized in this
study. Additionally, participants were eliminated
from analyses if they completed the study measures
in 10 min or less. After these two selection proce-
dures, 1,524 participants were retained.
Given that the primary purpose of this study was to
assess participant statuses of bully, bully-victim, and
victim as related to a number of other problem
behaviors (e.g. criminal thinking, aggression style), the
sample of 1,524 participants was further divided into
experimental and control groups. The experimental
group (N5480) consisted of those participants
who identied themselves as bullies, bully-victims,
or victims on the Revised Olweus Bully-victim
149 Characteristics of Bullies and Bully-Victims
Aggr. Behav.
Questionnaire [Olweus, 1996]. The control group
(N5480) was matched to the experimental group
on participant characteristics of sex, ethnicity, and
age. Thus, a total of 960 participants were included
in all subsequent analyses in this study. See Table I
for a description of the demographic variables of
the participants.
Participants reported on current psychological
characteristics (i.e. criminal thinking, aggression
style, psychopathy) as well as their histories of
involvement in criminal behaviors. With regard to
bullying, participants were asked to recall events
that took place during their last 2 years of high
school. Thus, participant accounts of their experi-
ences with bullying were retrospective. Brewin
et al. [1993] conducted a comprehensive review
of the use of retrospective accounts in research.
These authors concluded that as long as participants
were old enough and were asked about the
occurrence of specic events, participant accounts
were not adversely affected by the passage of
time. Furthermore, the authors noted that the
accuracy of their accounts was improved when
participants were asked to recall specic events
(e.g. bullying). The use of retrospective accounts in
this study was consistent with these ndings and
recommendations.
Materials
Demographic questionnaire. Participants res-
ponded to demographic questions regarding sex, age,
ethnicity, marital status, and class status (e.g. fresh-
man and sophomore).
The Revised Olweus Bully-Victim Question-
naire. The Revised Olweus Bully-victim Question-
naire [Olweus, 1996] was modied in several ways
for this study. For this study, all participants read
the initial denition of bullying from the Revised
Olweus Bully-victim Questionnaire. Fifteen ques-
tions made up the victim experiences portion of the
questionnaire and measured participants experi-
ences with different forms of being bullied. Nine of
these questions were taken from the Revised Olweus
Bully-victim Questionnaire. A sample question
adopted from the Revised Olweus Bully-victim
Questionnaire reads: I was bullied with mean names,
comments, or gestures with a sexual meaning. The
remaining questions from the Revised Olweus Bully-
victim Questionnaire assessed for the following
behaviors: called mean names or teased; ignored or
left out of activities; hit, kicked, pushed, shoved,
etc.; being stolen from or having belongings
damaged; being the source of unfounded rumors;
being forced to engage in activities; and being called
mean names due to race.
Recent studies indicate that internet bullying is
increasingly problematic and prevalent during the
high school years [Smith et al., 2008; Ybarra and
Mitchell, 2004b]; therefore, eight questions that
assessed being victimized by cyber bullying and
other forms of electronic bullying were included in
this study. Sample cyber bulling question reads:
I have been bullied by someone sending me a
threatening message via the internet (e.g. e-mail,
wall post, etc.) and I have been bullied by
someone spreading a rumor about me via the
internet (e.g. e-mail, wall posts, etc.). Other
behaviors assessed on the cyber bullying portion of
the questionnaire included sending pictures or
messages via the Internet that contained sexual
content; sending threatening messages, pictures, or
videos to someones telephone; putting pictures of
someone on the Internet without his or her
permission; and taking a picture of someone via
telephone and showing it to others without his or
her permission. Also included were 15 questions that
assessed the participants experience with being a
bully. The same behaviors assessed on the victim
portion of the scale were assessed on the bully
portion of the scale, with the only difference being
that the participant was to consider whether he or
TABLE I. Descriptive Data for Bullies, Victims, Bully-Victims,
and the Control Group
Bully Bully-victim Victim Control
M SD M SD M SD M SD
Participant
age (years)
19.42 1.6 19.22 1.4 19.53 1.7 19.52 1.6
n % n % n % n %
Participant gender
Male 60 47.6 45 35.2 46 20.4 151 31.5
Female 66 52.4 83 64.8 180 79.6 329 68.5
Class rank
Freshman 62 49.2 68 53.1 94 41.6 202 42.2
Sophomore 31 24.6 34 26.6 70 31.0 144 30.1
Junior 22 17.5 16 12.5 40 17.7 78 16.3
Senior 11 8.7 9 7.0 22 9.7 52 10.9
Graduate 0 0 1 0.8 0 0 3 0.6
Marital status
Married 1 0.8 2 1.6 3 1.3 10 2.1
Single 122 97.6 126 98.4 222 98.2 468 97.5
Divorced 1 0.8 0 0 1 0.4 1 0.2
Widowed 1 0.8 0 0 0 0 0 0
Separated 0 0 0 0 0 0 1 0.2
Ethnicity
White 119 94.4 119 93.0 211 93.4 449 93.5
Other 7 5.6 9 7 15 6.6 31 6.5
Note. Control 5individuals that were not a bully, victim, or bully-
victim.
150 Ragatz et al.
Aggr. Behav.
she had been the perpetrator of the specic behavior.
For both the victim and bully questions, respon-
dents were asked to consider their experiences
during the last 2 years of high school and up until
starting college. Responses for both the victim and
bully scales were measured on a ve-point scale
(1 5I wasnt bullied during my last 2 years of high
school, 2 5It has only happened once or twice, 3 52
or 3 times a month, 4 5About once a week,
5 5Several times a week).
If a respondent endorsed any of the following
answer options two or three times a month,
about once a week, or several times a week on
any of the victim questions, then the respondent
was classied in the victim category for this study.
If a participant chose any of the following answer
options two or three times a month, about once
a week, or several times a week on any of the
bullying questions, then the respondent was classi-
ed in the bully category. Participants that endorsed
both bully and victim questions with any of the
following three answer choices two or three times
a month, about once a week, or several times
a week classied in the bully-victim category.
Participants that endorsed the response options of
I was not bullied during my last 2 years of high
school or it has only happened once or twice on
both the victim and bully questions comprised the
control group.
Illegal Behavior Checklist. The Illegal Behavior
Checklist [IBC; McCoy et al., 2006] is a self-report
measure that consists of 22-items assessing four areas
of unlawful behavior (i.e. violence, property, drug,
and status). Some of the violent illegal acts measured
on the IBC include gang ghting, stalking, and
forcing someone to have sex. The IBC assesses
property crimes such as shoplifting, using credit cards
or checks illegally, and destroying property. The IBC
measures drug offenses such as selling marijuana,
using prescription drugs, and using hard drugs (e.g.
LSD). Finally, the IBC assesses status offenses such as
running away from home or consuming alcohol under
the age of 21 years. Respondents were provided with
instructions that advised them to select yes if they
had participated in the specied unlawful behavior in
the past and to select no if they had not been
involved in the specic unlawful behavior. To
calculate IBC total scores, the number of yes
responses was summed across the questions on the
IBC assessing illegal actions. Higher IBC total scores
indicated greater involvement in antisocial acts.
The Illegal Behaviors Checklist was modied by
adding the questions that make up the infrequency
validity scale of the ZuckermanKuhlman Personality
Questionnaire [Zuckerman, 2002]. The infrequency
validity scale consists of 10 items that measure
careless or random response styles. All question
responses were true or false. Responses of true were
scored as one point and responses of false were scored
as zero points. Individuals with scores of three or
higher were eliminated from study analyses, which is
similar to the manner in which the infrequency scale is
scored when included in the ZuckermanKuhlman
Personality Questionnaire. These questions were
embedded in the Illegal Behaviors Checklist.
Levensons Self-Report of Psychopathy Scale
(LSRP). Levensons Self-Report of Psychopathy
Scale [LSRP; Levenson et al., 1995] is a self-report
scale consisting of 26 items with all responses
measured on a four-point scale (15disagree strongly,
2 5disagree somewhat, 3 5agree somewhat, 4 5agree
strongly). Sixteen of the items assess primary
psychopathy (i.e. carelessness, self-centeredness, and
tendency to manipulate others). Ten of the items
assessed secondary psychopathy (i.e. irresponsibility
and tendency to be self-defeating). Scores for the
measure were summed by adding the number
corresponding to the response option for each
subscale. Several items were reverse scored on each
of the subscales. Higher scores indicated greater
endorsement of psychopathic traits measured.
Cronbachs alpha was .84 for the primary subscale
and .73 for the secondary subscale.
The Psychological Inventory of Criminal
Thinking Styles (PICTS). The Psychological
Inventory of Criminal Thinking Styles [PICTS;
Walters, 2006] is an 80-item self-report measure of
attitudes that encourage or support participation
in criminal acts. All responses are measured on a
four-point scale (4 5strongly agree, 3 5agree,
2 5uncertain, 1 5disagree). The scale consists of
the following eight criminal thinking subscales:
Mollication (i.e. blame external events for ones
involvement in criminal acts), Entitlement (i.e. belief
that one deserves special rights or attention), Cutoff
(i.e. become angry quickly and consequently engage
impulsively in antisocial behaviors), Power Orienta-
tion (i.e. strong need to be in control of situations),
Sentimentality (i.e. express care for others that is
articial and done to make oneself look good),
Superoptimism (i.e. believe that one can commit
criminal acts without consequences), Cognitive
Indolence (i.e. take shortcuts when working toward
a goal), and Discontinuity (i.e. distracted by and
inuenced by negative others into committing
criminal acts). The general criminal thinking score
(GCT) was calculated by summing the scores for all
of the criminal thinking subscale items.
151 Characteristics of Bullies and Bully-Victims
Aggr. Behav.
The PICTS also consists of two validity scales,
Confusion and Defensiveness. The Defensiveness
scale assesses whether an individual was forthright
and candid when responding to assessment items.
The Confusion scale measures the extent to which
an individual had difculty reading or showed a
random response pattern. The Cronbachs alpha for
the PICTS GCT was .95.
The PICTS originally was developed for use with
incarcerated populations; therefore, several items
were reworded for use with a non-incarcerated
sample. For example, the question Despite the
criminal life I have led, deep down I am basically a
good person was reworded, Despite some illegal
behavior I have done, deep down I am basically a
good person. This is similar to McCoy et al. [2006]
and Walters et al. [2009].
ReactiveProactive Aggression Questionnaire
(RPAQ). The ReactiveProactive Aggression Ques-
tionnaire [RPAQ; Raine et al., 2006] is a 23-item
self-report measure with all response options mea-
sured on a three-point scale (0 5never, 1 5sometimes,
2 5often). Twelve questions assess proactive aggres-
sion (i.e. aggression utilized as a means to achieve
a specic goal). Eleven questions assess reactive
aggression (i.e. immediate aggression committed
without having a clear goal). Separate scale scores
were calculated by summing the responses across
items on each of the subscales. Higher subscale scores
suggested a greater tendency to utilize a specic type
of aggression. Cronbachs alpha was .83 for the
proactive subscale and .82 for the reactive subscale.
Procedures
The University Institutional Review Board approved
this study. Respondents completed the study online via
a university-support website (SONA). The participants
were provided with extra credit for taking part in
the study if it was offered by their class instructor.
All respondents rst read an informed consent
detailing the study purpose followed by the option to
select the I agree button if they wished to participate.
Next, participants were directed to the anonymous
survey where they completed the Revised Olweus
Bully-victim Questionnaire, PICTS, RPAQ, IBC, and
LSRP. All measures were counterbalanced. Upon
completion of the online study, participants were
provided with a written debrieng.
RESULTS
All analyses were conducted using SPSS version
16.0 (Somers, NY). To conduct specic tests of our
hypotheses, we followed up signicant main effects
in our analyses of variance by conducting planned
orthogonal contrasts. User-dened contrasts were
conducted by way of the CONTRAST5SPECIAL
subcommand as described in the SPSS 16.0 Com-
mand Syntax Reference [SPSS, Inc., 2007].
Participant Gender and Bully Group Type
Table I summarizes the demographic characteristics
of the groups. A chi-square test for independence was
conducted to determine whether participant gender
varied by group (i.e. bully, bully-victim, victim,
control). There was a signicant participant gender
by group difference, w
2
(3, n 5960) 529.00, Po.01.
That is, male participants (n 560, 19.9%) were more
often classied as bullies compared with female
participants (n 566, 10.0%). In addition, female
participants (n 5180, 27.4%) were more often
classied as victims compared with male participants
(n 546, 15.2%).
Participant Gender, Bully Group Type, and
Criminal Thinking Styles
Next, we predicted that higher criminal thinking
scores would be observed for bullies and bully-
victims compared with victims and controls. We also
hypothesized that bully-victims would demonstrate
higher criminal thinking scores when contrasted
with bullies. To evaluate these hypotheses we
conducted a 2 (participant gender) 4 (group: bully
vs. bully-victim vs. victim vs. control), factorial
ANOVA with the PICTS GCT score as the
dependent variable.
The interaction effect between participant gender
and group was not statistically signicant. There
was a main effect for group, F(3, 951) 552.08,
Po.001, partial Z
2
5.14 (see Table II). Subsequent
linear contrasts were consistent with our hypothesis
that bullies would show, on average, higher criminal
thinking scores compared with the average criminal
thinking score for victims and controls combined,
F(1, 955) 517.19, Po.001, partial Z
2
5.02. Like-
wise, bully-victims showed, on average, higher
criminal thinking scores compared with the average
criminal thinking score for victims and controls
combined, F(1, 955) 528.90, Po.001, partial
Z
2
5.03. Next, when bully-victims were compared
with bullies, bully-victims demonstrated higher criminal
thinking scores, F(1, 954) 516.45, Po.001, Z
2
5.02.
Furthermore, there was a main effect for gender,
F(1, 951) 522.11, Po.001, partial Z5.02. More
specically, males reported higher criminal thinking
scores when contrasted with females (see Table III).
152 Ragatz et al.
Aggr. Behav.
Participant Gender, Bully Group Type, and
Aggression Styles
Next we conducted analyses to evaluate whether
differences existed between the different bullying
groups on aggression styles. Specically, we hypo-
thesized that bullies and bully-victims would show
higher scores for proactive and reactive aggression
compared with victims and controls. Additionally,
we expected that bully-victims would show higher
scores for reactive aggression compared with bullies.
Given mixed results from previous investigations,
it was not clear whether or not bullies and
bully-victims would differ on proactive aggression.
To evaluate these hypotheses, we conducted two,
2 (participant gender) 4 (group: bully vs. bully-
victim vs. victim vs. control), factorial ANOVAs
with the proactive and reactive aggression subscales
of the RPAQ as the dependent variables.
In the rst two-way independent ANOVA with
proactive aggression as the dependent variable, the
interaction effect between participant gender and
group was not statistically signicant. There was a
main effect for group, F(3, 949) 556.20, Po.001,
partial Z
2
5.15 (see Table II). Subsequent linear
contrasts were consistent with our hypothesis that
bullies would show higher scores on proactive
aggression compared with the average proactive
aggression score for victims and controls combined,
F(1, 952) 546.12, Po.001, partial Z
2
5.05. Like-
wise, bully-victims demonstrated higher proactive
aggression scores compared with the average pro-
active aggression score for victims and controls
combined, F(1, 952) 5145.71, Po.001, partial
Z
2
5.13. Next, when bully-victims were compared
with bullies, bully-victims showed higher average
proactive aggression scores, F(1, 952) 516.16,
Po.001, Z
2
5.02. There was also a main effect for
gender, F(1, 949) 553.32, Po.001, partial Z
2
5.05.
More specically, males reported more proactive
aggression compared with females (see Table III).
In the second two-way independent ANOVA with
reactive aggression as the dependent variable, the
interaction effect between participant gender and
group was not statistically signicant. However,
there was a main effect for group, F(3, 947) 538.20,
Po.001, partial Z
2
5.11 (see Table II). Subsequent
linear contrasts were consistent with our hypothesis
that bullies would show higher scores on reactive
aggression compared with the average proactive
aggression score for victims and controls combined,
F(1, 950) 519.39, Po.001, partial Z
2
5.02. Simi-
larly, bully-victims had higher reactive aggression
scores compared with the average reactive aggres-
sion score for victims and controls combined,
F(1, 950) 588.04, Po.001, partial Z
2
5.09. Next,
when bully-victims were compared with bullies, bully-
victims showed higher reactive aggression scores,
F(1, 950) 516.45, Po.001, Z
2
5.02. Furthermore,
TABLE II. Means for the Criminal Thinking Total Score, Aggression Subscales, Psychopathy Subscales, and Number of Criminal
Offenses as a Function of Being a Bully, Victim, Bully-Victim, or Control
Bully Bully-victim Victim Control
M SD M SD M SD M SD F P
PICTS total score 121.56 27.58 132.88 27.58 103.74 22.88 101.63 25.08 52.08 .001
Aggression subscales
Proactive 4.02 3.31 5.17 3.84 1.81 2.37 1.93 2.43 56.20 .001
Reactive 10.34 3.67 11.93 3.91 9.17 3.67 8.04 3.41 38.20 .001
Psychopathy subscales
Primary 33.64 7.00 33.88 6.44 28.40 6.92 29.23 7.09 21.55 .001
Secondary 23.03 4.79 24.61 4.10 21.00 4.57 20.45 4.16 31.57 .001
IBC total score 4.62 3.29 4.82 3.53 3.08 2.78 3.19 3.00 9.63 .001
Note. PICTS, Psychological Inventory of Criminal Thinking Styles; IBC, Illegal Behavior Checklist.
TABLE III. Means for the Criminal Thinking Total Score,
Aggression Subscales, Psychopathy Subscales, and Number of
Criminal Offenses as a Function of Participant Gender
Male Female
M SD M SD F P
PICTS total score 117.76 28.62 104.87 26.32 22.11 .001
Aggression subscales
Proactive 3.86 3.68 2.03 2.48 53.32 .001
Reactive 9.80 3.90 8.82 3.74 7.48 .01
Psychopathy subscales
Primary 33.30 7.32 28.84 6.81 44.49 .001
Secondary 21.97 4.47 21.27 4.62 1.52 .22
IBC total score 4.45 3.76 3.17 2.71 14.47 .001
Note. PICTS5Psychological Inventory of Criminal Thinking Styles.
IBC5Illegal Behavior Checklist.
153 Characteristics of Bullies and Bully-Victims
Aggr. Behav.
there was a main effect for gender, F(1, 947) 57.48,
Po.01, partial Z
2
5.01. More specically, males
reported more reactive aggression compared with
females (see Table III).
Participant Gender, Bully Group Type, and
Psychopathy
Next, we predicted that bullies and bully-victims
would show higher scores for primary psychopathy
compared with victims and controls. Additionally, we
expected that bullies and bully-victims would show
higher scores for secondary psychopathy compared
with victims and controls. We also predicted that
bully-victims would show higher scores for secondary
psychopathy compared with bullies. To evaluate
these hypotheses, we conducted two, 2 (participant
gender) 4 (group: bully vs. bully-victim vs. victim
vs. control), factorial ANOVAs with the primary and
secondary psychopathy subscales of the LSRP as the
dependent variables.
In the rst two-way independent ANOVA with
primary psychopathy as the dependent variable, the
interaction effect between participant gender and
group was not statistically signicant. A main effect
was found for group, F(3, 925) 521.55, Po.001,
partial Z
2
5.07 (see Table II). Subsequent linear
contrasts were consistent with our hypothesis that
bullies showed higher scores on primary psychopathy
compared with the average primary psychopathy
score for victims and controls combined, F(1, 928) 5
34.34, Po.001, partial Z
2
5.04. Likewise, bully-
victims showed higher levels of primary psychopathy
compared with the average primary psychopathy
score for victims and controls combined, F(1, 928) 5
50.52, Po.001, partial Z
2
5.05. Furthermore, there
was a main effect for gender, F(1, 925) 544.49,
Po.001, partial Z
2
5.05. Males reported, on average,
more primary psychopathy compared with females
(see Table III).
In the second two-way independent ANOVA with
secondary psychopathy as the dependent variable,
the interaction effect between participant gender and
group was not statistically signicant. There was
a main effect for group, F(3, 937) 531.57, Po.001,
partial Z
2
5.10 (see Table II). Subsequent linear
contrasts supported our hypothesis that bullies
would demonstrate higher scores on secondary
psychopathy compared with the average secondary
psychopathy score for victims and controls com-
bined, F(1, 940) 526.43, Po.001, partial Z
2
5.03.
Likewise, bully-victims showed higher secondary
psychopathy scores compared with the average
secondary psychopathy score for victims and
controls combined, F(1, 940) 582.48, Po.001,
partial Z
2
5.08. Next, when bully-victims were
compared with bullies, bully-victims showed higher
levels of secondary psychopathy, F(1, 940) 58.93,
Po.01, Z
2
5.01. The main effect for gender was not
statistically signicant.
Participant Gender, Bully Group Type, and
Criminal Behavior
For analyses examining participation in criminal
behavior, it was hypothesized that bullies and bully-
victims would report having been involved in the
highest number of criminal acts. To evaluate this
hypothesis, we conducted a 2 (participant gender) 4
(group: bully vs. bully-victim vs. victim vs. control)
factorial ANOVA with the number of criminal acts as
measured on the IBC as the dependent variable.
The interaction effect between participant gender
and group was not statistically signicant. There
was a main effect for group, F(3, 952) 59.63,
Po.001, partial Z
2
5.03 (see Table II). Subsequent
linear contrasts were consistent validated our
hypothesis demonstrating that bullies showed more
involvement in criminal acts compared with the
average involvement in criminal acts for victims and
controls combined, F(1, 955) 517.19, Po.001,
partial Z
2
5.02. Likewise, bully-victims endorsed
greater involvement in criminal acts compared with
the average involvement in criminal acts for victims
and controls combined, F(1, 955) 528.90, Po.001,
partial Z
2
5.03. There was also a main effect for
gender, F(1, 952) 514.47, Po.001, partial Z
2
5.02.
More specically, males reported taking engaging in
a greater number of criminal acts compared with
females (see Table III).
Next, a chi-square test for independence was
conducted to determine whether type of criminal
classication (i.e. violent, property, drug, status)
varied by group (i.e. bully, bully-victim, victim,
control). There was a signicant criminal offender
classication by group difference, w
2
(12, n 5960) 5
75.83, Po.01. Bully-victims were more often classi-
ed as violent offenders (n 526, 20.3%), followed
by bullies (n 516, 12.7%), victims (n 518, 8.0%),
and controls (n 536, 7.5%). Bully-victims also were
more likely to be classied as property offenders
(n 553, 41.4%) followed by bullies (n 550, 39.7%),
controls (n 5104, 21.7%), and victims (n 547,
20.8%). When looking at participation in drug
offenses, bully-victims were less often classied as
drug offenders (n 515, 11.7%) compared with
victims (n 549, 21.7%), bullies (n 529, 23.0%),
and controls (n 5113, 23.5%). Victims were more
154 Ragatz et al.
Aggr. Behav.
often classied as status offenders (n 591, 40.3%),
followed by controls (n 5190, 39.6%), bullies
(n 529, 23.0%), and bully-victims (n 529, 22.7%).
Finally, victims were more frequently classied as
non-offenders (n 521, 9.3%) followed by controls
(n 537, 7.7%), bully-victims (n 55, 3.9%), and
bullies (n 52, 1.6%).
A chi-square test for independence was conducted
to determine whether criminal classication (i.e.
violent, property, drug, and status) varied by partici-
pant gender. There was a signicant criminal offender
classication by participant gender difference, w
2
(4,
n 5960) 552.64, Po.01. That is, male participants
were more often classied as violent offenders (n 543,
14.2%) compared with female participants (n 553,
8.1%). When considering property offenses, male
participants also were more often classied as
property offenders (n 5115, 38.1%) compared with
female participants (n 5139, 21.1%). Female partici-
pants were more often classied as drug offenders
(n 5154, 23.4%) compared with male participants
(n 552, 17.2%). Finally, female participants also were
more often classied as status offenders (n 5269,
40.9%) compared with male participants (n 570,
23.2%).
Using Psychological and Behavioral Variables
to Predict Bully Group Type
Correlation coefcients were calculated for pro-
active aggression, reactive aggression, primary
psychopathy, secondary psychopathy, criminal
thinking (i.e. PICTS total score), and criminal
offenses (i.e. IBC total score) to assess for multi-
collinearity before conducting logistic regression
analyses. Bonferroni adjustments (Po.003) were
made to control Type I error (see Table IV).
Next, a stepwise logistic regression analysis was
conducted using the backward likelihood ratio
method to determine the utility of participant gender
(0 5male, 1 5female), aggression style (reactive,
proactive), psychopathy (primary, secondary), crim-
inal thinking, and criminal offenses in uniquely
predicting status as a bully or bully-victim (0 5bully,
1 5bully-victim). Results of the likelihood ratio test
showed that the nal model provided better t to the
data compared with the intercept-only model, w
2
(3,
N5248) 519.346, Po.001, Nagelkerke R
2
5.10.
Individual regression coefcients that were signi-
cant in the model included participant gender
(B5.61, Po.05), reactive aggression (B5.08,
Po.05), and criminal thinking (B5.01, Po.05) (see
Table V). The model correctly classied 76 of 122
(62.3%) bullies and 86 of 126 (68.3%) bully-victims.
The overall correct classication was 65.3%, an
improvement over the chance level.
Next, we conducted a second stepwise logistic
regression analysis using the backward likelihood
ratio method to determine the utility of participant
gender (0 5male, 1 5female), aggression style
(reactive, proactive), psychopathy (primary, second-
ary), criminal thinking, and criminal offenses in
uniquely predicting status as a victim or bully-victim
(0 5victim, 1 5bully-victim). Results of the like-
lihood ratio test demonstrated that the nal model
provided better t to the data compared with the
intercept-only model, w
2
(2, N5343) 5107.94,
Po.001, Nagelkerke R
2
5.37. Individual regression
coefcients that were signicant in the model
included proactive aggression (B5.22, Po.001)
and criminal thinking (B5.03, Po.001) (see
Table VI). The model correctly classied 191 of
217 (88%) victims and 66 of 126 (52.4%) bully-
victims. The overall correct classication was
74.9%, an improvement over the chance level.
TABLE IV. Correlations between, PICTS Total Score,
Aggression Subscales, Psychopathy Subscales, and Number of
Criminal Offenses
1. 2. 3. 4. 5.
1. PICTS total score
2. Proactive aggression .60

3. Reactive aggression .51

.60

4. Primary psychopathy .59

.53

.35

5. Secondary psychopathy .59

.46

.50

.49

6. Number of criminal offenses .49

.50

.37

.36

.38

Note. PICTS, Psychological Inventory of Criminal Thinking Styles.

Po.003.
TABLE V. Logistic Regression Analysis Predicting 248
Participants Status as Bully or Bully-Victim
Variable B SE
Wald
statistic df P
Odds
ratio
Participant gender .61 .270 5.09 1 .02 1.84
Reactive aggression .08 .039 4.33 1 .04 1.08
PICTS total score .01 .005 4.33 1 .04 1.01
Test w
2
df P
Overall model evaluation
Likelihood ratio test 19.35 3 .001
Goodness-of-t test
Hosmer and Lemeshow 9.41 8 .31
Note. 0 5bullies; 1 5bully-victims; 0 5male; 1 5female. PICTS,
Psychological Inventory of Criminal Thinking Styles. Cox and Snell
R
2
5.08. Nagelkerke R
2
5.10. Nonsignicant value for the Hosmer
Lemeshow test is desired and suggests the model t the data well.
155 Characteristics of Bullies and Bully-Victims
Aggr. Behav.
DISCUSSION
This study explored the differences between bullies,
bully-victims, victims, and uninvolved individuals
(i.e. controls) by reexamining several previously
researched variables (i.e. aggression, criminal beha-
vior) along with newly investigated psychological
dimensions (i.e. psychopathy, criminal thinking).
First, ndings showed that bullies and bully-victims
reported higher levels of aggression, criminal thinking,
and psychopathy compared with victims and con-
trols. Second, bullies and bully-victims reported
committing more serious criminal infractions (i.e.
violent, property). Third, ndings showed that
bully-victims were different from bullies. More
specically, bully-victims reported more criminal
thinking, more reactive aggression, and more
secondary psychopathy compared with bullies.
Proactive aggression, which refers to planned
aggression utilized to obtain a specic goal, has
been implicated in bullying and other antisocial acts
(e.g. homicide) [Dodge, 1991; Dodge and Coie,
1987; Miller and Lynam, 2006]. Reactive aggression
represents a type of automatic aggression that
results from being provoked. Previous research has
produced mixed results with regard to the type of
aggression used by bully-victims compared with
bullies. Some researchers have found higher pro-
active and reactive aggression in bully-victims
compared with bullies [Camodeca et al., 2002;
Salmivalli and Nieminen, 2002], whereas other
researchers have found lower proactive aggression
levels in bully-victims compared with bullies
[Unnever, 2005].
For this study, bully and bully-victim groups
were signicantly higher on proactive aggression
compared with the victim and control groups.
Findings from our study also demonstrated bully-
victims had higher levels of proactive aggression
compared with bullies, which is consistent with
previous research [Salmivalli and Nieminen, 2002].
Problematically, previous research demonstrates
bully-victims are likely to demonstrate decits in
their problem-solving capabilities [Cassidy and
Taylor, 2005; Haynie et al., 2001]. Furthermore,
bully-victims demonstrate attitudes supportive of
aggression [OBrennan et al., 2009]. Bully-victims
are also more inclined to engage in various antisocial
or criminal acts [Haynie et al., 2001; Menesini et al.,
2009; Stein et al., 2007]. Consequently, when
presented with a conict, bully-victims may resort
to aggressive means to attain various goals (e.g.
respect from peers, acceptance from peers). To
decrease the use of aggression to attain goals among
bully-victims, rst a functional analysis of the reason
why an individual engages in aggressive or criminal
acts should be conducted. Next, intervention pro-
grams could teach effective problem-solving strate-
gies and alternative communication approaches (i.e.
assertiveness), which allow bully-victims to attain
the same goals they were attempting to achieve with
aggression [see Fite and Colder, 2007].
Next, in line with our prediction, bully-victims
demonstrated the highest levels of reactive aggres-
sion compared with bullies, victims, and controls.
It is not surprising that bully-victims are inclined
to use reactive aggression. Specically, previous
research demonstrates that bully-victims are more
anxious than other groups [Espelage and Holt, 2006;
Swearer et al., 2001]. Furthermore, the combination
of impaired emotion regulation [Toblin et al., 2005]
and tendency to exhibit attitudes supportive of
aggression [OBrennan et al., 2009] could lead bully-
victims more inclined to interpret ambiguous acts of
others as hostile. Heightened physiological arousal
(i.e. anxiety) plus a tendency to perceive hostility in
others may bring out reactive aggression in bully-
victims. Treatment strategies targeting bully-victims
could focus on decreasing physiological arousal
by teaching relaxation strategies (e.g. progressive
muscle relaxation, diaphragmatic breathing). In
addition, effective treatment could use cognitive
restructuring techniques to assist individuals with
developing alternative ways of interpreting the
actions and motivations of others [see Fite and
Colder, 2007].
Maladaptive or criminal thinking has been correlated
with antisocial behavior [Simourd and Andrews, 1994;
Walters, 2006]. Furthermore, more serious offenders
(e.g. violent, property) have been found to have higher
criminal thinking scores compared with less serious
TABLE VI. Logistic Regression Analysis Predicting 343
Participants Status as Victim or Bully-Victim
Variable B SE
Wald
statistic df P
Odds
ratio
Proactive aggression .22 .057 15.31 1 .001 1.25
PICTS total score .03 .006 17.90 1 .001 1.03
Test w
2
df P
Overall model evaluation
Likelihood ratio test 107.94 2 .001
Goodness-of-t test
Hosmer and Lemeshow 14.91 8 .06
Note. 0 5victims; 1 5bully-victims; 0 5male; 1 5female. PICTS5
Psychological Inventory of Criminal Thinking Styles. Cox and Snell
R
2
5.27. Nagelkerke R
2
5.37. Nonsignicant value for the Hosmer
Lemeshow test is desired and suggests the model t the data well.
156 Ragatz et al.
Aggr. Behav.
offenders (e.g. drug, status) [McCoy et al., 2006]. No
previous research has investigated criminal thinking in
bullies or bully-victims. Finding from this study
demonstrated that bully-victims had the highest PICTS
GCT scores compared with bullies, victims, and
controls. The PICTS GCT score represents a compo-
site score of eight criminal thinking dimensions
(Mollication, Entitlement, Cutoff, Power Orientation,
Sentimentality, Superoptimism, Cognitive Indolence,
and Discontinuity). As this research demonstrates,
bully-victims are likely to exhibit a multitude of
criminal thinking errors (as evidenced by being
elevated on the composite score); hence, a compre-
hensive intervention program that uses cognitive
restructuring to address all of these thinking errors
could be warranted for bully-victims.
Psychopathy, another psychological dimension
investigated in this study, has been shown to relate
to participation in antisocial acts [Cornell et al.,
1996; Raine et al., 2006; Woodworth and Porter,
2002]. Yet, no previous research has examined the
relationship between psychopathic traits and bully-
ing. In this study, we utilized the LSRP [Levenson
et al., 1995], which denes psychopathy according to
two factors: primary psychopathy (e.g. self-centered,
manipulative) and secondary psychopathy (e.g.
irresponsibly, engaging in self-defeating behavior).
Bully-victims scored higher on measures of primary
psychopathy compared with victims and controls.
Study ndings also revealed that bully-victims had
higher scores on secondary psychopathy when
contrasted with all other groups. In summary,
bully-victims were elevated on both types of aggres-
sion and psychopathy in comparison with the other
groups examined.
Cima and Raine [2009] found individuals high in
psychopathy exhibit both proactive and reactive
aggression. Specically, their study demonstrated
reactive aggression was related to the fearlessness
(e.g. preference to take part in risky activities) and
blame externalization (e.g. tendency to blame others
when one perpetrates a problematic behavior)
components of psychopathy. Reactive aggression
was inversely related to the stress immunity (e.g.
exhibit more anxiety than is exhibited by others
people in a similar anxiety-provoking situation)
component of psychopathy. Cima and Raine con-
cluded from these ndings that reactively aggressive
individuals are likely to resort to risky aggressive
behavior, such as ghting, in situations where they
experience anxiety. In addition, reactively aggressive
individuals are likely to blame external provocations
for their impulsive aggressive responses. Proactive
aggression was found to be related to overall
psychopathy as well as the machiavellianism (i.e.
callous and self-centered), social potency (i.e. belief
in ones ability to inuence others), and social
nonconformity (i.e. disregard for social rules and
laws) dimensions. These ndings suggest proactively
aggressive individuals have little regard for the rules
of society and likely behave in a way that most
benets them. Furthermore, proactively aggressive
individuals may be effective at manipulating others
to take part in activities that benet the proactively
aggressive individual [Cima and Raine, 2009].
Findings from this study suggest that the bully-
victim group resembles the high psychopathy, versatile
aggressive group described by Cima and Raine [2009].
For instance, previous research has demonstrated that
bully-victims have impaired emotional regulation
[Toblin et al., 2005], heightened impulsivity [Haynie
et al., 200], and perpetrate an array of antisocial
acts [Haynie et al., 2001; Stein et al., 2007]. This prole
of psychosocial problems is consistent with the
constructs of reactive aggression and secondary
psychopathy. Furthermore, high anxiety seen among
bully-victims [Espelage and Holt, 2006; Swearer et al.,
2001] is consistent with high anxiety seen among
individuals that show reactive aggression. In addition,
previous research has shown bully-victims to be high
in callousness [Fanti et al., 2009; Viding et al., 2009],
a key component of primary psychopathy that has
been linked to both reactive and proactive aggression.
Therefore, the nding in this study that bully-victims
are high in both types of aggression and psychopathy
is consistent with previous research.
Individuals high in psychopathy represent a unique
population that has traditionally been difcult to treat
with psychological interventions. Given that bully-
victims are high in psychopathy, they may likewise be
difcult to treat. Harris and Rice [2006] suggest
utilizing a multisystemic therapy approach with high
psychopathic individuals. With such an approach,
teachers, parents, and peers become involved in
the treatment process. Specically, all individuals
involved are taught to use positive reinforcement to
encourage participation in prosocial behaviors and
discourage participation in antisocial acts.
Findings from our logistic regressions analyses
further showed that bully-victims were more likely
to be male, used more reactive aggression, and had
higher levels of criminal thinking than bullies.
In comparison with victims, bully-victims had high-
er levels of criminal thinking and used more
proactive aggression. It should be noted that the
model that differentiated victims from bully-victims
accounted for 37% of the variance in outcome,
whereas the model that differentiated bullies from
157 Characteristics of Bullies and Bully-Victims
Aggr. Behav.
bully-victims accounted for 10% of the variance in
outcome. These results suggest that although
criminal thinking is a key construct in understanding
participation in bullying, other psychological dimen-
sions need to be investigated in order to provide a
more thorough prole of bully-victims.
In summary, bully-victims are a complicated
group that likely requires a comprehensive treat-
ment program. Specically, bully-victims are high in
both forms of aggression (reactive and proactive)
and psychopathy (primary and secondary) suggest-
ing they react impulsively to provocations and are
planful of their retaliations. Bully-victims also have
a difcult time regulating their emotions [Toblin
et al., 2005] and interpreting the emotional reactions
of others [Dodge and Coie, 1987], which likely con-
tributes to their tendency to resort to reactive
aggression. Furthermore, this study and previous
works [Cassidy and Taylor, 2005; Haynie et al.,
2001; OBrennan et al., 2009] have demonstrated
that bully-victims exhibit beliefs supportive of
a criminal lifestyle. Effective interventions for bully-
victims will thus target multiple problem areas by
way of treatment components that include problem-
solving skills training, assertiveness training, relaxa-
tion, and restructuring of hostile and antisocial
beliefs.
Limitations
This study had several limitations. For example, it
was limited by the retrospective design, which
required college students to report previous high
school experiences with bullying or victimization.
Yet, reviews of retrospective accounts suggest
that participant recall is not necessarily affected
adversely by the passage of time, especially when
participants are older and when they are asked to
recall well-dened events [e.g. Brewin et al., 1993].
Another limitation of this research was that it was
collected in a survey format. This may limit our
ndings because participants may not have read
survey instructions or items carefully. If an interview
method were used in place of the survey, follow-up
questions or probes might offer a more in-depth
understanding of individual experiences with bully-
ing and victimization.
The LSRP measure used to assess psychopathy in
this study also may have limitations. The LSRP has
been found to correlate moderately with the PCL-R
[Hare, 2003], which is the gold standard for
assessment of psychopathy. However, it was not
possible to use the PCL-R in this study, as it requires
an extensive record review and a 60-min interview
for each participant. Thus, the LSRP was more suitable
for the college sample given that it is a self-report
measure. Although the LSRP may tap the psycho-
pathy construct in a somewhat different manner
than the PCL-R, both measures have shown moderate
correlations with participation in violent criminal
behavior and alcohol misuse [Brinkley et al., 2001].
Therefore, it is plausible that the LSRP is a useful tool
for gaining insight into personality and behavioral
dimensions that facilitate criminal behavior.
Finally, the small amount of variance explained in
the regression model differentiating bullies from
bully-victims points out the need to specify more
comprehensive models for bully-victim status.
Important variables for future investigation might
include callousness, attachment to adult caregivers,
parenting styles, involvement with child protective
services, being a witness to domestic violence, or
being a victim of abuse.
Future Directions
Future research on bullying may follow several
directions. One possibility is to conduct a long-
itudinal study of middle school students with regard
to their current bullying or victimization experi-
ences. A longitudinal study would allow us to
determine whether bully-victims are victims initially,
bullies initially, or if the two occur concurrently.
A second direction for future research might be to
examine how bullies, bully-victims, and victims
experience shame. Shame displacement has been
linked to bullying [Ahmed and Braithwaite, 2004];
therefore, future research could examine whether
bully-victims experience shame differently compared
with bullies or victims. For instance, shame manage-
ment theory suggests two possible methods for
dealing with shame. Shame acknowledgment occurs
when an individual identies his or her shame
feelings and takes personal accountability by offer-
ing some form of apology to the person whom he or
she has victimized. At appropriate levels, acknowl-
edgment is an effective way of handling shame,
whereas at exaggerated levels acknowledgment can
lead to feelings of low self-worth. Shame displace-
ment occurs when an individual refuses to admit
shame, leading to externalized anger responses.
It would be interesting to investigate how criminal
thinking mediates experiences of shame.
A third avenue of research may involve analyzing
typologies of bully-victims (e.g. physical, indirect,
and cyber). Understanding the different bully-victim
typologies is important, as each typology may
have unique psychological characteristics requiring
158 Ragatz et al.
Aggr. Behav.
different prevention or intervention strategies.
For instance, it has been suggested that interven-
tions be matched to type of aggression [Mark et al.,
2005]. Bullies and bully-victims tend to exhibit both
types of aggression (i.e. proactive, reactive), and
thus may benet from a comprehensive program
that includes multiple components (e.g. relaxation
training, assertiveness communication, problem-
solving, and cognitive restructuring).
In conclusion, this study demonstrated the unique
prole of bully-victims. Specically, bully-victims
were more likely to perpetrate serious criminal acts
and exhibited the highest levels of criminal thinking,
psychopathic traits, and proactive aggression
compared with other study groups. Findings from
this study point to the need for schools and other
institutions to be aware of bullying and its
consequences on many important aspects of intra-
and interpersonal development. Suitable goals may
include the early identication of bullying as well as
immediate and targeted intervention when bullying
does occur.
REFERENCES
Ahmed E, Braithwaite V. 2004. What, me ashamed? Shame
management and school bullying. J Res Crime Delinquency 41:
269294.
Arluke A, Levin J, Luke C, Ascione F. 1999. The relationship of
animal abuse to violence and other forms of antisocial behavior.
J Interpers Violence 14:963975.
Baldry AC, Farrington DP. 2000. Bullies and delinquents: Personal
characteristics and parental styles. J Commun Appl Soc Psychol
10:1731.
Brewin CR, Andrews B, Gotlib IH. 1993. Psychopathology and early
experience: A reappraisal of retrospective reports. Psychol Bull
113:8298.
Brinkley CA, Schmitt WA, Smith SS, Newman JP. 2001. Construct
validation of a self-report psychopathy scale: Does Levensons self-
report psychopathy scale measure the same constructs as Hares
psychopathy checklist-revised? Pers Indiv Differ 31:10211038.
Camodeca M, Goossens FA, Terwogt MM, Schuengel C. 2002.
Bullying and victimization among school-age children: Stability
and links to proactive and reactive aggression. Soc Dev 11:332345.
Camodeca M, Goossens FA, Schuengel C, Meerum Terwogt M.
2003. Links between social information processing in middle
childhood and involvement in bullying. Aggr Behav 29:116127.
Cassidy T, Taylor L. 2005. Coping and psychological distress as
a function of the bully victim dichotomy in older children. Soc
Psychol Educ 8:249262.
Cima M, Raine A. 2009. Distinct characteristics of psychopathy
relate to different subtypes of aggression. Pers Indiv Differ 47:
835840.
Cornell DG, Warren J, Hawk G, Stafford E, Oram G, Pine D. 1996.
Psychopathy in instrumental and reactive offenders. J Consult
Clin Psychol 64:783790.
Dembo R, Turner CW, Jainchill N. 2007. An assessment of criminal
thinking among incarcerated youth in three states. Crim Justice
Behav 34:11571167.
Dodge KA. 1991. The structure and function of reactive
and proactive aggression. In: Pepler D, Rubin K (eds).
The Development and Treatment for Childhood Aggression.
Hillsdale: Erlbum, pp 201218.
Dodge KA, Coie JD. 1987. Social-information processing factors in
reactive and proactive aggression in childrens peer groups. J Pers
Soc Psychol 53:11461158.
Dodge KA, Price JM, Bachorowski JA, Newman JP. 1990. Hostile
attributional biases in severely aggressive adolescents. J Abnorm
Psychol 99:385392.
Espelage DL, Holt MK. 2006. Dating violence and sexual harassment
across the bully-victim continuum among middle and high school
students. J Youth Adolescence 36:799811.
Fanti KA, Frick PJ, Georgiou S. 2009. Linking callous-unemotional
traits to instrumental and non-instrumental forms of aggression.
J Psychopathol Behav Assess 31:285298.
Fite PJ, Colder CR. 2007. Proactive and reactive aggression and peer
delinquency: Implications for prevention and intervention.
J Early Adolescence 27:223240.
Georgiou SN, Stavrinides P. 2008. Bullies, victims and bully-victims:
Psychological proles and attribution styles. School Psychol Int
29:574589.
Gini G. 2008. Associations between bullying behaviour, psycho-
somatic complaints, emotional and behavioural problems.
J Paediatr Child Health 44:492497.
Hare RD. 2003. Manual for the Revised Psychopathy Checklist,
2nd edition. Toronto, ON, Canada: Multi-Health Systems.
Harris GT, Rice ME. 2006. Treatment of psychopathy: A review
of empirical ndings. In: Patrick CJ (ed.). Handbook of
Psychopathy. New York: Guilford Press, pp 555572.
Hatch-Maillette MA, Scalora MJ, Huss MT, Baumgartner JV. 2001.
Criminal thinking patterns: Are child molesters unique? Int J
Offender Ther Comp Criminol 45:102117.
Haynie DL, Nansel T, Eitel P, Crump AD, Saylor K, Yu K, Simons-
Morton B. 2001. Bullies, victims, and bully/victims: Distinct
groups of at-risk youth. J Early Adolescence 21:2949.
Hernandez M. 2007. School bullying a focus since shootings.
Retrieved from http://www.venturacountystar.com/news/2007/
may/05/school-bullying-a-focus-since-shootings/
Kaltiala-Heino R, Rimpela M, Rantanen P, Rimpela A. 2000.
Bullying at school: An indicator of adolescents at risk for mental
disorders. J Adolescence 23:661674.
Kass J. 2000. Witnesses tell of Columbine bullying. Retrieved from
http://denver.rockymountainnews.com/shooting/1003col4.shtml
Levenson MR, Kiehl KA, Fitzpatrick CM. 1995. Assessing psycho-
pathic attributes in a noninstitutionalized population. J Pers Soc
Psychol 68:151158.
Lilienfeld SO, Fowler K. 2006.The self-report assessment of psycho-
pathy: Problems, pitfalls, and promises. In: Patrick C (ed.).
Handbook of Psychopathy. New York: Guilford, pp 107132.
Marini ZA, Dane AV, Bosacki SL, YLC-CURA. 2006. Direct and
indirect bully-victims: Differential psychosocial risk factors
associated with adolescents involved in bullying and victimiza-
tion. Aggr Behav 32:551569.
Mark W, Orobio de Castro B, Koops W, Matthys W. 2005. The
distinction between reactive and proactive aggression: Utility for
theory, diagnosis, and treatment. Eur J Dev Psychol 2:197220.
McCoy K, Fremouw W, Tyner E, Clegg C, Johansson-Love J,
Strunk J. 2006. Criminal-thinking styles and illegal behavior
among college students: Validation of the PICTS. J Forensic Sci
51:11741177.
Menesini E, Modena M, Tani E. 2009. Bullying and victimization in
adolescence: Concurrent and stable roles and psychological
health symptoms. J Genet Psychol 170:115133.
159 Characteristics of Bullies and Bully-Victims
Aggr. Behav.
Miller JD, Lynam DR. 2006. Reactive and proactive aggression:
Similarities and differences. Pers Indiv Differ 41:14691480.
Nansel TR, Overpeck M, Pilla RS, Ruan WJ, Simons-Morton B,
Scheidt P. 2001. Bullying behaviors among US youth. Prevalence
and association with psychosocial adjustment. J Am Med Assoc
285:20942100.
OBrennan LM, Bradshaw CP, Sawyer AL. 2009. Examining
developmental differences in the social-emotional problems among
frequent bullies, victims, and bully/victims. Psychol Schools 46:
100115.
Olweus D. 1996. Revised Olweus Bully-Victim Questionnaire.
Mimeo, Bergen, Norway: Research Center for Health Promotion
(HEMIL Center), University of Bergen.
Palmer EJ, Hollin CR. 2004a. Predicting reconviction using the
psychological inventory of criminal thinking styles (PICTS) with
English offenders. Legal Criminol Psychol 9:5768.
Palmer EJ, Hollin CR. 2004b. The use of the psychological inventory
of criminal thinking styles with English young offenders. Legal
Criminol Psychol 9:253263.
Patrick CJ. 2006. Handbook of Psychopathy. New York: The
Guilford Press.
Perren S, Hornung R. 2005. Bullying and delinquency in adoles-
cence: Victims and perpetrators family and peer relations. Swiss
J Psychol 64:5164.
Raine A, Dodge K, Loeber R, Gatzke-Kopp L, Lyman D, Reynolds C,
Stouthamer-Loeber M, Lui J. 2006. The reactive-proactive aggres-
sion questionnaire. Differential correlates of reactive and proactive
aggression in adolescent boys. Aggr Behav 32:159171.
Ragatz L, Fremouw W, Thomas T, McCoy K. 2009. Vicious dogs:
The antisocial behaviors and psychological characteristics of
owners. J Forensic Sci 54:699703.
Rigby K, Slee PT. 1993. Dimensions of interpersonal relations
among Australian schoolchildren and implications for psycho-
logical well-being. J Soc Psychol 133:3342.
Salmivalli C, Nieminen E. 2002. Proactive and reactive aggression
among school bullies, victims, and bully-victims. Aggr Behav 28:
3044.
Simourd DJ, Andrews DA. 1994. Correlates of delinquency: A look
at gender differences. Forum Correctional Res 6:2631.
Smith PK, Mahdavi J, Carvalho M, Fisher S, Russell S, Tippett N.
2008. Cyberbullying: Its nature and impact in secondary school
pupils. J Child Psychol Psychiatry 49:376385.
SPSS, Inc. 2007. SPSS 16.0 Command Syntax Reference. Retrieved
April 13, 2010, from http://support.spss.com/ProductsExt/SPSS/
Documentation/Manuals/16.0/SPSS%2016.0%20Command%20
Syntax%20Reference.pdf
Staton-Tindall M, Garner BR, Morey JT, Leukefeld C, Krietemeyer J,
Saum CA, Oser CB. 2007. Gender differences in treatment
engagement among a sample of incarcerated substance abusers.
Crim Justice Behav 34:11431156.
Stein JA, Dukes RL, Warren JI. 2007. Adolescent male bullies,
victims, and bully-victims: A comparison of psychosocial and
behavioral characteristics. J Pediatr Psychol 32:273282.
Swearer SM, Song SY, Cary P, Eagle J, Mickelson WT. 2001.
Psychosocial correlates in bullying and victimization: The
relationship between depression, anxiety, and bully-victim status.
J Emot Abuse 2:95121.
Toblin RL, Schwartz D, Hopmeyer Gorman A, Abou-ezzeddine T.
2005. Social-cognitive and behavioral attributes of aggressive
victims of bullying. Appl Dev Psychol 26:329346.
Unnever JD. 2005. Bullies, aggressive victims, and victims: Are they
distinct groups? Aggr Behav 31:153171.
Viding E, Simmonds E, Petrides KV, Frederickson N. 2009. The
contribution of callous-unemotional traits and conduct problems
to bullying in early adolescence. J Child Psychol Psychiatry 50:
471481.
Walters GD. 1996. The psychological inventory of criminal thinking
styles: Part III. Predictive validity. Int J Offender Ther Comp
Criminol 40:105112.
Walters GD. 1997. Predicting short-term release outcome using the
LCSF and PICTS. J Ment Hlth Correct 43:1825.
Walters GD. 2004. Predictors of early termination in a prison-based
program of psychoeducation. Prison J 84:171183.
Walters GD. 2005. Incremental validity of the psychological inventory
of criminal thinking styles as a predictor of continuous and
dichotomous measures of recidivism. Assessment 12:1927.
Walters GD. 2006. Appraising, researching and conceptualizing
criminal thinking: A personal view. Crim Behav Ment Hlth 16:
8799.
Walters GD, Geyer MD. 2004. Criminal thinking and identity in
male white collar offenders. Crim Justice Behav 31:263281.
Walters GD, Geyer MD. 2005. Construct validity of the psycho-
logical inventory of criminal thinking styles in relationship to the
PAI, disciplinary adjustment, and program completion. J Pers
Assess 84:252260.
Walters GD, Elliott WN, Miscoll D. 1998. Use of the psychological
inventory of criminal thinking styles in a group of female
offenders. Crim Justice Behav 25:125134.
Walters GD, Felix CM, Reinoehl R. 2009. Replicability and cross-
gender invariance of a two-dimensional model of antisociality in
male and female college students. Pers Indiv Differ 46:704708.
Woodworth M, Porter S. 2002. In cold blood: Characteristics of
criminal homicides as a function of psychopathy. J Abnorm
Psychopathy 111:436445.
Ybarra ML, Mitchell KJ. 2004a. Youth engaging in online
harassment: Associations with caregiver-child relationships,
internet use, and personal characteristics. J Adolescence 27:
319336.
Ybarra ML, Mitchell KJ. 2004b. Online aggressor/targets, aggressors,
and targets: A comparison of associated youth characteristics.
J Child Psychol Psychiatry 45:13081316.
Zuckerman M. 2002. Zuckerman-Kuhlman personality question-
naire (ZKPQ): An alternative ve-factorial model. In: de Raad B,
Perugini M (eds). Big Five Assessment. Ashland, OH: Hogrefe &
Huber Publishers, pp 376392.
160 Ragatz et al.
Aggr. Behav.
Copyright of Aggressive Behavior is the property of John Wiley & Sons, Inc. and its content may not be copied
or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission.
However, users may print, download, or email articles for individual use.

Potrebbero piacerti anche