relational aggression in children with preschool-onset ... · relational aggression in children...

NEW RESEARCH

Relational Aggression in Children WithPreschool-Onset Psychiatric Disorders

Andy C. Belden, Ph.D., Michael S. Gaffrey, Ph.D., Joan L. Luby, M.D.

Objective: The role of preschool-onset (PO) psychiatric disorders as correlates and/or riskfactors for relational aggression during kindergarten or first grade was tested in a sample of146 preschool-age children (age 3 to 5.11 years). Method: Axis-I diagnoses and symptom scoreswere derived using the Preschool Age Psychiatric Assessment. Children’s roles in relational aggressionas aggressor, victim, aggressive-victim, or nonaggressor/nonvictim were determined at preschool andagain 24 months later at elementary school entry. Results: Preschoolers diagnosed with POpsychiatric disorders were three times as likely as the healthy preschoolers to be classifiedaggressors, victims, or aggressive-victims. Children diagnosed with PO disruptive, de-pressive, and/or anxiety disorders were at least six times as likely as children without POpsychiatric disorders to become aggressive-victims during elementary school after cova-rying for other key risk factors. Conclusions: Findings suggested that PO psychiatricdisorders differentiated preschool and school-age children’s roles in relational aggressionbased on teacher report. Recommendations for future research and preventative interven-tion aimed at minimizing the development of relational aggression in early childhood byidentifying and targeting PO psychiatric disorders are made. J. Am. Acad. Child Adolesc.Psychiatry, 2012;51(9):889 –901. Key Words: relational aggression, preschool psychiatricdisorders, aggressive-victim, bullying

rp

T he experience of peer aggression as perpe-trator, victim, or both during childhood isone of the strongest social predictors of

developmental difficulties and maladjustment inlater childhood.1-6 Historically, investigations ofpeer aggression have focused on physical formsin males during middle childhood and adoles-cence.7-12 Physical aggression involves the intentto hurt, harm, or injure others using physicalforce, such as hitting, kicking, punching, push-ing, and forcibly taking things away frompeers.13,14 To assess nonphysical forms of peeraggression, thought to be more characteristic offemales, Crick and Grotpeter developed andtested an instrument that reliably measured anddifferentiated physical and relational forms ofpeer aggression.15 Relational aggression is de-fined as the intent to hurt or harm others throughnonphysical manipulation, threat, or damage toclose relationships, friendships, and/or socialstatus.16,17 That is, relational aggression is theexpression of aggression and manipulation ofothers through the use of social inclusion or
exclusion.
JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 51 NUMBER 9 SEPTEMBER 2012

Crick et al. have examined the factor structureof physical versus relational peer aggression.18

Results indicated that despite several overlap-ping characteristics, physical and relationalforms of peer aggression have discrete factorstructures. When examined simultaneously,measures of relational aggression accounted forunique portions of variance in social, emotional,and cognitive development outcomes above andbeyond physical peer aggression.18-20 Although aich body of literature has informed patterns ofhysical aggression in early childhood,21,22 the

pioneering work of Crick et al. examining rela-tional aggression resulted in a more comprehen-sive understanding of physical and social formsof aggression throughout development.23 Thesefindings also provided initial evidence that rela-tional forms of peer aggression occur much ear-lier in development than originally thought.24-27

Relational Aggression Among Preschool-AgeChildrenOnly recently has the study of relational aggres-
sion broadened from an almost exclusive focus
889www.jaacap.org

p

pscd

acbvpitaduatt

BELDEN et al.

on middle childhood and adolescence to thepreschool-age period of development. Despite abroad understanding of overt physical aggres-sion among preschoolers, it had been widelyaccepted that relational forms of peer aggressionwere a relatively uncommon occurrence amongpreschoolers (for exception, refer to Bjorkqvist etal.9). Because researchers had historically focusedon measuring overt aggression, it was thoughtthat peer aggression in preschoolers was pre-dominantly exhibited by boys in the form ofphysical aggression.28-30 Findings consistentlydemonstrated that preschool boys compared togirls were in fact more physically aggressiveduring peer interactions.31,12 These findings ledmany researchers to assume that the social livesof preschool-age girls were largely devoid of peeraggression. A second problematic assumptionheld before the early 1990s was that preschoolersdid not possess the social, emotional, and cogni-tive capacities to use more sophisticated forms ofpeer aggression characteristic of relational ag-gression. Both assumptions were challengedwhen empirical studies began examining rela-tional and physical forms of peer aggressionsimultaneously within samples of preschoolchildren.26,31

Studies using multiple methods and infor-mants have demonstrated validity, reliability,and short-term stability of relational aggressionmeasured in preschool children.12,26,27,31-33 Rela-tionally aggressive behaviors in preschool-agechildren tend to be overt and direct (and there-fore easily observed) as opposed to indirect,discrete, or subtle, as more often manifest inolder children. An example of overt relationalaggression used by preschool-age childrenwould be a child putting their hands over theirears indicating they are actively ignoring andrejecting a peer. Another example is when anaggressor directly tells the victim that he/shewill not be invited to his/her party unless thevictim does what the aggressor demands. Indi-rect relationally aggressive behaviors more com-monly used by school-aged children include be-haviors such as disseminating malicious rumorsabout victims to peers. Although studies exam-ining relational aggression in preschoolers haveindicated both genders engaged in this behavior,relational aggression is used more frequently bypreschool girls than boys.8,29-32,34-36 Preschoolgirls engage in and experience more sophisti-
cated, complex, and socially directed forms of f
JOURN

890 www.jaacap.org

eer relational aggression.12,37-39 It has been esti-mated that 70% of girls’ aggressive behaviorsdirected at peers are relationally focused andnonphysical.32 By measuring children’s use of

hysical and relational forms of peer aggression,everal studies have illustrated developmentalontinuity of peer aggression from early to mid-le childhood for both boys and girls.40-42

Peer Aggression in PreschoolersMonks et al. reported that, in their study, 25% ofpreschool children were aggressors and 22.1%were victims of peer aggresion.43 Preschool chil-dren who are persistently aggressive towardpeers (“aggressors”) show greater oppositional-ity, poorer school adjustment, greater emotiondysregulation, and more symptoms of inatten-tion and depression; they are also more likely tobecome antisocial in adolescence.6,44-47 Being per-sistently victimized during early childhood hasbeen associated with poorer school performanceand impaired social adjustment, greater loneliness,and increased social withdrawal and isolation, aswell as episodic reactive aggression.19,48-50 Of par-ticular importance is a distinct subgroup of chil-dren who are aggressors as well as victims of peeraggression, referred to as “aggressive-victims.”

Preschoolers classified by their teachersand/or peers as being aggressive-victims differin several important ways from peers classified asbeing “pure-aggressors” or “pure-victims.” Ag-gressive-victims are more likely to show reactiveaggression, in contrast to pure-aggressors, whoproactively use aggression to achieve a goal.51

Compared to “pure-aggressors” or “pure-victims,”aggressive-victims are described as being moreanxious, physically reactive, and annoying to otherchildren.52 Findings have also indicated thatggressive-victims have distinct temperamentalharacteristics that differ from children classified aseing pure-aggressors or pure-victims. Aggressive-ictims are more likely than pure-victims andure-aggressors to be impulsive, irritable, and

mpatient during interactions with peers.53 Theemperamental and behavioral characteristics ofggressive-victims are often the least sociallyesirable and are known risk factors for contin-ed involvement in peer aggression. The mal-daptive behavioral and temperamental charac-eristics of aggressive-victims, as well as findingshat these children often have significantly greater
unctional and developmental impairments, sup-
AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY


tanop

cafn

RELATIONAL AGGRESSION IN CHILDREN

port the finding of increased risk for and/or ratesof mental illness in this group.54,55

The prevalence of being an aggressive-victimis estimated to be approximately 6% to 8% inyoung children.56,57 Results from kindergartenstudents showed that 18% of boys and 8% of girlswere classified as aggressive-victims using teacher-reports.58 In addition to experiencing moresevere developmental impairment, aggressive-victims have a greater risk for manifesting psy-chiatric problems.59,60 Prior results demonstratedthat after controlling for pre-existing adjustmentproblems at age 5 years, aggressive-victims com-pared to aggressors or victims had significantlyhigher internalizing and externalizing behaviorproblem mean scores by the age of 7 years.56

Aggressive-victims are consistently reported ashaving the highest level of maladjustment amongall children involved in peer aggression, exhibitingmore symptoms of both internalizing and external-izing problems.54,55

PO Psychiatric Disorders as Risk Factors forRelational Aggression in PreschoolOf particular public health concern are findingsthat school children involved in relational ag-gression are more likely to manifest an array ofmental health problems that often continue intoadolescence and adulthood.59,61-63 Findings fromolder children suggest that correlations be-tween relational aggression and mental disor-ders may exist before children enter elementaryschool.5,6,64 To date, studies that have examinedemotional and behavioral problems associatedwith relational aggression in preschool childrenhave predominantly used more general dimen-sional measures of internalizing or externalizingsymptoms but have not examined more specificcategorical DSM-IV psychiatric disorders.25 Amounting body of literature has established thatreliable and valid DSM-IV Axis-I psychiatric dis-orders can be identified in children as young as 3years (reviewed by Egger et al.65). The impor-tance of early identification of psychiatric disor-ders during the preschool period continues togain attention based on new findings demon-strating the developmental continuity of thesedisorders from preschool to early adolsecence.66

The preschool period represents a unique phase oflife during which rapid social development takesplace and patterns of social interactions begin toform in the context of the “semi-structured” envi-
ronment of the preschool classroom. Therefore,


the social milieu of the preschool classroom,playground, and lunchroom provide the stagefor the emergence of aggressor and victim relatedsocial behavior. This preschool classroom alsorepresents a unique opportunity for observationsof these behaviors, as a result of minimal self-monitoring, limited cognitive capacities to antic-ipate future consequences of misbehavior, anddecreased awareness of social norms previouslydescribed. Following this, identifying and char-acterizing associations between early onset men-tal illness and relational forms of aggression maybe key to inform how these early behaviorsinfluence each other developmentally.67 Thus,he current study tested expected associationsnd group differences between children diag-osed with preschool-onset (PO) psychiatric dis-rders and their roles in relational aggression atreschool and 2 years later in elementary school.

METHODRecruitment and ParticipantsThis investigation used data from a National Institute ofMental Health (NIMH)–funded study entitled Validationof Preschool Depressive Syndromes (PDS). This ongoing,longitudinal, multi-method and multi-informant (i.e.,parents, children, and teachers) study was designed toexamine the nosology, etiology, and course of PO majordepressive disorder [MDD] (additional recruitment de-tails in Luby et al.68). Between May 2003 and March 2005,aregivers with children between 3.0 and 5.11 years ofge were recruited from pediatricians’ offices, daycareacilities, and preschools in a large metropolitan commu-ity using the Preschool Feelings Checklist (PFC).69 The

PFC is a brief validated screening tool for early-onsetemotional disorders. Excluded were children withchronic medical illnesses, neurological problems, perva-sive developmental disorders, and language and/or cog-nitive delays, as well as those out of the study age range.It is important to note that the recruitment techniquesused in this study were designed to oversample forpreschoolers with or at risk for MDD and/or attention-deficit/hyperactivity disorder (ADHD). Therefore, diag-nostic data from the present study cannot be used tocalculate the prevalence rates of PO psychiatric disordersin the general population.

A total sample of 306 caregiver–child dyads agreed toparticipate and completed their baseline assessment in alaboratory. Of the total sample of 306 children at baseline,a subsample of 202 children had complete teacher data.Children who stayed at home with a primary caregiveraccounted for 80 of the total 104 children with missingteacher data on the MacArthur Health and BehaviorQuestionnaire—Teacher Version (HBQ-T; described in
Measures section). The remaining group of 24 children
891www.jaacap.org

text a

BELDEN et al.

with missing teacher data were the result of caregiversrefusing to consent for us to contact children’s preschoolteachers or teachers failing to return a completed HBQ-T.For the cross-sectional analyses at baseline, children hadto be enrolled in formal (i.e., home daycare facilities werenot included) preschool or pre-kindergarten program. Ofthe 202 children with completed HBQ-T data at baseline,42 children were attending home daycare. A total of 14children were already enrolled in kindergarten at base-line (i.e., did not have preschool data available). Thus, thefinal sample size of preschoolers eligible for analysis was146. Children attending preschool or pre-kindergartenprograms at baseline and who were enrolled in kinder-garten or first grade 24 months after their baselineassessment were examined in the longitudinal analyses(n � 121). The 25 children with missing data at school age

TABLE 1 Demographic and Diagnostic Characteristics a

Demographics

Preschool NonNonVic(n � 6

Gender, nMale 33Female 36

Age, years, n3 174 405 12

Ethnicity, nBlack 17White 45Other 6

Income, n0–20K 1020,001–40K 1140,001–60K 8�60,001K 35

Highest Level of Education, nHigh school diploma or less 7Some college/2-year degree 204-Year degree 16Schooling beyond 4-year degree 26

Diagnostic Characteristics at time 1 (n)

Healthy (72)Disruptive only (17)Anxiety only (16)Depression only (9)Disruptive and anxiety (4)Disruptive and depression (10)Anxiety and depression (5)Disruptive and anxiety and depression (13)

Note: Differences in sample sizes and percentages not equal to 100% in

had either dropped out of the study or had teachers who

JOURN

892 www.jaacap.org

did not complete the HBQ-T. Descriptive data for demo-graphic and diagnostic variables used in the analyses areincluded in Table 1.

Procedure and MeasuresParent–child dyads participated in a 3- to 4-hour annualassessment. During this time, primary caregivers (94%mothers) were interviewed about their children’s behav-iors, emotions, and age-adjusted manifestations of psy-chiatric symptoms. Caregivers were also asked for per-mission to contact children’s current or most recentteacher. Teachers of consenting families were contactedwithin 7 to 10 days of the annual assessment and weresent a brief study description, directions for participating,

eline

essor/ PreschoolAggressor(n � 28)

PreschoolVictim

(n � 28)

PreschoolAggressive-Victim

(n � 21)

12 14 1116 14 10

3 7 613 15 812 6 7

7 11 819 11 9

2 6 4

6 7 85 6 44 6 1

10 8 7

4 4 311 14 9

5 5 68 5 3

15 15 52 4 73 2 12 0 20 0 22 3 20 0 14 4 1

nd in this table are the result of missing data from individual participants.

t Bas

Aggrtim9)

374

1052344

and questionnaires to be completed.



phcr

trtswpairmimca

yagwivcgwohaisst


Ratings of Childhood Aggressors and/or VictimsChildren were classified as aggressors, victims,aggressive-victims, or nonaggressor/nonvictims usingpreschool and elementary school teachers’ reports on theMacArthur Health and Behavior Questionnaire-TeacherVersion (HBQ-T 1.0).70 Aggressors (i.e., perpetrators ofrelational aggression) were assessed using six itemsfrom the HBQ-T, as follows: when mad at peer, keepsthat peer from being in the playgroup; tries to getothers to dislike a peer; tells others not to play with orbe a peer’s friend; tells peers that he/she won’t playwith peers or be peers’ friend unless peers do whathe/she asks; verbally threatens to keep a peer out ofthe playgroup if the peer does not do what he/shewants; and tells a peer that the peer will not be invitedto the aggressor’s birthday party unless that peer doeswhat the aggressor wants. For each item, teachers rank(0 � never; 1 � sometimes; 2 � often) children’sengagement in relational aggression as aggressors. Thesix items are averaged together to create a relationalaggressor mean score for each child.

Victims of relational aggression were identifiedusing four items from the peer victimization subscaleof the HBQ-T, as follows: other children refuse to lethim/her play with them; is actively disliked by otherchildren, who reject him/her from their playgroup; ispicked on by other children; and is teased and ridi-culed by other children. For each item, teachers ratedchildren as 1 � not at all like this; 2 � very little like;3 somewhat like; 4 � very much like. Mean scoreswere computed to provide continuous victim scores.

To create a categorical aggressor variable, at pre-school and again during elementary school, childrenwith an aggressor mean score in the top 20th percentileand a victim score in the bottom 80th percentile wereclassified as aggressors. Children scoring in the top20th percentile of the victim subscale and in the bottom80th percentile of the aggressor subscale were classifiedas victims. Children who scored in the top 20% of theaggressor and victim scales were classified as aggressive-victims. Children who scored in the bottom 80th percen-tile on both the aggressor and victim subscales of theHBQ-T were in the nonaggressor/nonvictim group. Thismethod resulted in four mutually exclusive groups atpreschool and elementary school: aggressor, victim, ag-gressive-victim, or nonaggressor/nonvictim. Similarclassification methods for differentiating aggressors, vic-tims, and aggressive-victims have been used andvalidated in several publications from independentstudies.31,61

The functional impairment subscale of the HBQ-Twas also used as a covariate in the final set of analyses.Impairment was included as a covariate to testwhether the hypothesized effect of PO psychiatricdisorder on school age relational aggression behaviorsremained significant after controlling for the potentialeffects of social impairment associated with children’s
psychiatric disorder present at school age. The HBQ-T g


impairment subscale uses teacher report to measurefunctional impairment that children are exhibiting inthe classroom. This subscale of the HBQ-T includesseven items that are rated using a likert scale (0 �none, 1 � a little, and 2 � a lot). Studies reporting the

sychometric properties of this subscale suggest that itas moderate to strong internal consistency and ac-eptable test–retest reliability across reporters and ageanges.

PO DSM-IV Psychiatric DisordersThe Preschool Age Psychiatric Assessment (PAPA) isan interviewer-based semi-structured diagnostic inter-view with established test–retest reliability that isdesigned for use in caregivers of children 2.0 to 6.0years of age.65 Although psychometric properties ofthe PAPA have been published only for children up toage 6 years, it is important to note that the PAPA hasbeen successfully used in children up to age 8.0 by anumber of research groups. The PAPA includes allrelevant DSM-IV criteria and their age-appropriatemanifestations. Diagnoses are derived by computeralgorithms that apply all of the DSM-IV criteria (withhe exception of duration criteria for MDD). The PAPAates the intensity, frequency, and duration of symp-oms as well as impairment from symptoms in threeeparate contexts (i.e., at home, at school, and else-here). Interviewers undergo 5- to 7-day training, andractice assessments are done until proficiency ischieved. Interviews were audio-taped for later qual-ty control and interviewer calibration. A master codereviewed 20% of each interviewer’s PAPA assess-ents; when discrepancies arose, items were re-coded

n consultation with a senior child psychiatrist. Toaintain high levels of interviewer reliability, weekly

oding meetings were conducted with a “master” raters recommended by the authors of the measure.

Data AnalysesOne-way univariate analysis of variance tests, �2 anal-

ses, and correlation analyses were conducted to ex-mine variation and/or differences in relational ag-ression during preschool and elementary school thatere associated with demographic variables. To exam-

ne the stability of children’s mean aggressor andictim scores at preschool and school age, Pearsonorrelation matrices were calculated. Multinomial lo-istic regression analyses were conducted to testhether children diagnosed with PO psychiatric dis-

rders were significantly more or less likely thanealthy peers to be classified as aggressors, victims,ggressive-victims, or nonaggressors/nonvictims dur-ng preschool and/or elementary school. For the finalet of analyses, multinomial logistic regression analy-es were conducted using covariates previously foundo influence children’s involvement in relational ag-
ression during elementary school. The following co-
893www.jaacap.org

vssacwmabawscwsaswsfiwswtwvsa

PDAPfp

a.

tcbpptbs

BELDEN et al.

variates were tested: children’s mean aggressor scoresand victim scores obtained at preschool, schoolchil-dren’s current level of functional impairment (teacherreport on the HBQ-T), as well as the total number ofdisruptive, anxiety, and depression symptoms experi-enced at school-age. The criterion variable for eachmodel was children’s role in relational aggression asan aggressive-victim versus nonaggressor/nonvictimduring elementary school.

RESULTSDemographic and Diagnostic Characteristics atBaseline and Preschoolers’ Relational AggressionChildren’s roles in relational aggression (i.e., ag-gressor, victim, aggressive-victim or nonaggres-sor/nonvictim) during preschool and/or ele-mentary school did not differ in relation tochildren’s gender, age, ethnic origin, family grossincome or their primary caregivers’ highest levelof education achieved. At baseline, 74 childrenhad a diagnosis of 1 or more preschool-onsetpsychiatric disorder. A total of 42 children had adisruptive disorder that, for the current study,included ADHD as well as oppositional defiantdisorder (ODD) and/or conduct disorder (CD).A total of 37 children had an anxiety disorderthat included generalized anxiety disorder(GAD), social anxiety disorder (SAD), and/orpost-traumatic stress disorder (PTSD); 41 chil-dren had been diagnosed with MDD. The re-maining 72 preschoolers had no psychiatric dis-orders and comprised the healthy comparisongroup for the following analyses. Table 1 pro-vides a further breakdown of diagnostic groupand comorbidity.

Descriptive Analyses Examining the Stability ofChildren’s Involvement in Relational AggressionPearson correlations indicated that children’s ag-gressor scores at preschool were significantlyassociated with their aggressor scores at schoolage (r � .34; p � .001). Similarly, children’s meanscores for the victim subscale of the HBQ-T atpreschool were significantly associated with theirmean scores on the victim subscale 2 years laterat school age (r � .31; p � .01). Descriptiveanalyses revealed that approximately 60% ofpreschool nonaggressor/nonvictims continuedto be classified this way by teachers when mea-sured again 24 months later. The remaining 40%
of preschoolers classified as nonaggressor/non- s
JOURN

894 www.jaacap.org

ictim had a different classification once they werechool age, 17% became aggressors, 8% were clas-ified as victims at school age, and 15% wereggressive-victims at school age. In all, 45% ofhildren classified as aggressors during preschoolere classified as nonaggressor/nonvictim wheneasured at school age. Of the children classified

s aggressors during preschool, 20% went on toe classified as aggressors once they were schoolge. The remaining 35% of preschool aggressorsere classified as aggressive-victims when mea-

ured at school age. Results indicated that 36% ofhildren classified as victims during preschoolere classified as nonaggressor/nonvictims at

chool age. Of the remaining children classifieds victims during preschool, 14% became aggres-ors, 22% retained their victim status, and 28%ere classified as aggressive-victims when mea-

ured at school age. Among the children classi-ed as aggressive-victims during preschool, 36%ere also classified as aggressive victims at

chool age. Of the remaining 64% of childrenho were aggressive-victims during preschool,

heir classifications changed as follows once theyere school age: 29% were nonaggressor/non-

ictim, 21% were aggressors, and 14% were clas-ified as victims based on teacher-report at schoolge.

resence versus Absence of PO Psychiatricisorders and Children’s Roles in Relationalggression during Preschool and at School Agereschoolers’ roles in relational aggression dif-

ered significantly between diagnostic groups atreschool [�2 (3, n � 146) � 6.68, p � .05].

Compared with healthy peers, preschoolers diag-nosed with one or more PO psychiatric disorderwere significantly more likely to be classified intoone of the following mutually exclusive groupsduring preschool: aggressive-victim (odds ratio[OR] � 3.75, 95% CI � 1.22-11.23, p � .05),ggressor (OR � 3.61, 95% CI � 1.06-12.87, p �05), or victim (OR � 3.64, 95% CI � 1.04-12.87,p � .05) when using nonaggressor/nonvictim ashe reference group. Most notably, findings indi-ated that among all preschoolers identified aseing aggressive-victims (n � 21), 76% had POsychiatric disorder(s). In sum, children with POsychiatric disorders were on average at least

hree times as likely as healthy same age peers toe classified by teachers as being either aggres-ors, victims, or aggressive-victims during pre-
chool.


CCnpp“ar

dovot


Children’s relational aggression roles weremeasured again 2 years after their baseline diag-nostic assessment when they were in elementaryschool. Schoolchildren’s relational aggressionroles during elementary school differed signifi-cantly in relation to their history of PO psychiat-ric disorder(s) [�2 (3, n � 121) � 10.46, p � .01].Schoolchildren diagnosed with PO psychiatricdisorders compared to schoolmates with no his-tory of PO psychiatric disorders were twice aslikely to be classified as aggressors than as non-aggressors/nonvictims (OR � 2.88, 95% CI �1.01-7.80, p � .05). Schoolchildren with PO psy-chiatric disorder(s) were four times as likelyas healthy peers to be classified as aggressive-victims compared to a nonaggressors/nonvic-tims (OR � 4.21, 95% CI � 1.60-11.09, p � .01).Schoolchildren’s risk for being classified as vic-tims of relational aggression did not differ signif-icantly between diagnostic groups.

PO Disruptive, Anxiety, or Depressive Disordersand Children’s Relational Aggression Rolesduring Preschool and Elementary SchoolPO Disruptive Disorders. Preschoolers’ roles in re-lational aggression differed between healthy anddisruptive disordered groups [�2(3, n � 116) �10.94, p � .01]. Compared with healthy peers,preschoolers with PO disruptive disorder(s) weresignificantly more likely to be classified as ag-

FIGURE 1 Preschool-onset (PO) disruptive disorders anschool age. Note: NS � not significant; OR � odds ratio

gressive-victims than aggressors (OR � 4.51, 95%



I � 1.17-17.37, p � .05), victims (OR � 3.20, 95%I � 0.89-12.02, p � .05), or nonaggressor/onvictims. It is important to note that disruptivereschoolers were equally as likely as healthyreschoolers to be classified by their teachers aspure aggressors” or “pure victims” when non-ggressors/nonvictims status was used at theeference group (Figure 1).

As seen in Figure 1, when measured againuring elementary school, schoolchildren previ-usly diagnosed with PO disruptive disorder(s)ersus schoolchildren who were healthy through-ut preschool differed significantly in their rela-ional aggression roles [�2 (3, n � 100) � 17.89,

p � .001). Schoolchildren diagnosed with POdisruptive disorders were more than eight timesas likely as children in the healthy comparisongroup to be classified as aggressive-victims.Schoolchildren with PO disruptive disorderswere more than five times as likely as healthypeers to be identified by teachers as aggressors.In contrast, schoolchildren with PO disruptivedisorders were four times less likely than healthypeers to be classified as victims when usingnonaggressor/nonvictim as the reference group.PO Anxiety Disorders. Preschoolers’ relational ag-gression roles did not differ significantly betweenanxiety disordered and healthy preschoolers(Figure 2). When testing PO anxiety disorders aspredictors of relational aggression roles at school

ildren’s relational aggression status at preschool and� .05, **p � .01, ***p � .001.

d ch. *p

age, the overall �2 testing for proportional differ-

895www.jaacap.org

BELDEN et al.

ences in the relational aggression roles of school-children who were healthy during preschoolversus classmates with a history of PO anxietydisorders was non significant (�2 (3, n � 89) �5.47, p � .07). However, one pairwise result wassignificant and is worth noting. That is, school-children diagnosed with PO anxiety disorderswere four times as likely as schoolmates whowere healthy preschoolers to be classified as

FIGURE 2 Preschool-onset (PO) anxiety disorders andschool age. Note: NS � not significant; OR � odds ratio

FIGURE 3 Preschool-Onset (PO) major depressive disopreschool and school age. Note: NS � not significant; O

JOURN

896 www.jaacap.org

aggressive-victims than nonaggressors/nonvic-tims (Figure 2).PO MDD. Healthy and PO MDD preschoolersdid not differ in their relational aggression rolesduring preschool (Figure 3). Roles in relationalaggression at school age differed between school-children who were healthy as preschoolers com-pared with schoolmates previously diagnosedwith PO MDD [�2 (3, n � 90) � 8.55, p � .05].

ren’s relational aggression status at preschool and� .05.

(MDD) and children’s relational aggression status atodds ratio. *p � .05.

child. *p

rderR �




School-age children with PO MDD were fivetimes more likely than the healthy comparisongroup to be classified as aggressive-victims thannonaggressor/nonvictims during elementaryschool (Figure 3).

PO Psychiatric Disorders as Predictors of LaterAggressive-Victim Status: Controlling forRelational Aggression Roles at Preschool,Psychopathology, and Functional Impairment atSchool AgeThe above findings indicated that schoolchildrenwith a history of PO disruptive, PO anxiety, andPO MDD were significantly more likely thantheir healthy peers to be classified as aggressive-victims than nonaggressor/nonvictims. To fur-ther explore this finding, a series of multinomiallogistic regression analyses were conducted totest whether PO psychiatric disorders continuedto predict schoolchildren’s relational aggressionroles when covarying for children’s relationalaggression behaviors during preschool as well asschoolchildren’s current experience of psychiat-ric symptoms and associated functional impair-ments. The following covariates were tested:schoolchildren’s mean aggressor and victimscores obtained at preschool, schoolchildren’scurrent level of functional impairment (teacher

FIGURE 4 Preschool-onset (PO) psychiatric disorders aafter covarying for school-age psychiatric disorder severiscores at preschool. Note: Results illustrated included theon the MacArthur Health and Behavior Questionnaire—Ton HBQ-T at preschool age; (3) Functional impairment scdisorder symptoms endorsed at school age; (5) Total num(6) Total number of major depressive disorder (MDD) sym**p � .01.



report), as well as the total number of disruptive,anxiety, depression symptoms experienced atschool age. The criterion variable for each modelwas children’s role in relational aggression as anaggressive-victim versus nonaggressor/nonvic-tim during elementary school.

Results indicated that PO disruptive disorderscontinued to be a significant predictor of theaggressive-victim classification during elemen-tary school (p � .01). Specifically, schoolchildrenwith PO disruptive disorder were eight times aslikely as schoolchildren who were healthy pre-schoolers to be classified as aggressive-victimsafter including covariates in the model (Figure 4).Similarly, schoolchildren with a history of POMDD were significantly more likely than thehealthy comparison group to be classified asaggressive-victims [�2 (df 21, n � 84) � 37.53, p �.01]. Schoolchildren diagnosed with PO MDDwere six times as likely as schoolchildren whowere healthy as preschoolers to be classified asaggressive-victims after covariates were includedin the model (Figure 4). Schoolchildren with ahistory of PO anxiety disorders compared withschoolchildren who were healthy preschoolerswere also significantly more likely to be classifiedby teachers as aggressive-victims during elemen-tary school [�2 (df 21, n � 76) � 42.81, p � .01].

kelihood of being an aggressive-victim at school agenctional impairment, and mean aggression and victiming covariates in the final model: (1) Aggressor score

er Version (HBQ-T) at preschool age; (2) Victim scoren HBQ-T at school age; (4) Total number of disruptive

of anxiety disorder symptoms endorsed at school-age;s endorsed at school age. OR � odds ratio. *p � .05,

nd lity, fufolloweachore oberptom

897www.jaacap.org

BELDEN et al.

Compared with schoolchildren who were healthyin preschool, schoolchildren diagnosed with POanxiety disorders were nine times more likely to beaggressive-victims than nonaggressor/nonvictimsafter including covariates.

DISCUSSIONThe association between childhood-onset psychi-atric symptoms/disorders and involvement inovert or physical forms of peer aggression (e.g.,bullying) throughout development has been wellestablished.59,61 However, a growing body ofliterature suggests that covert and nonphysicalforms of aggression, such as relational aggres-sion, have equally deleterious effects on chil-dren’s development, occur at relatively high fre-quencies, and start as early as the preschoolperiod of development. With a few exceptions, arelatively limited number of studies have exam-ined whether PO psychiatric disorders demon-strate associations with nonphysical forms ofpeer aggression consistent with findings examin-ing physical forms of peer aggression (e.g., bul-lying). The aim of the present study was toexamine whether preschool onset psychiatric dis-orders were concurrently related to preschoolers’involvement in relational aggression and/or pre-dicted their later involvement in relational ag-gression at school age.

Despite overlap between peer focused aggres-sive behaviors and ODD symptoms (e.g., spitefuland vindictive) as well as CD symptoms (e.g.,bullying), findings have illustrated that as few as6% of children with ODD and/or CD also hadhigh relational aggression scores (i.e., 1 SD abovethe sample mean). In contrast, the same studyfound that 14% of children without a diagnosis ofODD and/or CD had high relational aggressionscores.71 Despite children diagnosed with ODDand/or CD being more likely to be involved inrelational aggression, the majority of youth en-gaged in high levels of relational aggression donot meet symptom criteria for ODD and/or CD.

In the current study preschoolers diagnosedwith ADHD, ODD, and/or CD were no morelikely than healthy same-age peers to be classi-fied as “pure-aggressors or victims” of relationalaggression during preschool. However, disrup-tive disordered preschoolers were six times aslikely as healthy same age peers to be classifiedas aggressive-victims. This suggests that pre-
schoolers with disruptive disorders are fre-
JOURN

898 www.jaacap.org

quently the perpetrator and victim of relationalaggression. Once disruptive-disordered pre-schoolers were schoolchildren, they were signif-icantly more likely than schoolmates without ahistory of PO disruptive disorders to be “pure-aggressors” and aggressive-victims but signifi-cantly less likely to be victims of relationalaggression. This finding is consistent with re-sults from older children, which have demon-strated that as aggressors grow in physicalstrength and join social groups with other ag-gressors, their likelihood of becoming victimizeddecreases. This may be the result of other aggres-sors becoming more fearful and avoidant ofconfrontation with known aggressors.72

Arguably the most interesting and novel find-ings to emerge in the current study were childrendiagnosed with PO anxiety and/or depressivedisorders were no more likely than healthy pre-schoolers to be involved in relational aggressionas aggressors or victims during preschool or atschool age. However, this same group of children(with PO anxiety and/or depressive disorders)were more than six times as likely to be classifiedaggressive-victims at school age compared tohealthy preschoolers. These results emerged aftercontrolling for children’s involvement in rela-tional aggression as aggressors and victims dur-ing preschool as well as their current disruptive,anxiety, and depressive symptoms and theirfunctional impairment scores at school age. Thehigh risk for this unique outcome (aggressive-victim) is also consistent with prior findings thathave demonstrated during the beginning years ofelementary school, aggressive-victims show sig-nificantly greater internalizing symptoms than ag-gressors and victims of relational aggression. Incontrast to these findings, studies that measureaggressor and victim scores/status only (omit-ting an aggressive-victim score/group) in rela-tion to children’s concurrent anxiety and de-pressive symptoms typically demonstrate thatincreases in anxiety and depression are positivelycorrelated with children’s victim scores. Thesefindings related to early-onset anxiety and de-pression in the context of the extant literatureraise several interesting questions for future re-search. Of particular interest is the need forfuture studies examining the possibility of differ-ing trajectories for children’s involvement in re-lational aggression as a function of varying on-sets as well as current severity of specific
psychiatric disorders.


wapaoarsfiHrctl

tsdsimisgtirttticptmptmpvssosiiaepcri


Findings examining gender differences in re-lational aggression during preschool have beenmixed.9,17,23,29 Crick et al. found that preschoolgirls were more likely than boys to be involved inrelational aggression.26 In contrast, other studiesincluding the current study did not detect genderdifferences.39,64 When examining gender differ-ences in a sample of youth 9 to 17 years old,Keenan et al. concluded that gender similarities,and not differences in levels of relational aggres-sion, were the norm.71 Although speculative, thelack of gender differences in the present studymay have been related to the high percentage ofchildren with PO psychiatric disorders in thesample studied. It is possible that when includ-ing children with disruptive and other POpsychiatric disorders boys and girls’ use ofrelational forms of aggression may be moreequally distributed.

The present study has several limitations. Firstthe HBQ-T has no standardized cut-points fordetermining children’s involvement in relationalaggression as either perpetrator or victim. Con-sistent with prior literature, a 20% cutoff wasused at each time point. Thus, children wereassigned aggressor/victim status based on ob-served levels of aggression in the existing sam-ple. Nonetheless, children’s aggressor/victimstatus at time 1 was significantly predictive oftheir aggressor/victim status when measured 24months later and was associated with later men-tal health problems. Second, there were highrates of psychiatric disorder comorbidity in thepresent sample. Given the relatively small sam-ple size examining co-occurring disorders in re-lation to children’s aggressor/victim status re-sulted in group sizes too small for statisticalcomparisons. As a result, the current findingsshould be interpreted with a degree of caution.Third, children’s DSM-IV diagnostic group sta-tus (based on primary caregiver reports) as well astheir aggressor/victim status (based on teacherreport) was measured using a single informant.Although a multi-informant method is preferredfor both constructs, caregiver report for assessingresearch based diagnostic status in preschool-agechildren remains the current standard in thefield. Along these same lines, peer ratings andobservational measures, in conjunction withteacher ratings, would have been ideal for assess-ing preschoolers and schoolchildren’s involve-ment in relational aggression. Given the young
age of the sample, there are numerous challenges i


hen using peer-based nominations of relation-lly aggressive behaviors. Arguably, teacher re-ort of preschoolers’ involvement in relationalggression may provide the most reliable sourcef information related to these forms of peerggression in very young children. Finally, theecruitment methods used to obtain the presentample limits the generalizability of the currentndings to the general population of children.owever, findings from the current study war-

ant future studies in community-based and orlinical samples to test the generalizability ofhese findings to both healthy and clinical popu-ations of young children.

Findings suggest that increased attention tohe detection of psychiatric disorders in pre-chool populations, which currently remain un-etected in the vast majority of affected pre-choolers, may be a promising strategy fordentifying those at high risk for later involve-

ent in relational aggression as well as provid-ng a target for preventative intervention forchoolchildren’s involvement in relational ag-ression. These findings underscore two poten-ially key public health principles. The first is themportance of identifying and treating psychiat-ic disorders during the preschool period, givenhe established association to poorer peer rela-ionship outcomes.30,73,74 Second is the impor-ance of evaluating relational aggression behav-ors as early as the preschool period, given theirlear manifestation at this early juncture and theossibility of more effective intervention during

his time of rapid social and emotional develop-ent.40 In addition to this, and relevant to the

revention of school-age relational aggression, ishe need to account for history of early-onset

ental disorders in preschool populations as aossible mechanism to prevent later aggressive-ictim behaviors. That is, the current findingsuggest that the manifestation of psychiatricymptoms in preschool children may provide anbservable and targetable antecedent to moreevere forms of relational aggression (i.e., bully-ng) in schoolchildren. Based on the current find-ngs we conclude that efforts to reduce relationalggression in schools should focus on the earli-st possible detection of risk for or early onsetsychopathology. Interventions designed to spe-ifically target these subgroups, and focus onelieving psychiatric symptoms, appear to bendicated. Such strategies may in turn minim-
ze the occurrence of relational aggression at
899www.jaacap.org

BELDEN et al.

school age, an increasingly serious public healthconcern.20 &

Accepted June 28, 2012.

Drs. Belden, Gaffrey, and Luby are with the Early Emotional Develop-ment Program (EEDP) at the Washington University School of Medicinein St. Louis.

This study was funded by the National Institute of Mental Health(NIMH) grants R01 MH64769-01 (J.L.) and K01MH090515(A.B.).

The authors gratefully acknowledge Edward Spitznagel, Ph.D., ofWashington University–St. Louis for statistical consultation and MarilynEssex, Ph.D., of the University of Wisconsin for assistance with earlier
versions of this manuscript. They are also grateful to the EEDP staff, our
prosocial behavior in the prediction of. Child Dev. 1996;67:2317-2327.

JOURN

900 www.jaacap.org

preschool participants and their parents, and community recruiting siteswhose participation and cooperation made this research possible.

Disclosure: Dr. Luby has received grant or research support from theNational Insitute of Mental Health, the National Alliance for Researchon Schizophrenia and Depression, the Communities Healing Adoles-cent Depression and Suicide Coalition, and the Sidney R. BaerFoundation. She has served as a consultant to the Food and DrugAdministration Advisory Board. Drs. Belden and Gaffrey report nobiomedical financial interests or potential conflicts of interest.

Correspondence to Andy C. Belden, Ph.D., Washington UniversitySchool of Medicine, Department of Psychiatry, Box 8134, 660 S.Euclid, St. Louis, MO 63110; e-mail: [email protected]

0890-8567/$36.00/©2012 American Academy of Child andAdolescent Psychiatry

http://dx.doi.org/10.1016/j.jaac.2012.06.018

REFERENCES1. Berkowitz L. Pain and aggression—some findings and implica-

tions. Motiv Emot. 1993;17:277-293.2. Loeber R. Development and risk-factors of juvenile antisocial-

behavior and delinquency. Clin Psychol Rev. 1990;10:1-41.3. Parker JG, Asher SR. Peer relations and later personal adjust-

ment—are low-accepted children at risk? Psychol Bull. 1987;102:357-389.

4. Hanish LD, Guerra NG. Aggressive victims, passive victims, andbullies: developmental continuity or developmental change?Merrill-Palmer Q. 2004;50:17-38.

5. Barker EDP, Boivin MP, Brendgen MP, et al. Predictive validityand early predictors of peer-victimization trajectories in pre-school. Arch Gen Psychiatry. 2008;65:1185-1192.

6. Burk LR, Armstrong JM, Park J-h, Zahn-Waxler C, Klein MH,Essex MJ. Stability of early identified aggressive victim status inelementary school and associations with later mental healthproblems and functional impairments. J Abnorm Child Psychol.2010;39:225-238.

7. Block JH. Differential premises arising from differential socializa-tion of the sexes—some conjectures. Child Dev. 1983;54:1335-1354.

8. Monks CE, Smith PK, Swettenham J. Aggressors, victims, anddefenders in preschool: peer, self-, and teacher reports. Merrill-Palmer Quarterly. 2003;49:453-469.

9. Bjorkqvist K, Lagerspetz KMJ, Kaukiainen A. Do girls manipulateand boys fight? Developmental-trends in regard to direct andindirect aggression. Aggress Behav. 1992;18:117-127.

10. Crick NR, Bigbee MA, Howes C. Gender differences in children’snormative beliefs about aggression: how do I hurt thee? Let mecount the ways. Child Dev. 1996;67:1003-1014.

11. Baillargeon RH, Zoccolillo M, Keenan K, et al. Gender differ-ences in physical aggression: a prospective population-basedsurvey of children before and after 2 years of age. Dev Psychol.2007;43:13-26.

12. Ostrov JM, Keating CF. Gender differences in preschool aggres-sion during free play and structured interactions: an observa-tional study. Soc Dev. 2004;13:255-277.

13. Archer J, Coyne SM. An integrated review of indirect, relational,and social aggression. Pers Soc Psychol Rev. 2005;9:212-230.

14. Ostrov J, Crick N. How recent developments in the study ofrelational aggression and close relationships in early childhoodadvance the field. J Appl Dev Psychol. 2006;27:189-192.

15. Crick NR, Grotpeter JK. Relational aggression, gender, and social-psychological adjustment. Child Dev. 1995;66:710-722.

16. Crick NR. Relational aggression—the role of intent attributions,feelings of distress, and provocation type. Dev Psychopathol.1995;7:313-322.

17. Ostrov JM, Godleski SA. Toward an integrated gender-linkedmodel of aggression subtypes in early and middle childhood.Psychol Rev. 2010;117:233-242.

18. Crick NR. The role of overt aggression, relational aggression, and

19. Crick NR, Ostrov JM, Werner NE. A longitudinal study ofrelational aggression, physical aggression, and children’s social-psychological adjustment. J Abnorm Child Psychol. 2006;34:131-142.

20. Fite PJ, Stoppelbein L, Greening L, Preddy TM. Associationsbetween relational aggression, depression, and suicidal ideationin a child psychiatric inpatient sample. Child Psychiatry HumDev. 2011;42:666-678.

21. Keenan K, Shaw DS. The development of aggression intoddlers—a study of low-income families. J Abnorm Child Psy-chol. 1994;22:53-77.

22. Wakschlag LS, Briggs-Gowan MJ, Carter AS, et al. A developmen-tal framework for distinguishing disruptive behavior from nor-mative misbehavior in preschool children. J Child Psychol Psy-chiatry. 2007;48:976-987.

23. Card NA, Stucky BD, Sawalani GM, Little TD. Direct and indirectaggression during childhood and adolescence: a meta-analyticreview of gender differences, intercorrelations, and relations tomaladjustment. Child Dev. 2008;79:1185-1229.

24. Crick NR, Grotpeter JK. Children’s treatment by peers: victims ofrelational and overt aggression. Dev Psychopathol. 1996;8:367-380.

25. Crick NR, Casas JF, Mosher M. Relational and overt aggression inpreschool. Dev Psychol. 1997;33:579-588.

26. Crick NR, Casas JF, Ku HC. Relational and physical forms of peervictimization in preschool. Dev Psychol. 1999;35:376-385.

27. Ostrov JM, Crick NR, Keating CF. Gender-biased perceptions ofpreschoolers’ behavior: how much is aggression and prosocialbehavior in the eye of the beholder? Sex roles. 2005;52:393-398.

28. Schwartz D, Dodge KA, Coie JD, et al. Social-cognitive andbehavioral correlates of aggression and victimization in boys’play groups. J Abnorm Child Psychol. 1998;26:431-440.

29. Pellegrini A, Roseth C. Relational aggression and relationships inpreschoolers: a discussion of methods, gender differences, andfunction. J Appl Dev Psychol. 2006;27:269-276.

30. Sebanc AM. The friendship features of preschool children: linkswith prosocial behavior and aggression. Soc Dev. 2003;12:249-268.

31. Crick NR, Ostrov JM, Burr JE, Cullerton-Sen C, Jansen-Yeh E,Ralston P. A longitudinal study of relational and physical aggres-sion in preschool. J Appl Dev Psychol. 2006;27:254-268.

32. McEvoy MA, Estrem TL, Rodriguez M, Olson ML. Assessingrelational and physical aggression among preschool children:intermethod agreement. Topics Early Child Spec Educ. 2003;23:53-63.

33. Bonica C, Arnold DH, Fisher PH, Zeljo A, Yershova K. Relationalaggression, relational victimization, and language developmentin preschoolers. Soc Dev. 2003;12:551-562.

34. Ostrov JM. Deception and subtypes of aggression during earlychildhood. J Exp Child Psychol. 2006;93:322-336.

35. Ostrov JM, Crick NR. Forms and functions of aggression during
early childhood: a short-term longitudinal study. Sch PsycholRev. 2007;36:22-43.


mailto:[email protected]

http://dx.doi.org/10.1016/j.jaac.2012.06.018


36. Garner PW, Lemerise EA. The roles of behavioral adjustment andconceptions of peers and emotions in preschool children’s peervictimization. Dev Psychopathol. 2007;19:57-71.

37. Casas JF, Weigel SM, Crick NR, et al. Early parenting andchildren’s relational and physical aggression in the preschool andhome contexts. J Appl Dev Psychol. 2006;27:209-227.

38. Ostrov JM, Woods KE, Jansen EA, Casas JF, Crick NR. Anobservational study of delivered and received aggression, gender,and social-psychological adjustment in preschool: “This whitecrayon doesn’t work . . .”. Early Child Res Q. 2004;19:355-371.

39. Hart CH, Nelson DA, Robinson CC, Olsen SF, McNeilly-ChoqueMK. Overt and relational aggression in Russian nursery-school-age children: parenting style and marital linkages. Dev Psychol.1998;34:687-697.

40. Crick NR. The role of overt aggression, relational aggression, andprosocial behavior in the prediction of children’s future socialadjustment. Child Dev. 1996;67:2317-2327.

41. Vaillancourt T, Brendgen M, Boivin M, Tremblay RE. A longitu-dinal confirmatory factor analysis of indirect and physical aggres-sion: evidence of two factors over time? Child Dev. 2003;74:1628-1638.

42. Herrenkohl TI, McMorris BJ, Catalano RF, Abbott RD, HemphillSA, Toumbourou JW. Risk factors for violence and relationalaggression in adolescence. J Interpers Violence. 2007;22:386-405.

43. Monks CP, Smith PK, Swettenham J. Psychological correlates ofpeer victimisation in preschool: social cognitive skills, executivefunction and attachment profiles. Aggress Behav. 2005;31:571-588.

44. van Lier PAC, Crijnen AAM. Trajectories of peer-nominatedaggression: risk status, predictors and outcomes. J Abnorm ChildPsychol. 2005;33:99-112.

45. Campbell SB, Spieker S, Burchinal M, et al. Trajectories of aggres-sion from toddlerhood to age 9 predict academic and socialfunctioning through age 12. J Child Psychol Psychiatry. 2006;47:791-800.

46. Kumpulainen K. Children involved in bullying at elementaryschool age: their psychiatric symptoms and deviance in adoles-cence An epidemiological sample. Child Abuse Neglect. 2000;24:1567-1577.

47. Scholte RHJ, Engels RCME, Overbeek G, de Kemp RAT, Hase-lager GJT. Stability in bullying and victimization and its associa-tion with social adjustment in childhood and adolescence. JAbnorm Child Psychol. 2007;35:217-228.

48. Murray-Close D, Ostrov JM. A longitudinal study of forms andfunctions of aggressive behavior in early childhood. Child Dev.2009;80:828-842.

49. Dodge KA, Lochman JE, Harnish JD, Bates JE, Pettit GS. Reactiveand proactive aggression in school children and psychiatricallyimpaired chronically assaultive youth. J Abnorm Psychol. 1997;106:37-51.

50. Mathieson LC, Crick NR. Reactive and proactive subtypes ofrelational and physical aggression in middle childhood: links toconcurrent and longitudinal adjustment. Sch Psychol Rev. 2010;39:601-611.

51. Pellegrini AD. Bullies and victims in school: a review and call forresearch. J Appl Dev Psychol. 1998;19:165-176.

52. Griffin RS, Gross AM. Childhood bullying: current empiricalfindings and future directions for research. Aggress ViolentBehav. 2004;9:379-400.

53. Orpinas P, Horne AM, Staniszewski D. School bullying: changingthe problem by changing the school. School Psychol Rev. 2003;32:431-444.

54. Nansel TR, Overpeck M, Pilla RS, Ruan WJ, Simons-Morton B,Scheidt P. Bullying behaviors among US youth: prevalence and
association with psychosocial adjustment. JAMA. 2001;285:2094-2100.


55. Juvonen J, Graham S, Schuster MA. Bullying among youngadolescents: the strong, the weak, and the troubled. Pediatrics.2003;112:1231-1237.

56. Arseneault L, Walsh E, Trzesniewski K, Newcombe R, Caspi A,Moffitt TE. Bullying victimization uniquely contributes to adjust-ment problems in young children: a nationally representativecohort study. Pediatrics. 2006;118:130-138.

57. Juvonen J, Graham S, Schuster MA. Bullying among youngadolescents: the strong, the weak, and the troubled. Pediatrics.2011;112:1231-1237.

58. Perren S, Alsaker FD. Social behavior and peer relationships ofvictims, bully-victims, and bullies in kindergarten. J Child Psy-chol Psychiatry. 2006;47:45-57.

59. Kumpulainen K, Rasanen E. Children involved in bullying atelementary school age: their psychiatric symptoms and deviancein adolescence—an epidemiological sample. Child Abuse Ne-glect. 2000;24:1567-1577.

60. Swearer SM, Song SY, Cary PT, Eagle JW, Mickelson WT.Psychosocial correlates in bullying and victimization: the rela-tionship between depression, anxiety, and bully/victim status. JEmot Abuse. 2001;2:95-121.

61. Burk LR, Armstrong JM, Park J-H, Zahn-Waxler C, Klein MH,Essex MJ. Stability of early identified aggressive victim status inelementary school and associations with later mental healthproblems and functional impairments. J Abnorm Child Psychol.2011;39:225-238.

62. Haynie DDL. Bullies, victims, and bully/victims: distinct groupsof at-risk youth. J Early Adolesc. 2001;21:29-49.

63. Solberg ME, Olweus D. Prevalence estimation of school bullyingwith the Olweus Bully/Victim Questionnaire. Aggress Behav.2003;29:239-268.

64. Burk LR, Park J-H, Armstrong JM, et al. Identification of earlychild and family risk factors for aggressive victim status in firstgrade. J Abnorm Child Psychol. 2008;36:513-526.

65. Egger HL, Erkanli A, Keeler G, Potts E, Walter BK, Angold A.Test-retest reliability of the Preschool Age Psychiatric Assessment(PAPA). J Am Acad Child Adolesc Psychiatry. 2006;45:538-549.

66. Luby JL, Si X, Belden AC, Tandon M, Spitznagel E. Preschooldepression homotypic continuity and course over 24 months.Arch Gen Psychiatry. 2009;66:897-905.

67. Merrell KW, Gueldner BA, Ross SW, Isava DM. How effective areschool bullying intervention programs? A meta-analysis of inter-vention research. Sch Psychol Q. 2008;23:26-42.

68. Luby JL, Belden AC, Pautsch J, Si X, Spitznagel E. The clinicalsignificance of preschool depression: impairment in functioningand clinical markers of the disorder. J Affect Disord. 2009;112:111-119.

69. Luby J, Heffelfinger A, Mrakotsky C, Hildebrand T. PreschoolFeelings Checklist. St. Louis, MO: Washington University; 1999.

70. Essex MJ, Boyce WT, Goldstein LH, Armstrong JM, Kraemer HC,Kupfer DJ. The confluence of mental, physical, social, andacademic difficulties in middle childhood II: developing theMacArthur Health and Behavior Questionnaire. J Am AcadChild Adolesc Psychiatry. 2002;41:588-603.

71. Keenan K, Coyne C, Lahey BB. Should relational aggression beincluded in DSM-V? J Am Acad Child Adolesc Psychiatry.2008;47:86-93.

72. Boivin M, Petitclerc A, Feng B, Barker ED. The developmentaltrajectories of peer victimization in middle to late childhood andthe changing nature of their behavioral correlates. Merrill-PalmerQ. 2010;56:231-260.

73. Werner NE, Crick NR. Maladaptive peer relationships and thedevelopment of relational and physical aggression during middlechildhood. Soc Dev. 2004;13:495-514.

74. Mesman J, Koot HM. Early preschool predictors of preadolescent
internalizing and externalizing DSM-IV diagnoses. J Am AcadChild Adolesc Psychiatry. 2001;40:1029-1036.
901www.jaacap.org

relational aggression in children with preschool-onset ... · relational aggression in children...

Documents