CCiittyy UUnniivveerrssiittyy ooff NNeeww YYoorrkk ((CCUUNNYY)) CCUUNNYY AAccaaddeemmiicc WWoorrkkss Publications and Research City College of New York 2007 PPeeeerr VViiccttiimmiizzaattiioonn,, DDeepprreessssiioonn,, aanndd SSuuiicciiddaalliittyy iinn AAddoolleesscceennttss Anat Brunstein Klomek Columbia University Frank Marrocco Columbia University Marjorie Kleinman Columbia University Irvin Sam Schonfeld CUNY Graduate Center Madelyn S. Gould Columbia University How does access to this work benefit you? Let us know! More information about this work at: https://academicworks.cuny.edu/cc_pubs/303 Discover additional works at: https://academicworks.cuny.edu This work is made publicly available by the City University of New York (CUNY). Contact: [email protected] 166 Suicide and Life-Threatening Behavior 38(2) April 2008  2008 The American Association of Suicidology Peer Victimization, Depression, and Suicidiality in Adolescents Anat Brunstein Klomek, PhD, Frank Marrocco, PhD, Marjorie Kleinman, MS, Irvin Sam Schonfeld, PhD, MPH, and Madelyn S. Gould, PhD, MPH The association between specific types of peer victimization with depres- sion, suicidal ideation, and suicide attempts among adolescents was examined. A self-reportsurveywascompletedby2,342high-schoolstudents.Regressionanaly- ses indicated that frequent exposure to all types of peer victimization was related tohighriskofdepression,ideation,andattemptscomparedtostudentsnotvictim- ized. Infrequent victimization was also related to increased risk, particularly amongfemales.Themoretypesofvictimizationthehighertheriskfordepression and suicidality among both genders. Specific types of peer victimization are a potential risk factor for adolescent depression and suicidality. It is important to assess depression and suicidality among victimized students in order to develop appropriate intervention methods. Peer victimization is a prevalent problem mately 10 to 20% of high school students in among high school students (Brunstein Klo- theU.S.reportmoderateorfrequentvictim- mek, 2007; Nansel, Overpeck, Pilla, Ruan, ization (Klomek et al., 2007; Nansel et al., Simons-Marten, & Scheidt, 2001). Approxi- 2001). Nansel and colleagues’ national study assessed five specific types of victimization: belittlingaboutreligion/race;belittlingabout Dr.BrunsteinKlomekisanassistantpro- looks/speech; hitting, slapping, or punching; fessorofclinicalpsychologyatColumbiaUniver- rumor spreading or lying about the target; sity in the Division of Child and Adolescent Psy- and sexual comments/gestures (Nansel et al., chiatry;Dr.MarroccoisanInstructorofClinical Psychology in the Division of Child and Adoles- 2001). The prevalence of these types of vic- cent Psychiatry; Ms. Kleinman is a research sci- timization among adolescents ranged from 8 entist in the Division of Child and Adolescent to 20%; the most common being belittling Psychiatry; Dr. Schonfeld is a lecturer at Co- aboutlooks orspeech andthe leastprevalent lumbia University in the Department of Psychia- being belittling about religion or race tryandaProfessorattheCityCollegeoftheCity UniversityofNewYork;andDr.GouldisaPro- (Nanseletal.,2001).Beingbulliedviae-mail fessor at Columbia University in the Division of or the Internet (cyber victimization) is the ChildandAdolescentPsychiatryandDepartment newesttypeofvictimization(Patchin&Hin- ofEpidemiology(SchoolofPublicHealth),anda duja, 2006;Ybarra, 2004;Ybarra &Mitchell, researchscientistattheNewYorkStatePsychiat- 2004). The prevalence of this type of victim- ricInstitute. ThisprojectwassupportedbyNationalIn- zation varies, from 4 to 35% in adolescent stituteofMentalHealthgrantRO1-MH64632. samples(Patchin&Hinduja,2006;Ybarra& AddresscorrespondencetoAnatBrunstein Mitchell, 2004). Klomek, PhD, Division of Child & Adolescent The classification and definition of Psychiatry, Department of Psychiatry, Columbia specific types of victimization varies among University/ NYSPI, 1051 Riverside Drive, Unit 72,NewYork,NY10032;E-mail:klomeka@child studies (Wolke & Stanford, 1999), making psych.columbia.edu comparisonsamongstudiesdifficult(Hawker Klomek et al. 167 & Boulton, 2000). Overt victimization in- 2000; Kumpulainen, Rasanen, & Henttenen, cludes acts that are meant to harm a peer 1999; Mills, Guerin, Lynch, Daly, & Fitzpa- physically, verbal threats of such acts (e.g., trick, 2004; Neary & Joseph, 1994; Slee, threateningtobeatupapeer),andnamecall- 1995; van der Wal et al., 2003; Williams, ing (Crick & Grotpeter, 1995; Grotpeter & Chambers, Logan, & Robinson, 1996). In a Crick, 1996; Prinstein, Boergers, & Vern- meta-analysis, peer victimization was most berg, 2001). Similarly, direct victimization in- strongly related to depression in comparison cludesdiversephysicalandverbalmethodsof to other types of maladjustment (Hawker & victimization; for example, kicking, hitting, Boulton, 2000). Compared to nonvictims, threatening, name calling, and insulting (van victims are also found to exhibit high levels der Wal, de Wit, & Hirasing, 2003). Direct ofsuicidalideation(BrunsteinKlomek,2007; victimization has also included having be- Hold, Finkelhor, & Kantor, 2006; Kaltiala- longingstaken,havingliestoldaboutthetar- Heino, Rimpela, Marttunen, Rimpela, & get, having nasty tricks played on the target, Rantanen,1999;Rigby&Slee,1999;Roland, and having been threatened or blackmailed 2002; van der Wal et al., 2003) and are more (Wolke, Woods, Bloomfield, & Karstadt, likely to have attempted suicide (Brunstein 2000). Indirect victimization has been defined Klomek, 2007; Cleary, 2000; Eisenberg, as aggression that is enacted through third Neumark-Sztainer, & Story, 2003; Kim, Koh, parties (Bjorkqvist, 1994; Rivers & Smith, & Levanthal, 2005; Mills et al., 2004). 1994). It includes aspects of social isolation Studies have found significant interac- such as ignoring, excluding, and backbiting tions between gender and victimization with (van der Wal et al., 2003) as well as rejecting regard to depression and suicidality risk; the and having rumors spread about them (Bal- results, however, are not consistant. Kim et dry&Farrington,1999).Relationalvictimiza- al.(2005)foundthatfemalebutnotmalestu- tion includesbehaviors thatcause orthreaten dents who were victimized were at signifi- to damage peer relationships, particularly cantlygreaterriskforsuicidalideation.Simi- friendship and acceptance (Crick, Kasas, & larly, Katliala-Heino et al. (1999) found that Ku, 1999; Crick & Grotpeter, 1995, 1996; severe ideation was associated with frequent Wolke et al., 2000). Studies have compared victimization only among girls and there was different subsets of these constructs: direct no gender interaction for depressive symp- andindirectvictimization(Baldry&Farring- toms. van der Wal et al. (2003) found that ton, 1999; van der Wal et al., 2003); direct among both genders indirect victimization and relational victimization (Wolke et al., was associated with depression and suicidal 2000;vanderWaletal.,2003);overtandre- ideation. Direct victimization, however, was lational victimization (Prinstein et al., 2001); associated with depression and suicidal ide- physical, verbal, and indirect victimization ationamonggirlsbutnotboys.Inourearlier (Bjorkqvist, 1994; Rivers & Smith, 1994); study (Brunstein Klomek et al., 2007), we and physical and relational victimization found adifferent thresholdat whichbullying (Crick et al., 1999). Given this lack of clarity wasassociatedwithdepressionandsuicidality in the classification of peer victimization, in among females and males. Among females, the present study we examined discrete peer victimization at any frequency increased the victimizations as suggested by the World risk of depression, ideation, and attempts. HealthOrganization(WHO)studyonyouth Amongmales onlyfrequent victimizationin- health (Nansel et al., 2001). creased the risk of depression and ideation. Peer victims are found to manifest Most of the existing research on the moredepressivesymptomscomparedtonon- association between peer victimization with victims (Brunstein Klomek, 2007; Craig, depression and suicidality have examined 1998;Crick&Grotpeter,1996;Fekkes,Pijpers, overall victimization (Cleary, 2000; Eisen- & Verloove-Vanhorick, 2004; Hawker & berg et al., 2003; Kaltiala-Heino et al., 1999; Boulton, 2000; Kumpulainen & Rasanen, Kim et al., 2005; Mills et al., 2004; Rigby & 168 Adolescent Peer Victimization Slee,1999;Roland,2002;Slee,1995;vander participating students was 14.8 (1.2) years. Waletal.,2003).Onlyafewstudieshaveex- There were no significant differences be- aminedtheassociationbetweenspecifictypes tweenparticipantsandnonparticipantsinsex, of peer victimization with depression (Crick age, and race/ethnicity. & Grotpeter, 1996; Prinstein et al., 2001). Students were recruited with a waiver Specific types ofpeer victimization examined of parental consent for parents and active include weight-based teasing (Eiesnberg et written assent for youth. The recruitment al., 2003; Janssen, Craig, Boyce, & Pickett, procedures were based on those used in our 2004) and sexual victimization (Gidycz & earlier study (Gould, Velting, Kleinman, Lu- Koss,1989). Toour knowledge,no studyhas cas, Thomas, & Chung, 2004) and were de- examined the association between specific velopedin responsetowhatthe schoolscon- types of peer victimization with suicidality. sidered would best meet the needs of their The purpose of the current research communities. The current study received was to examine the association between spe- IRB approval of a waiver of consent, based cific types of peer victimization with depres- on Federal Regulations (Title 45; Part 46, sion, suicidal ideation, and suicide attempts Article 46.116(d)). Two mailings with an in- amongalargesampleofAmericanhighschool formation sheet describing survey content students. Specifically, we examined the preva- and procedures, a response form, and a lenceofsixspecifictypesofpeervictimization; stamped response envelope were sent to par- the associationof thesetypes of victimizations ents 6and 4 weeks beforesurvey administra- with depression, suicidal ideation, and suicide tion, providing parents opportunities to re- attempts by gender; and the co-occurrence of fuse their children’s participation and giving multiple types of peervictimization. them ample pertinent information about the project.Students’writtenassentwasobtained immediately before the survey. SUBJECTSANDMETHODS The study procedures, consistent with the Family Educational Rights and Privacy Participants Act and the Protection of Pupil Rights This study targeted adolescents aged Amendment, were approved by the IRB of 13 through 19 years, enrolled in nineth theNewYorkStatePsychiatricInstitute/Co- through twelfth grade in six high schools in lumbia University Departmentof Psychiatry. Nassau, Suffolk,and Westchestercounties in NewYorkState.Fiveschoolswerepublicco- Measures educational schools and one was a parochial all-boys school. These schools were part of a A self-report questionnaire assessed de- study examining whether asking about sui- pression, suicidal ideation, suicide attempts, cidality during a screening program creates and peer victimization. The assessment time distress or increases suicidal ideation (Gould frame was the past 4 weeks, with the excep- et al., 2005). This study included 2,341 of tion of measuring lifetime suicide attempts. 3,635students(64.4%participationrate)from Demographic Questionnaire. The de- the fall of 2002 through the spring of 2004. mographicquestionnaireelicitedinformation Reasonsfornonparticipationincludedparen- with regard to age, grade, gender, racial/eth- talrefusals(61.9%),studentrefusals(14.3%), nic background, and household composition. and absences (23.7%). The ethnic distribu- Depression. The Beck Depression In- tion of the participating sample was 80.3% ventory(BDI-IA,versionIamended;Beck& White, 5.1% Black, 7.3% Hispanic, 3.8% Steer, 1993) assessed cognitive, behavioral, Asian, and 3.5% other. A total of 58.1% of affective,andsomaticcomponentsofdepres- the students were boys (the inclusion of an sion. Loss of libido was not assessed. The all-male parochial school explains the high BDI has been used in over 200 studies, in- percentage of boys). The mean (SD) age of cludingthosewithadolescentssamples(Rob- Klomek et al. 169 erts, Lewinsohn, & Seeley, 1991; Strober, attempts were derived from the depression Green, & Carlson, 1981; Teri, 1982). Each moduleoftheDiagnosticInterviewSchedule responserangedfrom0(symptomsnotpresent) for Children (DISC-IV; Shaffer, Fisher, Lu- to3(symptomissevere),withamaximumtotal cas, Dulcan, & Schwab-Stone, 2000) and an score of 60. A cut-off point of 16 was em- earlier suicide screen (Shaffer et al., 2004). ployed to dichotomize BDI scores. This cut- These items have demonstrated good con- off has correctly classified 81% of adolescent struct validity (Gould et al., 1998; Shaffer et psychiatric patients with major depressive al., 2004). The assessment of an attempt in- disorder (Strober et al., 1981) and has been cluded questions about occurrences, injuries recommended to detect possible depression sustained, medical care sought, and hospital- in normal populations (Beck & Steer, 1993). ization(Meehan,Lamb, Saltzman,&O’Car- The internal consistency reliability (as mea- roll,1992).Theadolescentwasconsideredto sured by Cronbach’s alpha) of the BDI-IA is have a history of an attempt if he or she re- .89 (Beck, Steer, Ball, & Ranieri, 1996). portedanypastattempt,regardlessoftiming, Suicidal Ideation. The Suicide Ide- injury, or medical attention. ation Questionnaire (SIQ-JR) assesses sui- Specific Types of Peer Victimization. cidalthoughtsandisdesignedforlarge-scale, Several questions regarding victimization school-basedscreeningofadolescents(Reyn- werederivedfromtheWHOstudyonyouth olds, 1988). The 15-item SIQ-JR uses a 7- health (Nansel et al., 2001). Questions about point Likert-type scale, ranging from 0 (I victimization were preceded with the follow- never had this thought) to 6 (This thought was ing explanation. “We say a student is being inmymindalmosteveryday),assessingthefre- bullied when another student or group of quency of specific suicidal thoughts during students says or does nasty and unpleasant the past month. It assesses thoughts related things to him or her. It is also bullying when todeathanddying,passiveandactivesuicidal apupilisteasedrepeatedlyinawayheorshe ideation, and suicidal intent. Reliability of doesn’t like. But it is not bullying when two the SIQ-JR is high, ranging from .91 to .96 students of about the same strength quarrel (Keane, Dick, Bechtold, & Manson, 1996; or fight.” Victimization was assessed by six Reynolds, 1988; Reynolds & Mazza, 1999) questions asking respondents to report the for internal consistency and from .87 to .93 frequencywithwhichtheywerevictimizedin for test-retest reliability (Reynolds & Mazza, differentways.Thedifferenttypesofvictim- 1999). The SIQ-JR has demonstrated crite- ization assessed were: “Made fun of you be- rion validity (King et al., 1993; Reynolds, causeofyourreligionorrace”;“Madefunof 1988, 1990; Reynolds & Mazza, 1999), con- you because of your looks or speech”; “Hit, structvalidityinclinicalsamples(King,Gha- slapped,orpunchedyou”;“Spreadrumorsor ziuddin, McGovern, Brand, Hill, & Naylor, mean lies about you”; “Made sexual jokes, 1996; King, Hill, Naylor, Evans, & Shain, comments,orgesturestoyou”;“Usede-mail 1993; King, Katz, Ghaziuddin, Brand, Hill, or Internet to be mean to you.” The fre- & McGovern, 1997; Sibthorpe, Drinkwater, quency items were coded on a 5-point scale Gardner, & Bammer, 1995; Siemen, War- ranging from not at all to most days. Respon- rington, & Mangano, 1994), and predictive dents were classified as never victimized, vic- validity(Keaneetal.,1996).Suicidalideation timized less than weekly, or victimized fre- was considered serious if the adolescent quently (at least 3 to 4 times in the past 4 scored31orhigherontheSIQ-JR;orscored weeks). 5 or 6 on two or more of the six “critical” SIQ-JR items (Reynolds, 1988); or re- Statistical Analysis sponded with either of the two most serious response options of the BDI suicide item. Chi-squareanalyseswereconductedto Suicide Attempt History. Seven ques- examinegenderdifferences intheprevalence tionsaskingaboutlifetimeandrecentsuicide of specific types of peer victimization. A se- 170 Adolescent Peer Victimization ries of logistic regression models was con- RESULTS ducted to examine the association between specifictypesofvictimization(lessthanweekly Frequency of Specific Types and frequently) with depression, serious sui- of Victimization cidal ideation, and suicide attempts (as di- chotomized outcomes). The category “never The most common experience of be- victimized” was the reference group in these ing bullied involved having one’s looks or analyses.These regressionanalyses werefirst speech belittled ([infrequently] less than conducted separately for males and females weekly, 20.4%; frequently, 9.1%) (Table 1). and were adjusted for schools and grade. In- Being bullied via e-mail or the Internet oc- teractions between gender and the specific curredwiththelowestfrequency(infrequently, typesofvictimizationwereexaminedinaddi- 5%; frequently, 2.3%). tional models. A power analysis revealed suf- Malesweremorelikelythanfemalesto ficient power (greater than .80) to detect po- bebelittledbecauseofreligionorraceandto tential interactions. be bullied by being hit, slapped, or punched Another series of logistic regression (χ2=41.15, p<.001; χ2=47.53, p<.001, re- analyses was conducted to examine the rela- spectively) (Table 1). Females were more tion of each of the three outcome measures likely than males to be the subject of rumors totheco-occurrenceofthedifferenttypesof (χ2=26.10, p<.001); sexual jokes, comments, victimization. A count of the number of the or gestures (χ2=28.31, p<.001); and mean- different types of victimization was created ness by use of e-mail or the Internet (χ2= for each respondent. The maximum count 18.99, p<.001). was 5 or 6 types of victimization. We com- Alltypesofvictimizationsweresignifi- bined having been victimized in 5 or 6 ways cantly correlated (Table 2) but they were not in order to avoid small ns. For this analysis redundant,giventhemodestsizesofthecor- wecombinedfrequentandinfrequentvictim- relations.Wethereforeexaminedtheassocia- izationwithineachtypeofvictimization.Re- tion of specific types of peer victimization spondents who were never victimized served with depression, serious suicidal ideation, and as the reference group in the logistic regres- suicide attempts by gender (see Table 3). sion. In addition, within the victimization Belittled About Religion or Race. This groups we conducted tests for linear trends type of victimization was significantly associ- (Fleiss, Levin, & Paik, 2003). atedwithdepressionandsuicideattemptsfor We chose not to include random ef- females, but was not significantly associated fects for school or class in the regression withserioussuicidalideation(SSI)amongfe- analyses because the sample clusters (school) males. Among males, being bullied frequently and the randomization unit (class within but not infrequently was associated with de- school) had little impact on the outcomes or pression and SSI. This type of victimization correlates(gender,depression,serioussuicide at anyfrequency was not associatedwith sui- ideation/behavior, peer victimization), as in- cide attempts among males. Interactions be- dicated by the intraclass coefficients (ICCs), tweenvictimizationandgenderwerenotsig- which were all close to zero (ICCs<.06). nificant with regard to depression, SSI, and Moreover, meaningful differences between attempts. results of random effects regression models BelittledAboutLooksorSpeech. Among andouranalysiswerenotanticipatedbecause females,beingbelittledaboutlooksorspeech thereweremanyunitsofrandomization(181 at any frequency was significantly associated classes) of relatively small average size (Mur- with depression, SSI, and attempts. Among ray, 1998). Analyses were conducted using malesfrequent,butnotinfrequent,victimiza- the SPSS software package, version 12. Re- tion was associated withdepression, SSI, and sults were considered significant at α<.05. attempts. Significant interactions between 0 20 0 0 0 s 0 90 0 0 0 e 0 00 0 0 0 χ2 nderrenc =p. =p.=p. =p. =p. =p. GeDiffe 1.15, 4.77,7.53, 6.10, 8.31, 8.99, 4 4 2 2 1 y uentl(n) 109) 197)116) 109) 151) (50) q% 0( 1(3( 0( 9( 3 re 5. 9.5. 5. 6. 2. F Total=aN2,181) Lessthanweekly%(n) 8.0(175) 20.4(443)11.5(250) 14.0(305) 11.0(239) 5.0(109) ( 7) 6)0) 4) 8) 1) ern) 89 5381 76 78 02 Nev%( 0(1, 6(1,2(1, 0(1 1(1, 7(2, 7. 0.3. 1. 2. 2. 8 78 8 8 9 y quentl%(n) 1(19) 5(68)8(25) 0(54) 2(83) 0(27) e 2. 7.2. 6. 9. 3. r F Female=aN907) essthanweekly%(n) 5.7(52) 0.2(183)7.6(69) 8.1(164) 3.9(126) 7.2(65) ( L 2 1 1 6) 5)3) 9) 7) 6) Never%(n) 2.2(83 2.3(659.6(81 6.0(68 6.9(69 9.9(81 9 78 7 7 8 er equently%(n) 7.1(90) 0.2(129)7.2(91) 4.3(55) 5.3(68) 1.8(23) nd Fr 1 tionbyGe Male=aN1,274) Lessthanweekly%(n) 9.7(123) 20.5(260)14.3(181) 11.1(141) 8.9(113) 3.5(44) ngdata. Victimiza ( Never%(n) 3(1,061) 4(881)6(997) 6(1,075) 8(1,091) 7(1,205) eofmissi r 3. 9.8. 4. 5. 4. us Pee r8r67 8 8 9 eca u u b TABLE1PrevalenceofSpecificTypesof MadefunofyoubecauseofyoreligionorraceMadefunofyoubecauseofyolooksorspeechHit,slapped,orpunchedSpreadrumorsormeanliesaboutyouMadesexualjokes,comments,orgesturestoyouUsede-mailorInternettobemeantoyou aNote.N’svaryslightly 172 Adolescent Peer Victimization TABLE2 Correlation Matrix of the Specific Types of Peer Victimization Victimized Victimized Victimized about about byhit, Victimized Victimized religion/ looks/ slappedor byspreading bysexual Victimized Correlations race talks punched rumors jokes bye-mail Victimizedabout religion/race .290** .159** .174** .149** .110** Victimizedabout looks/talks .322** .358** .353** .248** Victimizedbyhit, slapped,punched .302** .302** .225** Victimizedbyspread- ingrumors .391** .383** Victimizedbysexual jokes .249** Victimizedbye-mail Note.**<.01 gender and victimization by looks or speech Amongmales,victimizationatanyfrequency indicated that females who were belittled was associated with depression and SSI but about looks or speech were at significantly only frequent victimization was associated higherriskfordepressioncomparedtomales withattempts.Interactionsbetweenthistype (OR =3.71, 95% CI=2.41–5.72, p< of victimization and gender were not signifi- infrequently .001; OR =10.81, 95% CI=6.20–18.85, cant with regard to depression, SSI, and at- frequently p<.001). Interactions between victimization tempts. and gender were not significant with regard Subject of Sexual Jokes, Comments, or to SSI and attempts. Gestures. Among females, being bullied in- PhysicalVictimization(BeingHit,Slapped, frequently or frequently was significantly as- or Punched). Among females, being physi- sociated with depression and attempts. Only cally bullied at any frequency was signifi- frequent victimization was associated with cantly associated with depression, SSI, and SSI. Among males only frequent victimiza- attempts. Among males frequent, but not in- tion was associated with depression and both frequent, physical victimization was associ- levels of victimization were associated with ated with depression, SSI, and attempts. Sig- SSI and attempts. Interactions between this nificant interactions between gender and type of victimization and gender were not infrequent physical victimization indicated statistically significant with regard to depres- that females who were physically victimized sion, SSI, and attempts. infrequently were at significantly higher risk Cyber Victimization. Among females, for depression than comparable males (OR= being bullied via Internet or e-mail infre- .311, 95% CI=.13–.73, p<.001). Interac- quentlyorfrequentlywassignificantlyassoci- tionsbetweenphysicalvictimizationandgen- ated with depression and SSI. Only frequent der were not significant with regard to SSI victimization was associated with attempts. and attempts. Among males, frequent but not infrequent Subject of Rumors or Mean Lies. victimization was associated with depression Among females, victimization by the spread and both levels of victimization were associ- ofrumorsorliesatanyfrequencywasassoci- ated with SSI. The small number of males ated with depression, SSI, and attempts. who were victimized via Internet or e-mail Klomek et al. 173 and made an attempt (n =2; n =1) timized in five or six ways had any of these infrequent frequent precluded a meaningful examination of the three outcomes (depression, serious suicidal risk of attempts. Interactions between this ideation, suicide attempts). type of victimization and gender were not significant with regard to depression, SSI, DISCUSSION and attempts. The most common experience of be- Impact of Co-Occurrence of Several Types ing bullied involved having one’s looks or of Victimizations speech belittled, similar to Nansel’s findings (Nansel et al., 2001). This result is not sur- Among males who were victimized, prising in light of teens’ interest and invest- 44.9%experiencedonetypeofvictimization, mentintheirappearance.Beingbulliedviae- 26.6% experienced two types, 15.2% experi- mailortheInternetoccurredwiththelowest enced three types, 7.5% experienced four frequency. Cyber victimization may have oc- types, and 5.8% experiences five or six types curredinfrequentlyinthepresentsamplebe- of victimization. Among females who were cause the data was collected between 2002– victimized, 40.9% experienced one type of 2004. Cyber victimization may have become victimization, 28.4% experienced two types, more frequent since then (Patchin & Hin- 18.1% experienced three types, 8.1% experi- duja,2006; Ybarra,2004;Ybarra &Mitchell, enced four types, and 4.5% experienced five 2004). or six types of victimization. Malesweremorelikelythanfemalesto Overall, the more types of victimiza- be physically bullied and to be belittled be- tion the higher the risk for depression, SSI, cause of religion or race. Females were more and suicide attempts among both genders likelythanmalestobethesubjectofrumors, (Table 4). A test for linear trends within the sexual gestures, and meanness by use of the victimization groups indicated that as the Internet. The gender differences we found types of victimization increased, the rates of are consistent with previous findings that depressionandsuicidalityincreasedinanon- physicalvictimizationismoreprevalentamong linear fashion (male depression: χ2=22.94, males (Baldry & Farrington, 1999; Prinstein df=3, p<.001; female depression: χ2=25.16, et al., 2001). Adolescent males may be more df=3, p<.001; male ideation: χ2=30.46, involved in physical victimization due to df=3,p<.001;femaleideation:χ2=27.97,df= their generally higher levels of aggression 3, p<.001; male attempts: χ2=11.65, df=3, (Achenbach & Edelbrock, 1981). Our results p<.05; femaleattempts: χ2=27.93, df=3, p< arealsoconsistentwithNansel’sfindingsthat .001). When an individual is exposed to five formales,bothphysicalandverbalvictimiza- or six types of peer victimization the risk of tion was common while for females verbal depression and suicidality increases dramati- victimization and rumors were common cally.Maleswhowerevictimizedinfiveorsix (Nansel et al., 2001). The high levels of ver- ways were 12 times more likely to be de- balvictimizationamongbothgendersmaybe pressed, nearly 20 times more likely to have explainedbytheincreaseinverbalaggression reported serious suicidal ideation, and ap- with age (Conner, 2004). proximately 18 times more likely to have at- All types of victimization were associ- temptedsuicidecomparedtomaleswhowere ated with depression and suicidality. Our nevervictimized.Similarly,femaleswhowere findings support previous reports that both victimized in five or six ways were approxi- direct and indirect victimization have nega- mately 33 times more likely to be depressed, tive consequences for internalizing problems 27timesmorelikelytohaveexperiencedsui- (Crick&Grotpeter,1996;vanderWaletal., cidal ideation, and 19 times more likely to 2003). The pattern of associations we found attempt suicide. Approximately 40% of the was consistent across the different types of males and 70% of the females who were vic- peervictimization.Onaverage,themorefre- 5)5) 3)7) 7)0) 3)5) 8)4) 5)9) 4)6) 6)9) 19 53 69 36 39 11 08 30 7.8. 6.2. 6.0. 7.8. 3.5. 0.2. 7.2. 3.2. 1 2 1 11 11 1 12 I) –– –– –– –– –– –– –– –– C 10 21 33 22 24 49 08 36 R 03 09 72 31 76 89 04 52 O% 2.1. 1.. .2. 2.6. 2.3. 1.2. 2.2. 2.3. Gender equently (95 3.80***(3.41*( 2.58*(3.36( 2.20(6.83**( 4.12***(0.68***( 6.04***(7.62***( 4.32**(6.04***( 3.75***(5.65***( 5.82***(8.49***( y r 1 b F n o ati 5)7) 6)3) 4)5) 1)3) 2)4) 9)5) 5)1) 9)7) ctimiz %(n) 16.7(136.8( 6.7(15.8( 4.7(26.3( 16.3(248.5(3 9.4(120.9(1 7.1(22.1(1 16.5(144.0(1 10.1(28.0( Vi esof 37)13) 75)56) 06)47) 36)75) 51)82) 65)74) 71)49) 05)80) yp 3.4. 3.6. 5.0. 2.5. 4.4. 3.3. 2.7. 5.5. cificT OR%CI) .80–1.07– .42–.88– .67–1.98–1 6.9–2.42– .79–1.20– .60–1.04– .73–2.55– .86–1.03– gSpe thanekly (95 5(1*( 6(0( 5(5***( 8(3***( 8(0*( 8(7*( 0(7***( 9(5*( ncin LessWe 1.62.1 1.22.4 1.84.5 1.23.7 1.82.4 1.41.9 1.44.3 2.02.4 e ri e nsExp %(n) 8.1(10)5.0(13) 3.3(4)9.6(5) 4.1(5)7.6(9) 5.8(15)6.1(47) 3.1(8)7.7(14) 2.7(7)8.7(16) 6.6(12)8.8(26) 3.9(7)0.3(7) e 2 1 2 3 1 e T ng CI) mo OR% —— —— —— —— —— —— —— —— a 5 mpts ever (9 te N 1)8) 6)2) 3)7) 9)6) 5)2) 6)0) 9)1) 0)6) eAt %(n) 8(52(11 5(20(4 2(26(4 4(36(5 7(14(2 8(160(3 9(45(10 0(24(3 uicid 4.14. 2.5. 2.5. 4.8. 1.3. 1.4. 4.12. 2.4. S and race eech n, or sp Ideatio eligion ooksor cidal ourr ourl ui y y S of of n, e e d TABLE3PrevalenceofDepressio MadefunofyoubecausDepressionaMalebFemaleSuicideideationaMalebFemaleSuicideattemptsaMalebFemaleMadefunofyoubecausDepressionaMalebFemaleSuicideideationaMalebFemaleSuicideAttemptsaMalebFemaleHit,slapped,orpuncheDepressionaMalebFemaleSuicideideationaMalebFemale