551289

research-article2014

VAWXXX10.1177/1077801214551289Violence Against WomenCoker et al.

Article

Dating Violence Victimization and Perpetration Rates Among High School Students

Violence Against Women 2014, Vol. 20(10) 1220­–1238 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1077801214551289 vaw.sagepub.com

Ann L. Coker1, Emily R. Clear1, Lisandra S. Garcia1, Ibitola O. Asaolu1, Patricia G. Cook-Craig1, Candace J. Brancato1, Corrine M. Williams1, Heather M. Bush1, and Bonnie S. Fisher2

Abstract This school-based sample provides the largest estimate of physical and psychological dating violence (DV) victimization and the only report of DV perpetration among high school students. Among 14,190 students in relationships, 33.4% disclosed DV by a partner (victimization) and 20.2% used these same behaviors against a partner (perpetration) in the past 12 months. Physical DV victimization (13%) was less frequently disclosed than psychological DV (23%). Rates of DV victimization and perpetration were highest among females, those receiving free or reduced-price meals, those not exclusively attracted to the opposite sex, students reporting parental or guardian partner violence, binge drinking, and bullying. Keywords bullying, dating violence, perpetration, victimization, youth Dating violence (DV) is now well recognized as having a significant public health impact on student mental and physical health and academic achievement (Coker et al., 2000; Silverman, Raj, Mucci, & Hathaway, 2001). Teen DV is defined to include physical, sexual, or psychological/emotional violence within a dating relationship. More recently stalking, either in person or electronically, by a current or former dating

1University 2University

of Kentucky, Lexington, KY, USA of Cincinnati, OH, USA

Corresponding Author: Ann L. Coker, Department of Obstetrics and Gynecology, University of Kentucky, A. B. Chandler Medical Center, 800, Rose Street, C-361, Lexington, KY 40536-0293, USA. Email: [email protected]

Downloaded from vaw.sagepub.com at UCSF LIBRARY & CKM on March 10, 2015

1221

Coker et al.

partner has been included as teen DV (Centers for Disease Control and Prevention, 2012a). While the Centers for Disease Control and Prevention’s (CDC) Youth Risk Behavior Surveillance System (YRBS) provides a robust population-based estimate of physical DV victimization among high school students across the United States, there are few large, population-based estimates of physical or psychological DV victimization and perpetration presented by sex, tactics used, and frequency of use. A review of the 12-month DV incidence rates included in studies of at least 500 high school students revealed a range of 6.12% (Champion, Foley, Sigmon-Smith, Sutfin, & DuRant, 2008) to 15.8% (Marquart, Nannini, Edwards, Stanley, & Wayman, 2007). Physical DV incidence rates were similar among females and males (Centers for Disease Control and Prevention, 2012a; Coker et al., 2000; Temple & Freeman, 2011). Among the few studies that have asked students about their use of physically aggressive acts against a dating partner (perpetration), incidence rates ranged from 6.4% to 18.9% (Champion et al., 2008; Coker et al., 2000; Temple & Freeman, 2011) with rates slightly higher among females than males. As confirmed in the recent Partner Abuse State of Knowledge project’s (PASK) systematic review of partner violence in high school samples (Desmarais, Reeves, Nicholls, Telford, & Fiebert, 2012), physical DV victimization rates were similar among females and males (Jezl, Molidor, & Wright, 1996; O’Leary, Smith Slep, Avery-Leaf, & Cascardi, 2008), yet the consequences of violence may differ by sex. Among the few studies estimating both physical DV victimization and perpetration, the majority found that perpetration and victimization co-occurred in both males and females (Billingham & Sack, 1986; Malik, Sorenson, & Aneshensel, 1997; O’Leary et al., 2008). Those engaged in mutual violence were more likely to sustain and initiate more frequent violence (Foshee, 1996). Few large studies have estimated psychological DV among adolescents; Foshee and colleagues (Foshee, Benefield, et al., 2009) found that almost one third of adolescents disclosed being psychologically abused by a partner in the previous year. Further research is needed to more comprehensively describe the frequency of psychological and physical DV victimization and perpetration among males and females. Understanding how DV rates may be correlated with demographic factors as well as other types of violence and risk-taking behaviors such as parental or guardian partner violence, bullying, and binge drinking can help researchers and practitioners develop more effective intervention strategies. This large, cross-sectional study of 14,190 high school students who were in a dating relationship in the past 12 months adds to the existing literature by providing population-based estimates of DV victimization and perpetration by type of violence (physical or psychological) as well as mutual violence by tactic used. In contrast with findings from the population-based YRBS, which includes only 1 to 2 questions on DV victimization, this anonymous survey includes 10 questions on physical and psychological violence victimization and perpetration by a dating partner. We additionally have sufficient study power to present rates of DV by specific tactics (e.g., physical and psychological violence), directionality (e.g., victimization, perpetration, and both victimization and perpetration), and sex. Finally, we evaluate prevalence rates of DV victimization and perpetration by demographic factors and other forms of

Downloaded from vaw.sagepub.com at UCSF LIBRARY & CKM on March 10, 2015

1222

Violence Against Women 20(10)

interpersonal violence, including parental partner abuse, binge drinking, and both bullying others and being bullied.

Method Participants All high school students (freshmen-seniors) attending 1 of 26 participating high schools were asked to complete a self-administered survey; 23,605 were present on the days of surveying across all 26 schools (schools were selected based on their geographic distribution across the state by region; 2 schools for each of 13 geographic regions). The survey was a 95-item paper-and-pencil anonymous questionnaire that took approximately 25 to 40 min to complete. Study personnel traveled to each high school and administered surveys in one of two ways during school hours: (a) a class administration setting in the school (n = 16) during a selected class (e.g., English classes for the whole day,) or (b) a school (n = 10) group administration during which, in one class period, all students were surveyed. Students were instructed not to include their name or birthdate on the Scantron® forms. Elements of the assent form were read to all students by study staff, who had received Collaborative Institutional Training Initiative training. Students were directed to review the assent form in the survey booklet and decide for themselves whether they wished to complete the survey. The survey was conducted using the model familiar to schools in the Youth Risk Behavior Survey (Catalano, Smith, & Rand, 2009). All surveys in this study were administered between January and May 2010. At the end of the survey booklet, resources were included, such as websites and toll-free numbers for national agencies available at all times to address domestic violence, sexual assault, depression, or suicide ideation should students find this information helpful. These resources were also provided on pencils the students used to complete the survey; they were instructed to keep these pencils. Study protocol was approved by the University of Kentucky Institutional Review Board 09-0680-F1V.

Consent All parents and guardians of enrolled students were mailed an informational letter describing the study purpose and information about the study at least 2 weeks prior to surveying. Parents and guardians who did not want their student to complete the survey were asked to call or email study staff with both the name and school of their child. Study staff worked with teachers to identify these students in individual classrooms during surveying at that school. These students received a booklet with the same cover as their peers, but the questionnaire was not included. Study staff provided the same instructions and assent information to all students; all students were asked to read the assent form, and students who wished to complete the survey were instructed to begin and bubble in their answers to the 95 questions on the Scantron® form. Students who did not wish to participate were asked to bubble in the letter H on the Scantron® form

Downloaded from vaw.sagepub.com at UCSF LIBRARY & CKM on March 10, 2015

1223

Coker et al.

for all 95 questions, and those with a blank booklet (indicating parental refusal) were asked to bubble in the letter J for the 95 questions. Having teachers remain in the classroom helped ensure smooth data collection.

Student Response Rates On the days we surveyed, 23,605 students were present in the 26 high schools; 20,806 marked a Scantron® form (88.1% of students attending the 26 high schools). Of these 20,806 students who marked a form, the overall refusal rate was 8.5%; 1,454 (7.0%) were student refusals, 82 (0.4%) were parental refusals, and 232 (1.1%) students started the survey but completed less than 30 of the 95 items (partial refusals). Of the remaining 19,038 students, 369 (1.9%) did not answer one of the demographic factors; 1,918 (10.1%) did not complete all 10 items measuring victimization and perpetration of DV or indicated that they were not in a relationship; and another 2,561 (13.5%) students answered the current relationship question to indicate that they were not in a dating relationship during the past 12 months and thus not at risk for DV measured during this time frame, resulting in 14,190 students for the current analysis. There were no significant demographic differences between those who agreed and completed the survey and those who did not, with one exception: Males were twice as likely (p < .0001) as females to refuse to complete the survey. As reported elsewhere (Clear et al., 2014), our statewide sample was representative of Kentucky when compared with the Kentucky Department of Education (KDE) 2009-2010 school year high school population with the exception that female high school students were over represented (53.5%), and seniors (16%) were under represented in our sample.

Measures Five items were used to measure DV perpetration and victimization (see Table 2). One item measured physical DV (hit, slap, or physically hurt you on purpose), and four items measured psychological abuse in the last 12 months. The same five items were rephrased to determine the student’s use of DV: “In the last 12 months, how many times have YOU” done the following. One additional item not included in the DV measure addressed intentional physical injuries resulting from unwanted sex or DV. The same response options for all items ranged from never, 1-2 times, 3-5 times, and 6 or more times; variables were coded numerically using the minimum of the interval response. We included those who answered that they had experienced the event in the past but not in the past 12 months as never, because we were estimating the rate of DV in the past 12 months. The Cronbach’s alphas were .82 for the DV victimization measure and .80 for perpetration. For all psychological DV items, indicator variables were created to represent the proportion reporting violence occurring 3 or more times by tactic: (a) controlling; (b) damaging property; (c) shout, yell, or insult; and (d) threaten to physically hurt. We opted for this more restrictive definition based on the frequency of these tactics

Downloaded from vaw.sagepub.com at UCSF LIBRARY & CKM on March 10, 2015

1224

Downloaded from vaw.sagepub.com at UCSF LIBRARY & CKM on March 10, 2015

All students Sex  Female  Male Grade  9th  10th  11th  12th Race  Nonwhite  White Proxy for income “Free or reduced lunch?”a  Yes  No Sexual attractionb   Attracted to opposite sex only   Attracted to same sex or both sexes Relationship status, past 12 monthsc   In one relationship now   In multiple relationships now   Not in a relationship now, but have been in the past 12 months Partner violence between parents/guardiansd  Yes  No Bullied at schoole   More than twice   1-2 times

Demographic factors

1,620 (11.4%) 2,501 (17.7%)

3,847 (27.1%) 10,343 (72.9%)

7,525 (53.0%) 1,243 (8.8%) 5,422 (38.2%)

12,196 (86.0%) 1,994 (14.0%)

6,217 (43.8%) 7,973 (56.2%)

2,331 (16.4%) 11,859 (83.6%)

4,109 (28.9%) 3,901 (27.5%) 3,781 (26.7%) 2,399 (16.9%)

7,897 (55.6%) 6,293 (44.4%)

N = 14,190

N (%)

All students

χ12 = 22.2, p < .0001 52.7% (2,028) 32.3% (3,336) χ 22 = 626.6, p < .0001 61.2% (991) 47.1% (1,177)

40.6% (2,521) 35.7% (2,843) χ12 = 21.1, p < .0001 35.7% (4,357) 50.4% (1005) χ 22 = 21.5, p < .0001 38.8% (2,921) 55.2% (686) 32.4% (1,757)

χ 37.8% (5,364) = 19.6, p< .0001 42.5% (3,356) 31.9% (2,008) χ12 = −9.7, p = .02 34.5% (1,419) 38.8% (1,512) 38.7% (12462) 40.5% (1,428) χ12 = 8.0, p = .005 40.8% (951) 37.2% (4,413) 2 1

Rate (n in strata)

DV victim or perpetrator

χ12 = 478.9, p < .0001 47.6% (1,831) 28.1% (2,907) 2 χ2 = 593.1, p < .0001 55.9% (906) 41.6% (1,040)

χ 22

χ12

χ12

χ12

χ12

χ12

33.4% (4,738) = 114.2, p < .0001 37.2% (2,935) 28.7% (1,803) = 25.3, p < .0001 30.4% (1,247) 34.1% (1,330) 34.7% (1,311) 35.4% (850) = 2.0, p = .15 34.7% (808) 33.1% (3,930) = 22.1, p < .0001 35.5% (2,207) 31.7% (2,531) = 135.4, p < .0001 31.5% (3,845) 44.8% (893) = 218.3, p < .0001 34.2% (2,573) 50.1% (623) 28.4% (1,542)

Rate (n in strata)

DV victim

(continued)

χ12 = 348.7, p < .0001 30.5% (1,174) 16.4% (1,692) 2 χ2 = 320.5, p < .0001 34.2% (554) 25.5% (637)

20.2% (2,866) χ12 = 184.8, p < .0001 24.3% (1,918) 15.1% (948) χ12 30.9, p < .0001 17.4% (713) 20.2% (789) 21.9% (829) 22.3% (535) χ12 = 43.7, p < .0001 25.2% (588) 19.2% (2,278) χ12 = 20.8, p < .0001 21.9% (1,364) 18.8% (1,502) χ12 = 109.9, p < .0001 18.8% (2,289) 28.9% (577) χ 22 = 133.6, p < .0001 21.6% (1,626) 29.7% (369) 16.1% (871)

Rate (n in strata)

DV perpetrator

Table 1.  Demographic Profile of High School Students Completing Baseline Survey, by Victimization or Perpetration of Physical or Psychological Dating Violence Among Those in a Dating Relationship in the Prior 12 Months.

1225

Downloaded from vaw.sagepub.com at UCSF LIBRARY & CKM on March 10, 2015

4,288 (30.5%) 9,783(69.5%) 119

χ12 = 360.2, p < .0001 44.8% (1,919) 28.4% (2,775) 119

χ12 = 21.0, p < .0001 50.3% (2,156) 32.3% (3,157)

27.7% (2,781) χ = 638.0, p < .0001 57.7% (716) 44.6% (1,184) 27.5% (2,831) 2 2

Rate (n in strata)

DV victim

χ 22 = 21.16, p < .0001 31.2% (3,205) 50.6% (1,344) 65.0% (807)

31.70 (3,185)

10,048 (70.9%) 21 1,241 (8.7%) 2,654 (18.8%) 10,279 (72.5%) 16

Rate (n in strata)

DV victim or perpetrator

N (%)

All students

16.6% (1,669)   χ 22 = 771.3, p < .0001 44.1% (547) 29.7% (788) 14.8% (1,525)   χ12 = 311.3, p < .0001 29.2% (1,250) 16.2% (1,584) 119

Rate (n in strata)

DV perpetrator

Note. DV = dating violence. aReceived free or reduced cost breakfast or lunch: “Do you receive a free or reduced-price meal through your high school?”— with yes/no answer options. bSexual attraction: “People are different in their sexual attraction to other people. Which best describes your feelings?”—only attracted to females, mostly attracted to females, equally attracted to females and males, mostly attracted to males, only attracted to males, and not sure. A dichotomous variable was created to determine exclusively heterosexually oriented students based on students’ responses as well as their sex. Those not reporting that they were attracted only to the opposite sex were grouped as not exclusively heterosexual. cCurrent dating relationship status: “During the past 12 months, have you been in a relationship with a boyfriend or girlfriend?” By a relationship, we mean either having a partner for planned event like a school dance or going to the movies, having a sex partner, or hanging out in a group with a boyfriend or girlfriend. Response options were (a) no, never been in a relationship (excluded); (b) no, not in the last 12 months (excluded); (c) not now, but have been in the last 12 months; (d) yes, in one relationship now; and (e) yes, in multiple relationships. dIntraparental Partner Violence “In your family, how often did you see or hear one of our parents or guardians being hit, slapped, punched, shoved, kicked or otherwise physically hurt by their spouse or partner?” Response options ranged from never to more than 10 times but were dichotomized as yes or no. eBeing bullied: “In the past 12 months, how many times have you been bullied by another high school student?” and bullying others “In the past 12 months, how many times have you bullied another high school student?” Response options again ranged from never to more than 10 times for both questions. fMissing numbers not included in the percentages. gBinge drinking: “In the past month, on how many days did you have 4 (if you are a female) or 5 (if you are a male) or more drinks of alcohol in a row (within a couple of hours)?” Response options ranged from I never drink to 20-31 days; a dichotomous variable was created to identify those students who reported binge drinking one or more days in the past 30.

 Never  Missingf Bullied others at school e   More than twice   1-2 times  Never  Missingf Binge drinking, more than once in past monthg  Yes  No  Missingf

Demographic factors

Table 1.  (continued)

1226

Violence Against Women 20(10)

because we were interested in identifying a patterned or chronic abuse characteristic of partner abuse. Because we cannot know the intentions of the boyfriend/girlfriend or the student in their report of using tactics labeled DV, we used the labels of experiencing and using DV to replace the language of victimization and perpetration. Finally, the overlap of victimization and perpetration was categorized as (a) both DV victimization and perpetration, (b) DV perpetration alone, (c) DV victimization alone, and (d) no DV (reference group). Several demographic variables (Table 1) were included in the survey. The specific DV items and response options are provided in Table 2. Our inclusion of parental or guardian partner violence, bullying, and binge drinking as correlates of DV victimization and perpetration greatly expanded the scope of existing literature and helped characterize co-occurrence of DV victimization and perpetration.

Statistical Analysis The analyzable data set included only those students indicating being in a dating relationship during the past 12 months. For these students, rates of DV victimization, and DV perpetration are provided overall and by demographic factors in Table 1. Rates of specific DV items are provided overall and by sex (Table 2); co-occurring DV victimization and perpetration rates are also provided by sex (Table 3). Comparisons between groups were made using generalized estimating equations (GEE; link = log and distribution = binomial, exchangeable working correlation matrix). GEEs allowed for the investigation of group differences, while accounting for the potential correlation that existed for students attending the same school; the unadjusted rates were presented with Wald chi-square statistics (obtained using GEE) for the overall effect on DV victimization and perpetration of each categorical variable (Table 2). Moreover, GEEs allowed for comparisons while controlling for grade level, receiving free meals, relationship status, race, sexual attraction, and parental partner violence. Adjusted prevalence rate ratios are presented with 95% confidence intervals in Table 2. Given that the outcome involving the co-occurrence of DV victimization/perpetration is a multiplelevel categorical variable (victim and perpetrator, perpetrator only, victim only, and no DV), multicategory logit models were used to make comparisons (PROC LOGISTIC link = glogit); adjusted odds ratios are presented with 95% confidence intervals (Table 3). Data were analyzed using Statistical Analysis System, SAS, Version 9.3. Given the large sample size and multiple tests, only p values

Dating violence victimization and perpetration rates among high school students.

This school-based sample provides the largest estimate of physical and psychological dating violence (DV) victimization and the only report of DV perp...
396KB Sizes 0 Downloads 8 Views