Children and Youth Services Review 29 (2007) 557 – 571 www.elsevier.com/locate/childyouth

Dating violence among emancipating foster youth Melissa Jonson-Reid a,⁎, Lionel D. Scott Jr. b , J. Curtis McMillen a , Tonya Edmond a a

Washington University in St Louis, United States b Georgia State University, United States

Received 14 September 2006; received in revised form 19 November 2006; accepted 7 December 2006 Available online 8 February 2007

Abstract Youth emancipating from foster care have been found to be at higher risk for many adverse outcomes. Only one prior published study exists, however, about the dating violence experiences of foster youth. Asking similar questions about dating violence, this study attempted to replicate the prior findings regarding prevalence and to explore associations between foster care placement history, maltreatment history and mental health/substance abuse and self-reported dating violence. The youth reported a higher prevalence of dating violence than found in studies of the general youth population. Self-reported PTSD symptoms and drug use were associated with higher likelihood of dating violence victimization. The majority of youth self-reporting as dating violence perpetrators also self-reported as victims. Implications for research and practice are discussed. © 2007 Elsevier Ltd. All rights reserved. Keywords: Foster care; Dating violence; Foster youth

1. Introduction As a recent review indicates, empirical research on dating violence among high school and middle school students is scant (Hickman, Jaycox, & Aranoff, 2004). In addition, only two prior studies are known that have examined dating violence in a child welfare population (Jonson-Reid & Bivens, 1999; Wekerle et al., 2001). Among children served by child welfare agencies, youth in foster care, particularly those “aging out” or emancipating from care, are considered at

⁎ Corresponding author. George Warren Brown, School of Social Work, Washington University in St Louis, St Louis, MO 63130. E-mail address: [email protected] (M. Jonson-Reid). 0190-7409/$ - see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.childyouth.2006.12.008

558

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

particularly high risk for a number of negative outcomes and behaviors (Jonson-Reid & Barth, 2000; McDonald, Allen,Westerfelt, & Piliavin, 1993; McMillen, Auslander, Elze, White, & Thompson, 2003; McMillen et al., 2005). Yet, only one prior study of dating violence among youth in foster care exists (Jonson-Reid & Bivens, 1999). Emancipating foster youth are an ideal population for dating violence prevention and intervention programs due to the existence of independent living preparation programs that can be a vehicle for such efforts nationwide. The lack of information regarding dating violence among foster youth, however, hampers the ability to advocate for and inform services in this area. This study seeks to help fill the gap in information on dating violence among foster youth by partially replicating and expanding a study of foster youth and dating violence in California (Jonson-Reid & Bivens, 1999). 1.1. Background 1.1.1. Prevalence It is possible to be a victim of dating violence, a perpetrator, or both. Past estimates of the occurrence of dating violence among middle and high school students vary widely — estimates range from 0.2% to 96% (Bergman, 1992; Gray & Foshee, 1997; Hickman et al., 2004; Howard & Wang, 2003a,b; Jezl, Molidor & Wright, 1996). The few studies that focus on dating violence among adolescents are difficult to compare due to varying samples and definitions of violence. Data from the National Crime Victimization Survey suggest a rate of 1.7% for physical or sexual abuse among girls ages 16–19 and only 0.2% for boys (Hickman et al., 2004). Bergman (1992) surveyed 641 high school students and found a prevalence rate of about 25% for physical or sexual victimization among girls and about 10% among boys. Gray and Foshee (1997) surveyed 185 middle and high school students and found a prevalence rate of physical violence of about 27% for acts that ranged from spitting to assault. A Canadian study of dating violence (Wekerle et al., 2001) compared high school youth reported for maltreatment (n = 224) with a non-reported sample (n = 978). Students reported victimization by type of dating violence, including threats, hitting, pushing, slapping/pulling hair and “Partner said things just to make me angry.” A percentage endorsement of between 4% and 24% was found for physical dating and threats and between 37% and 56% for the item about verbally inciting anger (Wekerle et al., 2001). The only known study of dating violence among middle and high school youth in foster care (n = 85) found a prevalence rate of 48% including physical, sexual and verbal abuse and threats of physical harm (Jonson-Reid & Bivens, 1999). 1.1.2. Gender Among adults, it is generally agreed that women are the primary victims of intimate partner violence (Tjaden & Thoennes, 2000). A conclusive statement about the relationship between gender and dating violence in adolescence is made more difficult due to the differing means of measuring violence used, fewer numbers and levels of rigor. Although the National Crime Victimization Rate Survey (Rennison & Welchans, 2000) found that girls/women had a much higher victimization rate, other studies have found that girls were only somewhat more likely to report being a victim or a perpetrator of dating violence than males (Bergman, 1992; Gray & Foshee, 1997). In most studies asking about dating sexual abuse victimization in general population samples, females reported higher rates (Hickman et al., 2004). In the prior study of foster youth, females reported a higher prevalence of victimization and a lower prevalence of perpetration (Jonson-Reid & Bivens, 1999).

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

559

1.1.3. Ethnicity Research on the association of ethnic/racial groups and the occurrence of dating violence is also inconclusive (Gray & Foshee, 1997; Jezl et al., 1996; O'Keefe, 1997). Rennison and Welchans (2000) reported higher rates of intimate partner violence among black men and women than for other racial groups. O'Keefe (1997) found racial differences in the likelihood of inflicting violence among males, but not among females once sociodemographic and contextual variables were controlled. Neither the study of foster youth nor the Canadian study which included a maltreatment sample examined racial or ethnic differences (Jonson-Reid & Bivens, 1999; Wekerle et al., 2001). 1.2. Risk and contextual factors Studies of dating violence among adolescents are just beginning to explore risk factors that may contribute to or co-occur with dating violence. The most commonly investigated risk factor is exposure to family violence (child maltreatment or domestic violence), but a few have also included potential risk factors like mental health disorders or substance use (Howard & Wang, 2003a,b; Wekerle et al., 2001). As aforementioned only one prior study of foster youth and dating violence is known to exist and this study did examine mental health, substance abuse or child welfare service history correlates. 1.2.1. Prior exposure to family violence Some theories suggest that the experience of maltreatment or exposure to family violence increases the risk of later violent outcomes either due to learned roles (Crittenden & Claussen, 2002; Maxfield & Widom, 1996), traumatic reactions (Wekerle & Wolfe, 1998) or attachment issues (Cicchetti, 1991). The empirical relationship between past exposure to violence and dating violence is, however, unclear (Wekerle et al., 2001). Measurement of prior family violence varies from self-report of past physical and sexual abuse (Gray & Foshee, 1997; Smith & Williams, 1992), to official report of maltreatment (Wekerle et al., 2001), to endorsing witnessing family violence without differentiating between domestic violence and child maltreatment (Jonson-Reid & Bivens, 1999). Some studies report a higher incidence in maltreated samples or subgroups having witnessed family violence (Smith & Williams, 1992; Wekerle & Wolfe, 1998; Wekerle et al., 2001), while others do not (Jonson-Reid & Bivens, 1999; O'Keefe & Treister, 1998). This may be due, in part, to the age at onset and whether or not the maltreatment recurs. 1.2.2. Child welfare factors Thus far, no studies have examined child welfare case characteristics like age at entry, placement types, or multiple spells in foster care in relation to dating violence. These characteristics, however, are frequently associated with other outcomes in child welfare populations (e.g., Courtney & Barth, 1996; Jonson-Reid & Barth, 2000; Jonson-Reid, 2002; Widom, 1991). 1.2.3. Mental health and substance abuse Dating violence may also be associated with mental health disorder or substance abuse among youth. Wekerle et al. (2001) found that PTSD symptoms were associated with victimization and perpetration among females with maltreatment report histories. PTSD symptoms were only associated with perpetration of dating violence among males with prior maltreatment reports. Howard and Wang (2003b) reported higher levels of substance use, suicidal behavior, risky sexual behavior and indicators of depression among girls who were victims of dating violence. Among boys, however, substance use was not a risk factor (Howard & Wang, 2003a).

560

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

1.2.4. Prevention and intervention As aforementioned, nationwide programming does exist to assist older youth in emancipation from foster care which provides a ready venue for prevention and intervention materials. At least one study suggests that prevention is effective with children with histories of maltreatment (Wolfe et al., 2003). Unfortunately most programs designed for teens have not undergone evaluation (Hickman et al., 2004). Little information is available to guide prevention efforts for foster youth. The only prior study of dating violence among foster youth (Jonson-Reid & Bivens, 1999) found that many of the youth reporting violence told at least one other person—usually a friend. Most expressed a desire for more information about teen dating violence. 1.3. The current study This study advances knowledge by attempting to replicate and expand upon the only other known study of dating violence among foster youth. The present study uses a larger sample drawing from all older youth in care rather than limiting participation to those attending an ILP program (see prior study Jonson-Reid & Bivens, 1999). This study examines differences in dating violence by self-reported maltreatment, foster care service characteristics, PTSD, and substance use. Theory (learning theory, attachment, and response to trauma), however, will be discussed in relationship to study hypotheses and findings. (1) What is the prevalence of dating violence among adolescents emancipating from foster care? Hypothesis 1. The prevalence of victimization among females will be higher than males. (2) Is self-reported maltreatment history associated with dating violence in foster youth? Hypothesis 1. Maltreatment history will be associated with higher rates of dating violence. (3) Are characteristics of foster care (placement type, stability, repeated spells, age at entry) associated with dating violence? Hypothesis 1. Multiple placements and/ or spells will be associated with dating violence. (4) Does the association between maltreatment and dating violence remain when controlling for reported PTSD and substance use? Hypothesis 1. The association between maltreatment and dating violence will decrease in magnitude once controlling for PTSD and substance use. (5) What actions do youth take (help-seeking) who are involved in dating violence? 2. Methods 2.1. Participants Data was taken from the baseline and three month follow-up interviews of a two-year longitudinal study of older youth in the care and custody of the Missouri Children's Division (MCD). From December 2001 to May 2003, the Missouri Children's Division provided to the research team the

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

561

names of youth who were in its custody and who would be turning 17 years old in the following month. The names and contact information of youths' respective caseworkers were also provided to the research team. Youth were from eight counties, which included six counties in and around St. Louis and two counties in Southwest Missouri that were added to make the sample more ethnically representative of youth in the state's foster care system. Caseworkers were then contacted to provide informed consent. Youth were screened for exclusion criteria, and were excluded from participation in the study if they met the following criteria: (a) possessed a documented full-scale IQ score below 70, (b) had exited MCD custody, (c) possessed a chronic medical condition that made it difficult to communicate, (d) were placed more than 100 miles beyond the borders of any of the eight counties, and (e) were on continual run-away status through 45 days past their 17th birthday. After case managers provided consent, prospective youth were contacted and asked if they wanted to participate. Of the 451 eligible youth, 406 (90%) completed the baseline interview. Of those eligible, 39 (8%) chose not to participate, 4 (1%) could not be contacted, and one interview for which consent and assent were obtained could not be completed. The baseline interview was completed in person at youth's place of residence by trained full-time professional interviewers, and youth were paid $40. The three-month follow-up interview was completed by telephone by trained full-time professional interviewers, and youth were paid $20. PTSD, maltreatment history, and substance use were assessed in the baseline interview. Dating violence history was assessed in the three-month follow-up interview. The Hilltop Human Subjects Committee of Washington University at St Louis approved all procedures. 2.2. Sample At baseline, the longitudinal study consisted of 406 youth, 87% (n = 354) of whom completed the three-month follow-up interview. Results of attrition analysis showed that the 52 non-completers of the three month follow-up interview did not significantly differ from completers on major study variables (gender, ethnic group membership, foster care placement setting, psychiatric history, maltreatment history, etc.). For the purposes of this analyses, youth who reported never having had a girlfriend or boyfriend or going on a date were dropped from the sample (n = 15). The final sample for the present study was 339 youth. The participants were 17 years old. The mean age at which youth entered the foster care system was 10.9 (4.46) years, with the majority reporting placement in the care and custody of state authorities on only one occasion (77.6%). Seventy-four percent of youth reported some type of maltreatment — physical abuse (45%), physical neglect (45%), sexual abuse (34%), and/or emotional abuse (46%). Fifty-two participants (15%) qualified as meeting criteria for Post-Traumatic Stress Disorder (PTSD) in the past-year or lifetime. Any substance use (alcohol or drugs) in the past six months was reported by 61.7% of the sample; drug use was reported by 49.5% of the sample and poly-substance use (alcohol, marijuana, and another illegal substance) was reported by forty-four participants (12%). 2.3. Measures Age, gender and race were provided by the youth. 2.4. Foster care history Age at first entry into care (in years), number of placements into foster care (1 or more spells), remaining in the same home for the past 12 months (yes/no), and current and lifetime placement

562

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

type were assessed through youth self-report. Current placement type categories were: (1) relative/friend/school; (2) foster home; (3) group home/mental health or substance abuse residential center/institution; (4) other. “Ever placed” categories were collapsed into (1) Ever on the street, (2) Ever in a detention center/mental health/substance abuse residential center/ institution (but not on the street); (3) Other. Length of time in care for most recent spell was created by subtracting the youth's current age from the year first placed or the year most recently placed into foster care. 2.5. Maltreatment history Physical abuse, physical neglect, and emotional abuse histories were assessed with the Childhood Trauma Questionnaire (CTQ; Bernstein & Fink, 1998). The CTQ uses five items to assess each type of maltreatment. Youth indicated the extent to which they had been victims of physical abuse (e.g., “I was punished with a belt, a board, a cord, or some other hard object.”), physical neglect (e.g., “I didn't have enough to eat.”), and emotional abuse (e.g., “People in my family said hurtful or insulting things to me.”) on a 5-point, Likert scale ranging from never true (1) to very often true (5). The CTQ has demonstrated adequate reliability in a community sample (Scher, Stein, Asmundson, McCreary, & Forde, 2001) and previous study of older youth in foster care (Auslander et al., 2002). In this study, alpha coefficients were .88 for physical abuse, .78 for physical neglect, and .87 for emotional abuse. For clarity of results, Bernstein and Fink's (1998) cut-off score of 10 or above was utilized to identify cases of moderate or severe abuse and neglect. To assess sexual abuse history, three items adapted from Russell (1986) and included in a prior study of older foster care youth (Auslander et al., 2002) were used. Youth were asked to indicate (a) if they were ever made to touch someone's private parts against their wishes, (b) if anyone had ever touched their private parts (breasts or genitals) against their wishes, and (c) if anyone ever had vaginal, oral, or anal sex with them against their wishes? Youth responses were dichotomized into: history of abuse (1) vs. no abuse history (0) overall and by type. 2.6. Post-traumatic stress disorder Lifetime post-traumatic stress disorder (PTSD) was assessed by the Diagnostic Interview Schedule (DIS) for DSM-IV (Robins, Cottler, Bucholz, & Compton, 1995). The DIS-IV is a structured diagnostic interview schedule designed for use with lay interviewers that assesses the recency, onset, and duration of DSM-IV diagnoses. The DIS-IV has demonstrated moderate reliability and validity for lifetime and current diagnoses (Robins, Helzer, Croughan, & Ratcliff, 1981). 2.7. Substance use Alcohol use in the past six months was assessed with questions from the Diagnostic Interview Schedule for Children and Adolescents (DICA-IV) (Reich, Welner, & Herjanic, 2002). Youth were asked the number of times they drank alcohol in the past six months. Response options ranged from 1 or 2 occasions (1) to 40 or more occasions (6). Use of other substances was assessed with modified portions of the Diagnostic Interview Schedule for DSM-IV (Robins et al., 1995). Respondents indicated whether they had ever used drugs from nine different categories (marijuana, amphetamines, sedatives, opiates, cocaine, hallucinogens, inhalants—excluding

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

563

nitrous oxide or amyl nitrate, and club drugs). The last two substances were added for this study and were not part of the original DIS-IV. Almost all of the youth who drank on several occasions were also part of the “any use of drugs” group. “Any drug use” was included in the present analyses. 2.8. Dating Youth were asked a series of questions about their dating habits including if they had ever been on a date, the number of dating partners they had, the number of boyfriends or girlfriends they had, and frequency of dating. If youth replied that they had never had any dates or the number of dating partners was 0 or the number of dating partners was missing and the number of boyfriends and girlfriends was 0 then they were excluded from the sample. 2.9. Dating violence Dating violence was assessed by using the modified version of Bergman's (1992) Questionnaire on Dating Violence that was used by Jonson-Reid and Bivens (1999). Four types of dating violence were assessed: sexual violence (being forced to do something sexual against your wishes — e.g., kissing, touching); physical violence (being hurt physically by a date — e.g., shaken, slapped, punched, kicked, or pushed); threatening violence (being threatened with harm by a date); threatening self-harm (date threatened to hurt self to control behavior); and verbal violence (being repeatedly put down and called names by a date). Questions about being victimized were supplemented by similarly worded questions about inflicting violence for each type of abuse. Youth were also asked if they ever continued to remain in a relationship after experiencing some type of dating violence. 2.10. Help-seeking Youth who reported any involvement in dating violence were asked a series of follow-up questions. Youth were asked whether they continued to remain in a violent dating relationship (yes or no). If they answered yes, they were asked if the violence continued to occur. They were then asked if they told anyone about their dating violence involvement. Youth who responded “yes” were asked to indicate if they told any of the following persons: teacher, friend, counselor/ therapist, child welfare worker, family member, foster parent, religious leader, and/or other unspecified person. Those who reported having told someone were asked if the person was able to help, and if so, how. 2.11. Data analysis Data analyses were conducted using SAS 9.1 for Windows. Chi-square, Cochran Mantel– Haenszel chi-square, Fisher exact tests and independent-samples t tests, were used to examine differences in dating violence victimization, perpetration status and help-seeking based on sample characteristics. Because youth could endorse more than one form of violence and the group reporting violence was small, dating violence victimization was examined by subtype in bivariate analysis only with logistic regression used only for overall (yes/no) dating violence victimization. Independent variables were entered in a stepwise fashion to examine the relative influence of the factors. Demographic characteristics were entered first, followed

564

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

by maltreatment history in Step 2, foster care characteristics in Step 3, substance abuse and PTSD in Step 4, and number of lifetime dating partners in Step 5. Due to the small numbers endorsing perpetration, only descriptive analyses of dating violence perpetration are presented. 3. Results 3.1. Experience of dating violence Of the 339 older youth in foster care, 18.3% (n = 62) reported involvement (victim and/or perpetrator) in some type of dating violence. About 17% (n = 57) reported being a victim-coerced sexual act or physical harm (6.2%, n = 21); threatened with harm (5%, n = 17); date threatened self-harm (6.5%, n = 22); verbal abuse (10.3%, n = 32). Eight percent (n = 27) reported being a perpetrator of dating violence-coerced sexual act or physical harm (1.7%, n = 6); threatened date with harm (0.9%, n = 3); threatened self-harm (0.9%, n = 3); verbal abuse (5.9%, n = 20). Six percent (n = 22) of youth reported being both a victim and perpetrator. Results of one-sample chi-square tests of any reported type of victimization or perpetration are provided in Table 1. Youth who were Caucasian, female, victims of sexual abuse, youth who met criteria for PTSD or depression, and youth who reported drug or poly-substance use were more likely to report being victims of dating violence. Only alcohol or drug only use were associated with being a perpetrator of dating violence. Age at first entry into foster care was also examined (not shown in Table 1). Victims and non-victims of dating violence did not significantly differ in the age at which they entered foster care, t(337) = − 1.02, p = .31 (M = 10.24, SD = 4.53 for non-victims; M = 10.35, SD = 4.08 for victims). Similarly, perpetrators and non-perpetrators of dating violence did not significantly differ in the age at which they entered foster care, t(337) = − 1.31, p = .19 (M = 10.29, SD = 4.50 for non-perpetrators; M = 10.44, SD = 3.85 for perpetrators). 3.2. Dating violence and maltreatment There was no difference in dating violence self-report by prior maltreatment status overall, but there was a significant association between a report of past molestation and self-reported dating victimization (see Table 1). There was also a significant association between reporting a history of more than one form of maltreatment and dating violence victimization (21% v 12%; ×2 = 4.04, df = 1, p = .04). Further within the group of youth reporting prior maltreatment (n = 251), girls were more likely to report victimization than boys (23% v 11%), while within the non-maltreated group (n = 88) there was no gender difference in the rate of reported victimization (14% v 13%) (CMH ×2 = 4.61, df = 1, p = .03). 3.3. Self-reported dating violence victims Because the numbers of self-reported victims was too small to use multivariate analyses to examine subtypes, subtype analysis was limited to bivariate associations using the same variables used in the multivariate model. Violence was divided into three categories: (1) any physical or sexual harm (n = 21); (2) threats but no physical or sexual harm (n = 18); and (3) verbal violence only (n = 18). The only statistically significant association was between gender and subtype. Females were much more likely (×2 (2) = 14.76, p = .0006) to be among the respondents reporting

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

565

Table 1 Sample characteristics (N = 339)

Gender % Female % Male Ethnic/racial group % Of color % Caucasian # Times placed in foster care % 1 time % 2 or more times Age at first placement into care % b6 years % 6–11 years % 12–17 years Current placement type % Relative or friend % Foster home % Group home/residential treatment/hospital % Other Ever resided % Residential/institution/hospital % On the street % Other Stable residence last 12 mo % Yes % No History of maltreatment (any type) % Yes % No History of physical abuse % Yes % No History of physical neglect % Yes % No History of sexual abuse % Yes % No History of emotional abuse % Yes % No Ever PTSD % Yes % No Ever used drugs % Yes % No

Sample

Victims

57 43

21 12

56 44

13 21

78 22

15 24

17 25 58

13 16 18

28 28 40 4

12 16 20 23

37 43 20

14 20 13

37 63

11 20

74 26

18 14

45 55

20 14

45 55

16 17

χ2 value

Perpetrators

4.69 ⁎ 4.13 ⁎

0.09 8 7 0.00 8 8

3.30+

0.00 8 8

0.61

1.18 5 7 9

0.44

0.06 5 10 9 0

2.62

4.86 ⁎

0.96 7 10 6

0.85 9 6

1.56

0.10 8 7

0.04

28 10

44 55

21 13

15 85

36 13

50 50

26 8

4.49 ⁎

4 10 0.86

35 65

χ2 value

17.62 ⁎

0.78 6 9 3.47+ 12 6

3.73+ 15.43 ⁎ 18.36 ⁎

1.44 10 6 1.30 12 7

14.90 ⁎

14 2

⁎ p ≤ .05; + .05 b p b .10.

some form of physical/sexual violence (95% were female) compared to threats but no physical/ sexual violence (39% were female) or verbal violence alone (72% were female). Because of the similarity of variables of interest across subtypes and the small sample size, logistic regression analysis was conducted on any reported victimization. A variable had to be

566

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

statistically significant or near significant in one of the steps to be included. The model was constructed (n = 339; violence = 57) using a stepwise approach to entering variables as described in the Methods section. The final model (Likelihood Ratio ×2 (10) = 64.24, p b .0001) fits the data relatively well. Due to a significant interaction term with race, a higher tendency to self-report among females is more pronounced among the white respondents. Self-report of prior molestation and foster care history variables (ever on the street and in stable residence last 12 months) were significant until controls for any PTSD history and drug use were entered (see Table 2). Having had three or more dating partners increased the likelihood of reporting dating violence by over 4 times. 3.4. Self-reported dating violence perpetrators Overall there were no significant differences in characteristics of youth who reported victimization compared to youth who self-reported perpetration. Most youth, however, reported using only verbal abuse (20 out of 27) and 22 out of 27 perpetrators also identified as victims. Because of the difficulty in separating victims from perpetrators in this sample, a multivariate model of perpetration was not constructed. The five “perpetrator only” subjects were generally similar to the perpetrators who also reported being victims. There was a higher rate of PTSD history among the five “perpetrator only” subjects compared to the mixed perpetrator/victims (60% v 13%). Also, 100% of the five “perpetrator only” subjects reported some drug use. 3.5. Follow-up questions Overall, 54% of the self-reported victims of dating violence reported continuing the same dating relationship. Of the variables used in modeling victimization, only a past history of being Table 2 Odds ratios from logistic regression model: any self-reported dating violence victimization (n = 339; DV = 57) Odds ratio Gender Female 5.13⁎ 3.33⁎ Race/ethnic Of color 1.93 2.00 Female⁎of color 0.16⁎ 0.17⁎ Model chi-square 16.80, df = 3, p = .0008 Maltreatment Sexual abuse 2.88⁎ Model chi-square 27.29, df = 4, p = .0001 Placement history Ever on street Ever in institution Stable for 12 mo Model chi-square 37.87, df = 7, p b .0001 PTSD history Drug use Model chi-square 52.34, df = 9, p b .0001 Dating partners 3 or more Model chi-square 64.24, df = 10, p b .0001; Max rescaled r-square = .29 ⁎p ≤ .05; +.05 b p b .10.

3.71⁎

3.58⁎

3.62⁎

3.25 0.13⁎

2.61 0.14⁎

3.06+ 0.14⁎

2.46⁎

2.22⁎

2.45⁎

2.33⁎ 1.09 0.46⁎

1.59 0.96 0.51+

1.68 1.05 0.59

2.07+ 2.96⁎

2.29⁎ 2.67⁎ 4.95⁎

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

567

in a hospital or residential treatment center was associated with remaining in the violent relationship (60.9% v 20%, Fisher Exact p = .03). Sixty percent of the youth who remained in the violent relationship told someone about the violence and 56% of those youth reported that it was helpful. Among those who stayed in the relationship, 57% (n = 17) reported that the violence continued to occur. There was no association between help-seeking and self-report of violence continuing. 3.5.1. Help-seeking About 49% (n = 27) of dating violence victims and 45% (n = 12) of dating violence perpetrators reported telling someone about it at least once. Of those who disclosed the dating violence, 7 (23%) told a friend, 2 (7%) told a child welfare worker, 13 (43%) told a family member, 5 (17%) told a foster parent, and 8 (27%) told another unspecified person. Twenty (69%) youth reported that the person they told was helpful. Help was described in one of the following ways: “helped me end the relationship” (14%), “helped me control my behaviors” (9%), or “helped my partner control the violent behaviors” (48%). 4. Discussion The present study adds to the scant literature on dating violence among older youth in foster care. The current overall prevalence rate is under half of that reported in the prior published study of dating violence among foster youth. This variation may be due to the fact that the prior study administered the survey after a presentation on dating violence (Jonson-Reid & Bivens, 1999). Because of differing means of measuring dating violence in other studies, the prevalence rate in this study is also difficult to compare to studies of other youth populations. Limiting the current study's measure of prevalence to physical or sexual harm (about 5.9%) for comparison, the present rate was substantially higher than that reported from the National Crime Victimization Survey (1.7% for girls and less than 1% for boys) (Rennison & Welchans, 2000). Consistent with our hypothesis and prior studies (Rennison & Welchans, 2000), females reported a higher level of victimization than males. Gender differences in rates, however, were confined to those reporting a history of maltreatment. This descriptive finding is consistent with Wekerle et al. (2001). Probably due to sample size, there was no significant interaction between gender and maltreatment in the multivariate model. There was, however a significant main effect for gender and interaction effect between gender and race. This gender/race interaction is consistent with some prior literature on children in foster care related to delinquency and sexual risk behaviors (Auslander et al., 2002; Jonson-Reid & Barth, 2000). While we found no association with gender and report of perpetration, these results must be interpreted with caution. There were very few cases reporting anything other than the use of verbal abuse and studies of adults tend to find greater gender differences in more severe forms of violence (Tjaden & Thoennes, 1998). Further, in our study perpetration almost always cooccurred with victimization suggesting that the subjects reporting perpetration perceived themselves to be in mutually abusive relationships. Our hypothesis regarding a relationship between maltreatment and dating violence was partially supported. In bivariate analyses, maltreated females reported a higher rate of victimization than maltreated males. It is not known if maltreatment in childhood may be differentially associated with dating violence by gender. In multivariate analyses, a self-reported history of sexual abuse but not other maltreatment was significant. Studies of women have found associations between prior maltreatment as a child and both adult rape and adult partner violence

568

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

(Banyard, Arnold, & Smith, 2000; Collins et al., 1999; Tjaden & Thoennes, 1998). For some youth, considerable time had elapsed since they were placed into care. It is not known how this may have influenced their recall of maltreatment. Future studies should attempt to link administrative child abuse reporting data to see if and how maltreatment type is associated with dating violence among foster youth. The hypothesis that characteristics of a youth's foster care history would be associated with dating violence was also partially supported. Prior to controlling for PTSD or substance use, having spent time on the street was associated with a higher likelihood of self-reported victimization. Consistent with Widom's study of delinquency following foster care (1991), a stable placement for the past 12 months was associated with a decrease in self-reported dating violence until controls for PTSD and substance use were added. Although placement stability became non-significant when controls for PTSD and drug use were added, the coefficient value did not change a great deal across steps. Thus this change may be, in part, an artifact of the small sample size for this outcome. There are several theoretical reasons to expect an association with placement stability including this as a proxy for a decrease in recurrent victimization or improved ability to form and maintain positive attachments (Cicchetti, 1991; Crittenden & Claussen, 2002; Wekerle & Wolfe, 1998). On the other hand, it may be that placement stability becomes less of a factor among those youth who are unable to exit care prior to emancipation. Future studies should continue to assess the impact of placement characteristics. 4.1. Mental health and substance abuse Similar to Wekerle et al. (2001), we found a consistent relationship between PTSD and selfreport of victimization in a dating relationship. Our study did not find a relationship between PTSD and self-report of perpetration of dating violence overall, but it was more common among the five youth who self-reported as “perpetrators only”. Of course, in a cross-sectional study it is impossible to determine whether or not the PTSD history is tied to early experiences or related to the reported dating violence. Any drug use was associated with an increased likelihood of experiencing dating violence. Howard and Wang (2003a,b) found that drug use was associated with victimization reported by girls but not boys. In our study, only 17 boys reported victimization making it impossible to do separate models by gender. Further, in a cross-sectional study it is not possible to know whether or not drug use co-occurs with dating violence or multiple partner behavior rather than being a risk factor for dating violence behaviors. 4.2. Ongoing violence Over half of the youth reporting relationship violence remained in that relationship. Although lower than the prior study of dating violence among foster youth (74%; Jonson-Reid & Bivens, 1999), the rate of remaining in the relationship in the current study is four times higher than reported in a study of dating violence in a general high school population (Jezl et al., 1996). Other studies have found that dating violence during adolescence is a stronger predictor of victimization of women during college than victimization in childhood (Smith, White, & Holland, 2004). The fact that a high proportion of foster youth both experience dating violence and remain in these abusive relationships may place them at even greater risk of domestic violence as adults.

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

569

4.3. Help-seeking A prior investigation of dating violence among foster youth found that the majority of youth wanted more information about dating violence (Jonson-Reid & Bivens, 1999). While youth in the present study were not surveyed about desire for intervention, about half of the youth reporting dating violence in the current study shared this fact with someone else and most reported this as helpful in some way. It is not known if youth who seek help during adolescence face less risk of later violence in adult relationships. 4.4. Limitations The literature on dating violence among foster youth is too scant to assess generalizability. The increased risk of dating violence, however, is consistent with studies of other negative partner behaviors during foster care (Auslander et al., 2002). Care must also be taken in applying current findings only to those youth emancipating from foster care. Children and youth who enter foster care but are able to exit to reunification, relative care or adoption may have very different experiences. Our study was cross-sectional in nature, meaning that causal associations between risk factors and dating violence could not be made. Further our variables were based on self-report, meaning that foster care history and maltreatment history are based on the subjects' recall. This reliance on recall is typical of most studies of dating violence (see Smith & Williams, 1992; Wekerle et al., 2001). Future studies should attempt to cross-reference prior history with other sources like child abuse and neglect reports and foster care records. Despite the limitations, this article adds valuable information regarding the high incidence of dating violence among foster youth and possible correlates and follow-up issues that can be used in prevention and intervention programs. Implications for research and practice are now discussed. 4.5. Implications More research on dating violence among adolescents in general as well as among youth in foster care is needed to better understand the origins and etiology of this problem. It is important to understand how and if partner violence differs during adolescence compared to adulthood. Longitudinal studies are needed to understand what risk factors are particularly salient for subpopulations of children (i.e., females, differing maltreatment type histories, etc.) in the child welfare system. Services information is needed to understand if and when the provision of child welfare, mental health or other services may moderate the association of risk factors with outcomes like dating violence. Theory building in the area of dating violence is just emerging and suffers from the lack of longitudinal study as well as studies comparing different forms of violence. For example, the fact that most abused and neglected children do not seem to experience later dating violence appears to counter global applications of theoretical explanations like the “cycle of violence” or victim– victimizer role theories (Crittenden & Claussen, 2002; Widom, 1991). Studies that simultaneously measure multiple forms of violence (e.g., violent delinquency, dating violence, violence toward animals, adolescent perpetrators of child abuse, violent bullying, etc.) might find greater support for such theories. In other words, different youth may be at risk for developing different types of violent behaviors or victim roles. Such research will require collaboration between typically isolated areas of violence research.

570

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

Although the full extent of the problem of dating violence among foster youth is unclear, emerging evidence as well as theory suggests that the prevalence is higher among foster youth. It seems wise, therefore, to provide targeted intervention. The ability to implement a dating violence prevention/intervention program for youth in foster care is greatly helped by the existence of national funding and support for independent living programs. The fact that most youth who seek help in this situation do so from non-professionals, however, suggests that targeted programming will be more effective in combination with a public health approach to sensitizing the general population to this issue and means of finding help. It is hoped that the present study will help encourage ongoing research as well as program development in this area. Acknowledgement This research was funded by the National Institute of Mental Health R01 MH 61404. References Auslander, W. F., McMillen, J. C., Elze, D., Thompson, R., Jonson-Reid, M., & Stiffman, A. (2002). Mental health problems and sexual abuse among adolescents in foster care: Relationship to HIV risk behaviors and intentions. AIDS and Behavior, 6(4), 351−359. Banyard, V., Arnold, S., & Smith, J. (2000). Childhood sexual abuse and dating experiences of undergraduate women. Child Maltreatment, 5(1), 39−48. Bergman, L. (1992). Dating violence among high school students. Social Work, 37(1), 21−27. Bernstein, D., & Fink, L. (1998). Childhood Trauma Questionnaire: A retrospective self-report. San Antonio: The Psychological Corporation. Cicchetti, D. (1991). Fractures in the crystal: Developmental psychopathology and the emergence of self. Developmental Review, 11, 271−287. Collins, K., Schoen, C., Joseph, S., Duchon, L., Simantov, E., & Yellowitz, M. (1999). Health concerns across a woman's lifespan: The Commonwealth Fund 1998 Survey of Women's Health, The Commonwealth Fund. Available [online]: http://www.cmwf.org/publications/publications_show.htm?doc_id=221554 Courtney, M., & Barth, R. (1996). Pathways of older adolescents out of foster care: Implications for independent living services. Social Work, 41, 75−84. Crittenden, P., & Claussen, A. (2002). Developmental psychopathology perspectives on substance abuse and relationship violence. In A. Wall & C. Wekerle (Eds.), The violence and addiction equation theoretical and clinical issues in substance abuse and relationship violence. Brunner-Routledge. Gray, H., & Foshee, V. (1997). Adolescent dating violence: Differences between one-sided and mutually violent profiles. Journal of Interpersonal Violence, 12, 126−141. Hickman, L., Jaycox, L., & Aranoff, J. (2004). Dating violence among adolescents. Prevalence, gender distribution, and prevention program effectiveness. Trauma, Violence and Abuse, 5, 123−142. Howard, D., & Wang, M. (2003). Psychosocial factors associated with adolescent boys' reports of dating violence. Adolescence, 38(151), 519−533. Howard, D., & Wang, M. (2003). Risk procedures of adolescent girls who were victims of dating violence. Adolescence, 38(149), 1−14. Jezl, D., Molidor, C., & Wright, T. (1996). Physical, sexual and psychological abuse in high school dating relationships: Prevalence rates and self-esteem issues. Child and Adolescent Social Work Journal, 13(1), 69−87. Jonson-Reid, M. (2002). Exploring the relationship between child welfare intervention and juvenile corrections involvement. American Journal of Orthopsychiatry, 74(4), 559−576. Jonson-Reid, M., & Barth, R. (2000). From placement to prison: The path to adolescent incarceration from child welfare supervised foster or group care. Children and Youth Services Review, 22(7), 493−516. Jonson-Reid, M., & Bivens, L. (1999). Foster youth and dating violence. Journal of Interpersonal Violence, 14(12), 1249−1262. Maxfield, M., & Widom, C. (1996). The cycle of violence. Revisited 6 years later. Archive of Pediatrics and Adolescent Medicine, 150, 390−395.

M. Jonson-Reid et al. / Children and Youth Services Review 29 (2007) 557–571

571

McDonald, T., Allen, R., Westerfelt, A., & Piliavin, I. (1993). Assessing the longterm effects of foster care: A research synthesis (Special report 57–63). WI: Institute for Research and Poverty. McMillen, C., Auslander, W., Elze, D., White, T., & Thompson, R. (2003). Educational experiences and aspirations of older youth in foster care. Child Welfare, 82(4), 475−495. McMillen, J., Zima, B., Scott, L., Auslander, W., Munson, M., Ollie, M., et al. (2005). Prevalence of psychiatric disorders among older youths in the foster care system. Journal of the American Academy of Child and Adolescent Psychiatry, 44(1), 88−95. O'Keefe, M. (1997). Predictors of dating violence among high school students. Journal of Interpersonal Violence, 12, 546−568. O'Keefe, M., & Treister, L. (1998). Victims of dating violence among high school students. Are the predictors different for males and females? Violence Against Women, 4, 195−223. Reich, W., Welner, Z., & Herjanic, B. (2002). Diagnostic interview for children and adolescents-IV. Multi-Health Systems Incorporated. Rennison, C. M., & Welchans, S. (2000, May). Intimate partner violence. Bureau of Justice statistics special report: U.S. Department of Justice. Robins, L., Cottler, L., Bucholz, K., & Compton, W. (1995). Diagnostic interview schedule for DSM-IV. St. Louis, MO: Washington University in St. Louis. Robins, L., Helzer, J., Croughan, J., & Ratcliff, K. (1981). National Institute of Mental Health diagnostic interview schedule: Its history, characteristics, and validity. Archives of General Psychiatry, 38, 381−389. Russell, D. (1986). The secret trauma: Incest in the lives of girls and women. New York: Basic Books. Scher, C., Stein, M., Asmundson, G., McCreary, D., & Forde, D. (2001). The Childhood Trauma Questionnaire in a community sample: Psychometric properties and normative data. Journal of Traumatic Stress, 14(4), 843−857. Smith, P., White, J., & Holland, L. (2004). A longitudinal perspective on dating violence among adolescent and collegeage women. American Journal of Public Health, 93(7), 1104−1109. Smith, J., & Williams, J. (1992). From abusive household to dating violence. Journal of Family Violence, 7, 153−165. Tjaden, P., & Thoennes, N. (1998, Nov). Prevalence, incidence, and consequences of violence against women: Findings from the National Violence Against Women Survey. Research in brief. Washington DC: National Institute of Justice. Tjaden, P., & Thoennes, N. (2000). Extent, nature, and consequences of intimate partner violence: Findings from the National Violence Against Women Survey.Washington DC: Department of Justice (US); Publication No. NCJ 181867. Retrieved Nov. 2, 2006 from, http://www.ojp.usdoj.gov/nij/pubs-sum/181867.htm Wekerle, C., & Wolfe, D. (1998). The role of child maltreatment and attachment style in adolescent relationship violence. Developmental Psychopathology, 10(3), 571−586. Wekerle, C., Wolfe, D., Hawkins, D., Pittman, A., Glickman, A., & Lovald, B. (2001). The value and contribution of youth self-reported maltreatment history to adolescent dating violence: Testing a trauma mediational model. Development and Psychopathology, 13, 847−871. Widom, C. (1991). The role of placement experiences in mediating the criminal consequences of early childhood. American Journal of Orthopsychiatry, 61, 195−209. Wolfe, D., Wekerle, C., Scott, K., Straatman, A., Grasley, C., & Reitzel-Jaffe, D. (2003). Dating violence prevention with at-risk youth: A controlled outcome evaluation. Journal of Consulting and Clinical Psychology, 71, 279−291.