Rates and Risk Factors for Sexual Violence Among an Ethnically Diverse Sample of Adolescents FREE

Adolescents and young adults are 4 times more likely to be victims of sexual assault than women in all other age groups,^1- 2 with a lifetime prevalence of rape/attempted rape as high as 42%.^1,3- 17 In most of these cases (80%-92%), the perpetrator is a date or acquaintance of the victim.^10,18- 22

A robust risk factor for the occurrence of sexual violence is a history of sexual victimization as either a child (aged <14 years) or an adolescent (aged ≤18 years).^{5,13- 14,21- 23} Other identified risk factors include earlier age at menarche and/or first date,¹² number of dating and sexual partners,^5,17,24- 25 history of dating violence,²⁶ and a sexually active peer group.¹² The social context of the date, date-specific behaviors, and the stage of the relationship also have been related to the risk of vulnerability to sexual violence, especially among white young women.^{9,15,17,21,27- 28} For example, when the male initiates the date, pays the date expenses, and drives,^11,24,28 the risk for sexual violence is increased. In addition, studies have suggested that forced sex is more likely to occur later in the dating relationship compared with earlier.²⁹

An important catalyst leading to the occurrence of sexual victimization by a date or acquaintance is the use of alcohol by the victim, perpetrator, or both.^{5- 6,10,16,24,26,28,30- 34} Research suggests that alcohol use within the dating or social situation may lead to the misinterpretation of friendly nonsexual cues as sexual invitations, diminish coping responses, and limit the ability of a woman to ward off a potential attack.^17,34 These studies are based largely on white college women^{1,5- 6,9- 10,16- 17,24,26,32- 33} and, thus, do not address the role of substance use in sexual violence among ethnically diverse populations. Abbey et al¹⁷ found that African American college women were less likely than white college women to report the concurrent use of alcohol at the time of sexual violence.

In general, there is insufficient knowledge about the role of risk factors for sexual violence among ethnically diverse samples. Few studies have investigated racial/ethnic differences in sexual victimization despite suggestions that they may differ.^5,27 Two studies of adolescents and adults found no race/ethnicity differences,^35- 36 but another study found that African American college women were somewhat more likely to report experiencing sexual assault than their white counterparts.¹⁷ Thus, studies using ethnically diverse populations are needed to identify potential risk factors so that prevention and intervention programs for all young women may be developed.

We sought to describe the occurrence of date/acquaintance rape and other coercive sexual experiences in the past year among an ethnically diverse population and to evaluate contributing risk factors. Risk factors examined included demographic, reproductive, victim-specific (history of abuse and dating practices in the past year), date-specific (including the nature of the relationship), and substance use behaviors of the victim and perpetrator. Finally, we wanted to determine whether unique profiles existed for victims of rape/attempted rape vs verbally coerced sex when compared with young women who did not report any unwanted sexual experience. We hypothesized that younger adolescent age, a longer dating relationship, and alcohol use during the date would be significantly associated with date/acquaintance rape.

Those eligible to participate were adolescent and young female adults aged 14 to 23 years who presented for health services at the Mount Sinai Adolescent Health Center, New York, NY, between October 3, 2000, and February 28, 2002; self-identified as white, African American, Puerto Rican, or Dominican; were not married and/or living with a male partner (because forced sex would not be considered date/acquaintance rape); were not currently pregnant; did not demonstrate an obvious cognitive or mental impairment; and reported being on 1 or more dates in the past year. Other racial/ethnic groups were not included because of low numbers. Recruitment occurred consecutively during randomly selected weekly clinic sessions; 908 women screened following clinic registration were deemed eligible. Time constraints led 137 subjects (15%) to refuse participation; 35 subjects (4%) agreed to participate but did not complete or return their questionnaires. It is possible that some eligible women were missed because of clinic traffic patterns. Based on the comparison of personal codes generated by the participants, 36 duplicate respondents were removed. Eleven questionnaires were eliminated because of systematic patterning of answering. Thus, analyses were conducted with 689 adolescent and young adult women. Eligible subjects who refused participation completed a short questionnaire; no differences in demographic or dating characteristics were detected for those individuals who completed vs did not complete the questionnaire.

The Institutional Review Board approved the study, and all participants provided written informed consent. A bilingual research assistant asked young women in the clinic waiting area if they would like to participate in a study of dating behaviors. Participants completed the self-administered questionnaire in either English or Spanish while in the waiting or examination room. To assure anonymity, subjects were instructed to avoid placing identifying information on the questionnaire. Participants received $15 for their effort and were given handouts containing information about relationship violence, alcohol and other drug abuse, and agencies that provided services to sexual assault survivors.

The self-report measure was finalized after pilot testing (n = 50). Accurate Spanish translation was evaluated using forward-backward translation protocols. Only 4 subjects completed the Spanish version. Thus, no systematic comparisons could be made between the English and Spanish versions. Four broad categories of information were collected: demographic and reproductive characteristics, sexual and physical victimization by a dating partner, date-specific behaviors and social context of that date, and substance use by the victim and perpetrator.

Items included subject age, parental education and employment status, race/ethnicity, current living arrangements, school enrollment, last grade completed, religious attendance, birthplace, preferred language, partner age discrepancy (female age subtracted from male partner’s age), age at menarche and first sexual intercourse, gravidity, and lifetime number of sexual partners. A 12-item ethnic identity scale³⁷ that uses a 4-point response option (from strongly agree to strongly disagree) to assess the extent to which subjects endorsed the importance and value of their ethnic heritage was used. For example, subjects were asked to agree or disagree with having made an effort to learn more about their ethnic group, to think about how their ethnic group affects their lives, and to determine their attachment to their ethnic group. A sum score was created, with lower scores reflecting greater ethnic affiliation.

The Sexual Experiences Survey^38- 39 was used to assess sexual violence and unwanted sexual experiences in the past 12 months and during the lifetime. This 10-item measure is the most commonly used instrument to evaluate sexual victimization among young women,^{9- 10,15- 16,24,28,38- 39} and reliably classifies the occurrence of sexual victimization when used anonymously.³⁸ Consistent with other studies, subjects were classified, based on the most severe item to which they said yes, into 1 of 4 categories: no victimization, unwanted sexual contact (touching or kissing without permission), rape/attempted rape (use of physical force, alcohol, or other drugs or position of power), and verbal sexual coercion (being talked or pressured into having unwanted intercourse).^{9- 10,15- 16,24,38- 39} To classify a young woman as having experienced a history of abuse, each adolescent was asked to report the first age of occurrence of sexual victimization on the lifetime index of the Sexual Experiences Survey.^38- 39 If she reported an age of 14 years or younger, it was considered to be child abuse; otherwise, she was identified as experiencing prior victimization as an adolescent.

Each young woman also completed the verbal and physical abuse subscales from the Dating Violence Questionnaire,²⁶ which assessed violence from her most recent dating partner and prior verbal and physical violence experienced from the identified perpetrator, if applicable. Specifically, each young woman was asked to report the frequency (0-≥4 times) with which this dating partner verbally (criticized, humiliated, or shouted) or physically (pushed, shoved, restrained, or slapped) abused her. Prior physical violence was categorized into none, mild to moderate, or severe. Consistent with other research,^1- 2 severe physical abuse from this dating partner was defined as 1 or more incidents during which she was choked, strangled, punched, or threatened with a knife or gun⁴⁰; mild to moderate violence was defined as 1 or more episodes of being pushed or shoved, being held from leaving, having objects thrown at her, being verbally threatened with physical violence, or being subjected to reckless driving. Past verbal aggression by a dating partner was categorized as none, mild to moderate, or severe. Subjects who did not report any prior verbal abuse from this dating partner were coded as 0, those whose total score ranged from 1 to 20 were coded as mild to moderate, and those with total scores 2 SDs or more above the mean were coded as severe.

Subjects were asked to describe their most recent date (how long they had known this person, number of prior dates with this person, and whether this person was their exclusive sexual partner), their date’s age, who initiated the date, who paid the expenses for the date, about date type (single, double, or group date), whether the date was chaperoned, what transportation was used, where they went (movie, sporting event, or dinner), and about the occurrence or nonoccurrence of “parking” and/or going to an apartment or house when no other people were present. The level of involvement in the relationship was evaluated using a 10-point rating scale; higher ratings indicated a more serious commitment. Subjects who reported an unwanted sexual experience in the past 12 months were asked parallel questions that allowed them to describe in detail the behaviors and context of that date during which the unwanted sexual contact occurred.

We assessed the frequency of lifetime substance (tobacco, hard liquor, marijuana, and other illicit drug) use and use in the past 12 months for the following categories: 0, 1 to 2, 3 to 5, 6 to 9, 10 to 19, 20 to 39, and 40 or more times.^41- 42 We also assessed alcohol and other illicit drug use during their most recent date by the young woman and her date, including the number of drinks consumed before and during the date, the use of any illicit drugs, pressure experienced to use alcohol and/or other illicit substances, and how the subject felt as a result of using substances. If a subject reported an unwanted sexual experience, parallel questions were elicited about substance use during the context of this date.

All data were manually entered, with 20% of surveys evaluated for accuracy. Subjects were grouped according to whether they reported 1 or more vs no unwanted sexual experiences in the past 12 months. Specific variables associated with assault were then identified by comparing demographic, reproductive, and victim-specific risk factors of sexually assaulted vs nonassaulted subjects using the χ² or t test. Date-specific behaviors associated with assault were identified by comparing dating behaviors associated with dates that included sexual victimization with those of participants on their most recent date who did not report assault. Separate logistic regression analyses using simultaneous entry in 4 blocks (demographic/reproductive, victim-specific risk factors, dating behaviors and social context characteristics of that date, and substance use by the victim and perpetrator) were conducted to identify correlates of verbal sexual coercion and rape/attempted rape. These 2 victim groupings were chosen based on past literature and to account for the fact that rape is a legal definition that varies across states. We did not conduct analyses on unwanted sexual contact because of the limited number of young women who reported this level of victimization. Variables significant at the P<.10 level were considered for entry into the logistic regression analysis, and the colinearity among significant variables was assessed. We included potential interactions in all models, but none were significant. Multivariate odds ratios with associated 95% confidence intervals were computed and reported. All data were analyzed using a commercially available software program (SPSS for Windows, 11.0; SPSS Inc, Chicago, Ill).

Approximately 30% (n = 203) of the young women reported having an unwanted sexual experience during the previous 12 months; 21 (3%) reported that their sexual victimization occurred during their most recent date. Specifically, 41 (6%) of the 689 subjects reported unwanted touching, 71 (10%) reported verbal sexual coercion, 91 (13%) reported rape/attempted rape, and 486 (71%) did not report any sexual victimization. Of those who reported sexual victimization, 35% (192/542) were between the ages of 14 and 19 years and 28% (41/149) were between the ages of 20 and 23 years.

Few differences in demographic and reproductive characteristics emerged (Table 1). No differences in victimization groups were noted for age, those who reported being born in the United States, employment status, those enrolled in school, having repeated 1 or more grades in school, or early menarche (aged ≤10 years).

Victimization as an adolescent was associated with rape/attempted rape and verbal sexual coercion (Table 2). We did not find differences in age at first date (14.2 years for no victimization vs 14.8 years for rape/attempted rape), but did find that those who reported rape/attempted rape were less likely to report having a bisexual dating history (12% vs 6%; P<.10)

Several date-specific behaviors and social context variables were related to assault status (Table 3). Those young women who reported rape/attempted rape were significantly less likely to report that this was a single date (57% vs 74%; P<.01) and that this person asked them out for this date (69% vs 85%; P<.001). However, no differences between those who reported verbal sexual coercion were found for these same variables. Those who reported rape/attempted rape (45%) and verbal sexual coercion (49%) were significantly more likely to report that they went to the date’s house to be alone compared with those who did not report any victimization (28%). Going to a park or to a friend’s house to be alone and whether or not the date drove were unrelated to sexual victimization. Those who experienced verbal sexual coercion were also significantly more likely to report high levels of past verbal aggression from this dating partner.

We found that those young women who reported rape/attempted rape were significantly more likely to report a history of using hard liquor (71%), marijuana (45%), and other illicit drugs (9%) when compared with those who did not report an unwanted sexual experience (58%, 34%, and 4%, respectively; P<.02). Similar findings were detected for those who reported verbal sexual coercion (72%, 56%, and 9%, respectively; P<.02). No difference in alcohol use by the victim during the date was found (18% vs 23% of rape/attempted rape victims). Interestingly, those who reported verbal sexual coercion were significantly less likely to drink during this date (6%, P<.01). Young women who experienced rape/attempted rape were significantly more likely to report pressure by the date to use either alcohol or marijuana (31% and 18%, respectively) compared with those not victimized (4% and 4%, respectively; P>.001). Those who reported verbal sexual coercion were also more likely to report experiencing pressure to use alcohol (12%, P<.01).

Perpetrator use of alcohol or marijuana was significantly associated with rape/attempted rape. Young women who reported that their date drank 1 or more drinks (40%) or used marijuana (25%) were significantly more likely to report rape/attempted rape than those who did not report any victimization (20% and 14%, respectively;P<.01). Of interest, only perpetrator use of marijuana was associated with verbal sexual coercion (24%, P<.03). Too few subjects (<1%) reported the use of illicit drugs such as ecstasy during the date to perform these analyses.

Distinct profiles of risk emerged for each type of sexual victimization when compared with no victimization in the past year (Table 4). As expected, prior sexual victimization as an adolescent, either verbal or physical victimization by the dating partner, and going to the date’s house or apartment to be alone were associated with sexual assault. Although the length of the relationship was not significantly related to sexual victimization, a lowered level of romantic involvement increased the likelihood of reporting rape/attempted rape and verbal sexual coercion. Contrary to expectations, the use of alcohol by the victim or dating partner was not significantly associated with either type of sexual assault.

Almost 1 in 4 urban adolescents who presented to a primary care health setting reported having experienced verbal sexual coercion or rape/attempted rape by a date or acquaintance in the past year. In contrast to other research, we did not detect increased risk among adolescents compared with young adult women.^1- 2 However, unique profiles associated with verbal sexual coercion and rape/attempted rape among this urban population were identified.

Demographic characteristics were associated with rape/attempted rape, but not with verbal sexual coercion. Similar to previous findings,¹⁷ we found that the highest proportion of those reporting a forced sexual experience in the past year were African Americans (100 [35%] of 289), followed by Dominicans (29 [25%] of 118). In addition, decreasing ethnic identity was associated with increased reports of rape/attempted rape. Ethnic identity is a psychological construct reflecting various aspects of identification and membership in an ethnic group, including the way one feels about his or her ethnic background and an awareness of cultural values and mores.³⁷ It may be that there are no differences in the rates of sexual violence among young women of different ethnicities because those who experience a high degree of affiliation with their identity and culture are less likely to disclose forced sexual experiences even when using an anonymous survey. Alternatively, strong feelings of belonging to one’s culture may be protective against sexual assault because they may contribute to partner selection and mitigate other social and date-specific risk factors. Future research is needed to elucidate the interrelationships among ethnic identity, date-specific behaviors, and disclosure of sexual violence.

The lack of relationship between alcohol use by the perpetrator and/or victim among young women who reported rape/attempted rape is an important finding. Three explanations for the fact that our findings differ from studies of white college women are possible. First, women younger than 21 years and those who do not attend college may have less access to alcohol than women who are enrolled in college. Second, minority adolescents and young women may be less likely than their white peers to drink alcohol on a date. Finally, reports of sexual victimization among adolescents may vary based on alcohol use because either memory is better when not intoxicated or they feel more responsible and less victimized when they have been drinking.

Prior research has demonstrated a relationship between abuse as a child and sexual victimization as an adolescent.²² Consistent with other researchers,^43- 44 we found that prior victimization as an adolescent was a strong predictor of sexual victimization within this ethnically diverse sample of young women.

Several variables within the social context of the date were related to the reporting of verbally coerced sex. Specifically, an increasing age discrepancy between the partners, having at least 6 prior dates, and experiencing verbal abuse from this dating partner were all associated with verbally coerced sex. Concomitant reports of not drinking during the date, of pressure to use alcohol, and of verbal aggression suggest that male partners may have been using verbal and psychological abuse as a coercive strategy to control these young women.

Reports of experiencing prior severe physical violence from the dating partner were the strongest association with reports of experiencing rape/attempted rape, while mild to moderate physical violence increased the likelihood of reporting any type of sexual victimization. Prior research among a Midwestern sample of adolescents found that sexual violence occurred in tandem with physical violence.⁴⁵ Another study among college-aged women found that those who were physically assaulted in the past year were significantly more likely to be sexually assaulted in the same year.⁴⁵ Although our data do not allow for the determination of the causal pathways between physical and sexual violence, we speculate that for some women, having a dating partner who is physically aggressive may be a precursor to sexual violence later in the relationship. Regardless, prevention programs should include specific strategies that teach young women how to effectively mitigate the verbal abuse and other coercive verbal strategies that their partners may use.

Several limitations deserve mention. We cannot be sure that each young woman who experienced an unwanted sexual experience during the past year disclosed these experiences to us. However, Koss and Dinero⁵ report that anonymous, written, self-report surveys, compared with face-to-face interviews, result in higher rates of participation among young adults. Second, our results may not generalize to adolescents who reside in other geographic areas, to those who are not presenting for health care, and to those at risk for sexual violence from same-sex relationships. Finally, these data are cross-sectional and subject to recall bias, especially for history of sexual victimization.

More than 1 in 4 urban adolescents and young adult women who sought care at an adolescent health care facility reported having experienced sexual assault by a date or acquaintance within the past year. Distinct risk profiles for rape/attempted rape and verbal sexual coercion compared with those who had not experienced forced sex were identified. Decreased ethnic identity and previous physical aggression by the dating partner were important correlates for rape/attempted rape, but not for verbal sexual coercion. Taken together, our findings suggest the need to educate young women on how to effectively manage verbal and psychological abuse and other types of coercive behaviors that may be exerted by male partners during the context of a date. Future studies are needed to determine the temporal relationship between physical and sexual aggression within dates and dating relationships and the exact role of ethnic identity in the experience and disclosure of sexual assault.

Literature examining the occurrence of date/acquaintance rape and other coercive sexual violence, and the contributing role of alcohol, is compromised by the following: (1) a failure to sample sufficiently large numbers of minority adolescents and young adults to examine differences among women of different ages, (2) studies’ exclusion of young adult women who do not attend college, (3) use of different time frames for recall (lifetime occurrence of sexual assault vs current drinking and dating practices), and (4) an absence of studies that investigate the role of alcohol use in sexual violence among those of diverse ethnic relationships.

We found that among an ethnically diverse population of urban adolescent females, 1 in 4 reported having experienced sexual assault by a date or acquaintance within the past year. Distinct risk profiles for rape/attempted rape and verbal sexual coercion compared with those who had not experienced forced sex were identified. In contrast to previous research using mostly college-aged white samples, victim or partner use of alcohol or other illicit drugs during the date did not predict either type of sexual assault.

Correspondence: Vaughn I. Rickert, PsyD,Center Center for Community Health and Education, Heilbrunn Department of Population and Family Health, Mailman School of Public Health at Columbia University, 60 Haven Ave, Suite B-3, Room 306, New York, NY 10032 (vir2002@columbia.edu).

Accepted for Publication: May 27, 2004.

Funding/Support: This study was supported by grant AA 12704 from the National Institutes of Health, Bethesda, Md (Dr Rickert).

Acknowledgment: We thank the many adolescent participants who shared their experiences with us; the research assistants (Ahnya Mendes, Ada Villatora, and Maritza Martinez) who spent hundreds of hours collecting and entering the data; the professional staff of the Mount Sinai Adolescent Health Center, who allowed us to disrupt their busy clinical schedules; and Susan L. Rosenthal, PhD, and Ralph J. DiClemente, PhD, for their helpful comments on earlier drafts of the manuscript.