Health Risk Behavior and Sexual Assault Among Ethnically Diverse Women

Abstract

Sexual assault is associated with a number of health risk behaviors in women. It has been hypothesized that these risk behaviors, such as hazardous drinking, may represent women's attempts to cope with psychological distress, such as symptoms of depression and anxiety. However, extant research has failed to evaluate these relationships among ethnic minority samples or identify the mechanisms responsible for this association. The current study examined sexual assault history and two health risk behaviors (hazardous drinking and engaging in sexual behavior to regulate negative affect) in a diverse sample of 1,620 college women. Depression and anxiety were examined as mediators of the relationship between sexual assault and health risk behaviors. There was evidence of moderated mediation, such that for European American women, but not for ethnic minority women, both forms of psychological distress were significant mediators of the sexual assault/hazardous drinking relationship. In contrast, among all ethnic groups, the relationship between sexual assault and both forms of psychological distress was mediated by the use of sexual behavior as an affect regulation strategy. Results support a need to evaluate the assault experiences of ethnically diverse women, as well as the impact of the assault on their postassault experiences including health risk behaviors and psychological adjustment. Additionally, results suggest that practitioners should carefully assess health risk behaviors among victims of sexual assault and be aware that there may be differences in the risk factors and motives for these behaviors among women of various ethnic backgrounds.

Sexual assault is a significant public health problem among college women. Indeed, recent studies of college women find that 11– 20% report that they have experienced completed sexual assault (vaginal, anal, or oral penetration) perpetrated by threat or force, or that occurred when they were incapable of consenting (Kahn, Jackson, Kully, Badger, & Halvorsen, 2003; Littleton & Radecki Breitkopf, 2006; McCauley, Ruggiero, Resnick, Conoscenti, & Kilpatrick, 2009). Experiencing sexual assault also places college women at risk of a number of deleterious mental health outcomes, including the development of posttraumatic stress disorder (PTSD) and depression (e.g., Zinzow et al., 2010).

In addition, sexual assault victims are more likely than nonvictims to engage in a number of health risk behaviors. For example, sexual assault victims are more likely to engage in hazardous alcohol use than nonvictims—with hazardous alcohol use appearing to be a risk factor for experiencing sexual assault, as well as a consequence of sexual victimization (e.g., Brener, McMahon, Warren, & Douglas, 1999; McCauley et al., 2009; Najdowski & Ullman, 2009a; Nguyen, Kaysen, Dillworth, Brajcich, & Larimer, 2010; Ullman, 2003). Additionally, sexual assault victims are more likely than non-victims to engage in sexual behavior as an affect regulation strategy (Orcutt, Cooper, & Garcia, 2005). Engaging in these behaviors can be harmful to women in a number of ways. For example, common consequences of hazardous drinking among college students include having blackouts and hangovers, driving while intoxicated, suffering physical injuries due to accidents, and engaging in unintended sexual activity (Clements, 1999; Jackson, Sher, & Park, 2006; Mallett et al., 2011). In addition, engaging in sexual behavior as an affect regulation strategy has been associated with having more sexual partners, including casual partners (Cooper, Shapiro, & Powers, 1998; Gebhardt, Kuyper, & Greunsven, 2003; Patrick & Maggs, 2010). Having more sexual partners in turn is linked to increased risk of sexually transmitted infections (STI) among college students (Sipkin, Gillam, & Grady, 2003; Vivancos, Abubakar, & Hunter, 2008). Finally, engaging in these health risk behaviors may serve to increase women's vulnerability to further sexual victimization through a number of mechanisms, including increasing women's likelihood of encountering a sexually assaultive partner and decreasing their ability to defend against an assault (e.g., Orcutt et al., 2005; Testa, Hoffman, & Livingston, 2010; Ullman, Najdowski, & Filipas, 2009).

Given the association between sexual assault and health risk behaviors, it is important to understand the potential mechanisms behind these associations. The predominant hypothesis is that these behaviors represent strategies to cope with psychological distress, including distress regarding sexual assault experiences, other traumas, and general distress (Bryant-Davis, Chung, & Tillman, 2009; Nguyen et al., 2010). Supporting this possibility, Orcutt and colleagues (2005) found that having higher levels of psychological distress (depression, anxiety, and hostility) mediated the relationship between a history of childhood sexual abuse and engaging in sexual behavior as an affect regulation strategy. Similarly, in a longitudinal study, Ullman and colleagues (2009) found that having higher levels of PTSD symptoms mediated the relationship between experiencing multiple prior sexual victimizations and engaging in more problem drinking. Finally, Goldstein, Flett, and Wekerle (2010) found that drinking to cope with depression significantly mediated the relationship between a history of child maltreatment and drinking-related consequences in a sample of college women. Remarkably, although there are a number of studies examining health risk behaviors among victims of sexual violence, few studies have attempted to evaluate the potential mediators of this relationship (e.g., psychological distress). In addition, extant studies have focused primarily on victims of childhood sexual abuse, as opposed to victims of sexual assault in adolescence or adulthood. Finally, few studies have evaluated these relationships in college women.

Ethnic Differences in Sexual Victimization and Health Risk Behaviors

Another significant limitation of research examining sexual assault and its association with health risk behaviors has been an almost exclusive focus on the experiences of ethnic majority women. Most studies have utilized samples of almost entirely European American women, or did not evaluate if ethnic differences existed (e.g., Kahn et al., 2003; Littleton & Radecki Breitkopf, 2006). It is likely that cultural values are important in shaping women's conceptualization of, and responses to, sexual assault, including cultural beliefs regarding women's sexuality and appropriate behaviors for women (e.g., drinking alcohol, engaging in casual sexual behavior; Bryant-Davis et al., 2009). In addition, broader societal attitudes toward women of a particular ethnic group (e.g., stereotyping of some women as more sexually promiscuous/exotic) likely are influential in shaping women's conceptualization and responses to sexual violence (Bryant-Davis et al., 2009; Townsend, Thomas, Neilands, & Jackson, 2010). Finally, cultural norms and attitudes may affect women's victimization risk, likelihood of engaging in health risk behaviors, and likelihood of using alcohol or sexual behavior as strategies to cope with psychological distress (Bryant-Davis et al., 2009).

African American women

Turning to extant research regarding sexual victimization and health risk behaviors among ethnic minority women, there are a few studies that have examined sexual assault experiences among African American college women. These studies find that African American women generally report overall rates of sexual assault similar to those of European American women (Kalof, 2000). However, some research has suggested that African American college women may be more likely to experience a sexual assault involving physical force than European American women (Bryant-Davis et al., 2009; Urquiza & Goodlin-Jones, 1994). Conversely, it has been hypothesized that African American women may be less likely than European American women to experience a sexual assault that occurred when they were impaired due to substance use (Bryant-Davis et al., 2009).

With regard to health risk behaviors among African American college women, they have consistently been found to engage in less hazardous alcohol use than women of other ethnic backgrounds (Clements, 1999; Randolph, Torres, Gore-Felton, Lloyd, & McGarvey, 2009; Wechsler et al., 2002). Additionally, studies have supported that African American women are less likely to associate with peers who engage in heavy drinking, are less likely to report that their peers approve of drinking, and report fewer positive alcohol expectancies—as compared to European Americans (Kahler, Read, Wood, & Palfai, 2003; Randolph et al., 2009). The model of contingent consistency peer influence has attempted to explain ethnic differences in hazardous drinking among college students. This model posits that college drinking behavior is influenced by actual peer norms, perceived peer norms, and personal attitudes (Piane & Safer, 2008). In this model, African American students would be theorized to be at lower risk of engaging in hazardous alcohol use because they have lower actual and perceived peer norms for drinking, as well as less positive personal attitudes toward drinking. Given that African American college women are likely at lower risk of hazardous drinking and have less positive attitudes toward drinking, it is also possible that they may be less likely to use alcohol to cope with psychological distress, such as following sexual assault. However, this possibility has not been empirically tested.

In contrast, there is some evidence that African American college women may be more likely to engage in sexually risky behavior than European American women. For example, several studies have found that African American college women report having more sexual partners than European American women (Buhi, Marhefka, & Hoban, 2010; Espinosa-Hernández & Lefkowitz, 2009; Randolph et al., 2009), as well as more STIs and unintended pregnancies (Buhi et al., 2010). In addition, one study found that African American adolescents were more likely to report coping motives for sexual behavior than European Americans (Cooper et al., 1998). Thus, risky sexual behaviors may be more normative among African American college women as compared to women of other ethnicities; engaging in sexual behavior to cope with negative affect may also potentially be more normative, although it should be noted that research on sexual risk behavior among African American college women is quite limited.

Latina women

Studies of sexual assault among Latina college women find that they report fairly comparable rates of sexual assault as European American women (Kalof, 2000; Urquiza & Goodlin-Jones, 1994). Similarly, Latina college women report comparable rates of hazardous drinking as European American college women (Clements, 1999; Randolph et al., 2009). This second finding is particularly interesting given that many Latin cultures have sanctions against drinking among women (Wahl & Eitle, 2010).

One possible theoretical explanation for this discrepancy is the acculturative stress model. This model posits that adolescents/young adults in immigrant families are likely to engage in the acculturative process to a greater degree than other family members (e.g., parents and grandparents), resulting in discrepancies between familial and child values, as well as possible rejection of familial values by children/adolescents (Kim, 2009). Immigrant adolescents/young adults may also engage in problem behaviors as a way to rebel against familial authority and gain peer acceptance (Thai, Connell, & Tebes, 2010). In addition, greater exposure to the stressors associated with acculturation (e.g., discrimination, new social norms) increases adolescents/young adults’ risk of engaging in problem behaviors (Dion & Giordano, 1990; Potochnick & Perreira, 2010; Thai et al., 2010). Indeed, studies of Latino/a adolescents have documented that second- and third- generation adolescents (who are likely exposed to mainstream U.S. culture to a greater degree than first-generation adolescents) engage in more problem drinking than first-generation adolescents (Wahl & Eitle, 2010).

Although risky sexual behaviors have not been extensively studied among Latino/a college students, extant studies suggest that Latino/a college students do not differ from European Americans in their likelihood of engaging in risky sexual behavior or in their sexual attitudes (Espinosa-Hernández & Lefkowitz, 2009; Randolph et al., 2009). Again, this finding could be explained within an acculturative stress framework, given that attending college likely exposes Latina students extensively to mainstream U.S. culture and may enhance the acculturative process, as well as increase acculturative stress.

Asian American women

Studies of sexual assault among Asian American college women support that they are less likely than women of other ethnicities to report having been the victim of sexual assault (Archambeau et al., 2010; Nguyen et al., 2010; Urquiza & Goodlin-Jones, 1994). However, it is not clear if Asian American women are at lower risk of assault, or if they are less likely to disclose their experiences to others, including researchers. Indeed, Asian American college students are more likely than women of other ethnicities to hold attitudes that potentially inhibit disclosure, such as viewing rape victims as responsible for preventing the assault and endorsing rape myths (e.g., a woman can stop a rape if she wants to, women lie about being raped; Bryant-Davis et al., 2009; Lee, Pomeroy, Yoo, & Rheinboldt, 2005). However, it should also be noted that Asian American college women tend to report lower rates of hazardous drinking than women of other ethnicities, particularly European American and Latina students (Luczak, Wall, Shea, Byun, & Carr, 2001; Randolph et al., 2009; Wechsler et al., 2002), and thus may be at lower risk of experiencing an assault following impairment or incapacitation. Indeed, one recent study found that 6% of Asian American college women reported an incapacitated rape experience, as compared to 11% of European American women (Nguyen et al., 2010). However, their study did not examine sexual assaults obtained by force as well, and thus, the proportion of sexual assaults that involved incapacitation in these two groups of women could not be determined.

As previously noted, Asian American college students report lower rates of hazardous drinking than students of other ethnicities. There is also some evidence that Asian American college students may be less likely than others to use alcohol as an affect regulation strategy (Spada & Moneta, 2004). A number of potential protective factors reducing Asian American students’ risk of hazardous drinking have been identified. For one, Asian American women have been found to be more strongly influenced by peer drinking norms (which are likely lower than among some other ethnic groups) than students of other ethnicities (Thai et al., 2010; Vaughan, Corbin, & Fromme, 2009). Asian American college students also have less positive drinking expectancies overall than European American students (Han & Short, 2009). Finally, a sizable percentage of individuals of Asian descent (30–50%) lack an enzyme that plays a role in alcohol metabolism, resulting in unpleasant physiological symptoms following drinking, including flushing, tachycardia, and nausea, and thus a reduced likelihood of engaging in hazardous drinking (Hendershot et al., 2009; Luczak et al., 2001). However, it should be noted that acculturation factors and stress may play an important role in hazardous alcohol use among Asian Americans as well, with high levels of acculturative stress leading to more hazardous drinking among certain groups of Asian American students (Nguyen et al., 2010).

In contrast, sexual risk behavior remains largely unstudied among Asian American college students. One study found that, although there were no significant differences, Asian American college women fell between European American and African American women with regard to their mean number of sexual partners, and they were most similar to African American women in their condom use (there were no ethnic differences in condom use in this study; Randolph et al., 2009). Therefore, it is clear that further research is necessary in this area, including research examining these behaviors following sexual assault.

The Current Study

Thus, our current study sought to investigate sexual assault and health risk behaviors with a large, ethnically diverse sample of college women. Our first goal was to evaluate rates of adolescent and adult sexual assault among European American, African American, Latina, and Asian women. We also examined ethnic differences in rates of binge drinking and substance-induced impairment during the assault, given the possibility that African American women may be less likely than European American women to report being the victim of an incapacitated assault. In addition, we evaluated differences in two specific health risk behaviors: hazardous drinking and use of sexual behavior to regulate negative affect. Given the lack of known research examining the mechanisms for the association between sexual violence and health risk behaviors, we also evaluated whether depressive and anxiety symptoms mediated the association between having a sexual assault history and engaging in the risk behaviors. A final goal was to evaluate if there were differences in the strength of this mediated relationship among women of different ethnic backgrounds; that is, to evaluate whether moderated mediation existed.

We hypothesized that sexual victimization would be prevalent among women of all four ethnic groups, with only Asian American women reporting lower rates of sexual victimization than European American women. We also predicted that African American women would be less likely to report that they were binge drinking prior to the assault or that they were impaired during the assault than European American women. In addition, we hypothesized that the prevalence of hazardous drinking would vary among women of the four ethnic groups, with African American and Asian American women reporting lower rates of hazardous drinking than European American women. However, given the lack of prior research, no specific hypotheses were made regarding ethnic differences in frequency of use of sexual behavior to regulate negative affect. We further hypothesized that having a sexual assault history would be associated with a greater likelihood of engaging in the two health risk behaviors (hazardous drinking and use of sexual behavior to regulate negative affect). In addition, we predicted that experiencing higher levels of depressive and anxiety symptoms would mediate the association between having a sexual assault history and engaging in higher levels of the two health risk behaviors. Finally, given the preliminary nature of the research regarding predictors of these health risk behaviors among ethnic minority women, specific hypotheses regarding the nature of any moderated mediation relationships were not made.

Method

Participants

Participants were drawn from a sample of 1,744 women enrolled at one of the three U.S. southeastern universities who completed an online survey for course credit across two academic semesters. In the current study, the sample was restricted to the 1,620 women (92.9% of the sample) who indicated that their ethnicity was European American, Latina, African American, or Asian American. Excluded participants either left this item blank (n = 34) or indicated they were of other ethnicities (e.g., Native American, multiethnic; n = 90). These participants were excluded due to insufficient numbers of these women to conduct comparisons. Among participants, 1,171 (72.3%) described themselves as European American, 143 (8.8%) as African American, 144 (8.9%) as Latina, and 162 (10%) as Asian American. The mean age of participants was 20.5 years (SD = 3.8 years). Among participants, 46 (3.9%) European American women, 14 (9.8%) African American women, 42 (29.2%) Latina women, and 57 (35.2%) Asian American women reported that they were U.S. immigrants. A total of 324 (20%) of participants had been a victim of completed sexual assault (completed oral, vaginal, or anal penetration perpetrated by threat, force, or when the victim was incapable of consenting).

Measures

Ethnicity

Participant ethnicity was determined by self-report. Specifically, participants were asked to indicate what they considered themselves and were given a number of options. Participants were coded into specific ethnic groups based on their responses: European American (White/Caucasian/European American), African American (Black/African American/Caribbean Islander), Latina (Mexican/Mexican American/Latina), or Asian American (Asian/Pacific Islander/Asian American).

Sexual assault history

Two behaviorally specific items derived from the Sexual Experiences Survey (Koss & Gidycz, 1985) were administered to assess experiences of rape or completed sexual assault since the age of 14. The items assessed experiences of unwanted sex with a man or men (vaginal, oral, anal intercourse, or object penetration) obtained by force or threat, or that occurred when the individual was incapacitated or unconscious. Thus, women who experienced completed sexual assault (i.e., involving vaginal, oral, or anal penetration) due to threat, force, or that occurred when they were incapacitated were classified as sexual assault victims.

Binge drinking prior to the assault

Binge drinking among women who had experienced a sexual assault was assessed with the following item, “How much alcohol had you consumed at the time of the experience (1 drink = 1 pint of beer, 1 shot, or 1 small mixed drink)?” Participants who reported that they had consumed four or more standard drinks at the time of the assault were classified as having engaged in binge drinking (National Institute on Alcohol Abuse and Alcoholism, 2006).

Substance-induced impairment or incapacitation during the assault

In addition to asking about drinking prior to the assault, participants were asked if they were using any other drugs with the following item, “Were you using illegal substances at the time of the experience?” Participants were also asked to indicate the ways in which they were impaired during the experience and were given the following options: asleep, unconscious, had trouble walking, had trouble speaking, had trouble moving limbs. Like prior analyses (e.g., Littleton, Grills-Taquechel, & Axsom, 2009), victims were classified as impaired or incapacitated due to substances if they responded affirmatively to the substance-facilitated assault item on the sexual assault measure, reported that they were drinking alcohol or using drugs prior to the assault, and reported that they were impaired in some way during the assault. In this way, only victims who were impaired or incapacitated during the assault due to substances (as opposed to using substances but not impaired, or impaired not due to substance use) were classified as impaired or incapacitated due to substance use.

Depressive symptoms

The 20-item Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977) was administered to assess current depressive symptoms. For each item, individuals indicated how often they have felt that way in the past week on a 4-point Likert-type scale bounded by 0 (rarely or none of the time/less than one day) and 3 (most or all of the time/5-7 days). A sample item is, “I felt sad.” Scores can range from 0 to 80, and scores of 16 or above indicate significant depressive symptomatology (Radloff, 1977). The 4-week test–retest reliability was found to be .67, and Cronbach's α in community and patient samples has ranged from .84 to .90 (Radloff, 1977). Cronbach's α for this measure in the current study was .93. Supporting the validity of the measure, scores have been found to correlate moderately to strongly with other depression measures (self-report and clinician-administered) and to be sensitive to changes in symptomatology following treatment (Weissman, Sholomskas, Pottenger, Prusoff, & Locke, 1977).

Anxiety symptoms

The Four Dimensional Anxiety scale (FDAS; Bystritsky, Linn, & Ware, 1990) was administered to assess current anxiety symptoms. The FDAS is a 35-item, self-report measure of the affective, cognitive, behavioral, and physiological components of anxiety. For each item, individuals indicated how often they have felt in the described manner in the past week on a 5-point Likert-type scale bounded by 1 (not at all) and 5 (extremely). A sample item is, “Feeling nervous?” Cronbach's α of the measure in both a treatment-seeking and in a community sample was .92 (Bystritsky et al., 1990; Stoessel, Bystritsky, & Pasnau, 1995), and in the current study was .94. Supporting the validity of the measure, individuals seeking treatment for anxiety disorders scored significantly higher on the measure than individuals recruited from a medical setting (Stoessel et al., 1995).

Hazardous drinking

The 5-item version of the Alcohol Use Disorders Identification Test (AUDIT; Babor, Higgins-Biddle, Saunders, & Monteiro, 2001; Miles, Winstock, & Strang, 2001) was administered to assess hazardous drinking in the past year. A sample item is, “How often during the past year have you found that you were not able to stop drinking once you had started?” In the current study, the Cronbach's α of this measure was .77. Supporting the measure's validity, it was found to have a sensitivity of 79% and a specificity of 95% for alcohol dependence and harmful use as compared to a structured clinical interview (Piccinelli et al., 1997). A score of 5 or above on this measure indicates hazardous use (Miles et al., 2001). It should be noted that 193 (11.9%) women in our sample reported no past year alcohol use and did not complete this measure. These participants’ AUDIT score was coded as 0.

Sexual behavior as an affect regulation strategy

The 4-item Use of Sex to Reduce Negative Affect scale was administered to assess this construct (Orcutt et al., 2005). In the current study, participants were asked to indicate how often they had engaged in sexual behavior, including kissing and petting, in the last year for the listed reasons on a 5-point Likert-type scale anchored by 1 (almost never/never) and 5 (almost always/always). A sample item is “To feel better when feeling lonely.” Cronbach's α among a community recruited sample was .81 (Orcutt et al., 2005), and in the current study was .92. It should be noted that 105 (6.5%) participants did not report engaging in any sexual activity in the past year and did not complete this measure. These participants’ score on the measure was coded as a 4 (the minimum score).

Procedures

Data were collected using an online survey. Study procedures have been described previously (e.g., Littleton, Axsom, & Grills-Taquechel, 2009; Littleton, Grills-Taquechel, et al., 2009). Briefly, potential participants were recruited using announcements on Psychology Department research participant management websites at three large U.S. southeastern universities. Posted information stated that participants would be asked to complete a confidential survey about their negative sexual experiences, coping, and psychological health. Participants were excluded if they were male or under 18 years of age. Some form of student identification was collected to award course credit and to prevent duplicate participation (student identification information was removed from the data files prior to downloading).

Potential participants were given a brief description of the study and information about available counseling resources and were asked to provide their electronic consent. Behaviorally specific questions were administered to determine whether participants had experienced sexual assault. Participants also completed self-report measures of their current depressive and anxiety symptoms. In addition, participants completed measures of their past year drinking and use of sexual behavior as an affect regulation strategy. The study was approved by the institutional review boards of the three universities and followed the guidelines for ensuring the confidentiality of online data outlined by Reips (2002).

Analysis Plan

First, univariate analyses were conducted to compare the four ethnic groups’ scores on study variables. Between group differences were evaluated using Bonferroni-adjusted pairwise comparisons. A similar procedure was used to compare sexual assault victims of the four ethnic groups’ scores on study variables. Finally, sexual assault victims and nonvictims’ scores on study measures were compared utilizing t tests.

To conduct the moderated mediation analyses, we chose a two-step analysis plan. First, the four mediated models were evaluated in the overall sample. Specifically, anxiety and depressive symptoms were evaluated as mediators of the relationship between sexual assault and hazardous drinking and as mediators of the relationship between sexual assault and use of sexual behavior as an affect regulation strategy. To conduct mediation analyses, we utilized the bootstrap procedure recommended by Preacher and Hayes (2004) and Shrout and Bolger (2002). Bootstrapping entails creating a large number of bootstrap (pseudo) samples of randomly sampled observations from the data set that are drawn with replacement. The model paths are then estimated for each of these bootstrap samples (Shrout & Bolger, 2002). Results from the bootstrap analyses are then used to construct an estimate for the model paths and a confidence interval of these estimates. If the confidence interval does not contain zero, this supports a significant model path (Preacher & Hayes, 2004; Shrout & Bolger, 2002). Specifically, this procedure allows one to generate an estimate and confidence interval for the path from the predictor variable to the mediator and the path from the mediator to the outcome. In addition, an estimate and confidence interval for the overall mediated path, that is, the product of these first two paths, is estimated. Finally, the direct effect of the predictor on the outcome after controlling for the mediator is estimated. This procedure to evaluate mediation was chosen given evidence that traditional tests of mediation have low power (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002; Shrout & Bolger, 2002).

Next, the moderated mediation models were explored using the procedures recommended by Preacher, Rucker, and Hayes (2007). We evaluated whether the size of the mediated path differed significantly among ethnic groups. First, bootstrapping was used to generate estimates and confidence intervals of the estimate of the size of the mediated path—that is, the path from the mediators (depressive and anxiety symptoms) to risk behaviors (hazardous drinking and use of sexual behavior as an affect regulation strategy)—for women within each of the four ethnic groups (European American, African American, Latina, and Asian American). For these analyses, a significant confidence interval indicated significant mediation among women of that particular ethnic group. Next, we evaluated if the size of the mediated path—that is, the path from the mediators (depressive and anxiety symptoms) to risk behaviors (hazardous alcohol use and use of sexual behavior as an affect regulation strategy)—among each group of ethnic minority women differed significantly from the mediated path among ethnic majority (European American) women. Specifically, bootstrapping was used to generate estimates and confidence intervals of the estimate evaluating whether the size of the mediated path for women of each ethnic minority group differed from that of the reference group (European American women). For these analyses, a significant confidence interval indicated that the mediated path differed significantly for women of that particular ethnic group from that of the reference group (European American women) and thus the existence of moderated mediation. Because the majority of prior research has been conducted with European American women and because European American women made up the majority of the current sample, we chose them as the reference group.

All analyses were conducted using the MPlus program (version 6.1; Muthén & Muthén, 1998–2010). A total of 1,000 draws were used for all bootstrap analyses, considered to be the minimum number necessary for estimating the upper and lower bounds of the 95% confidence intervals (Edwards & Lambert, 2007). In addition, bias-corrected bootstrap values were estimated, which adjust for differences between the product from the full sample and the median of the products estimated from the bootstrap sample (Edwards & Lambert, 2007). Before conducting moderated mediation analyses, all continuous variables were mean centered. Finally, only unstandardized estimates are reported because standardized estimates are not interpretable in the analyses under consideration (A. F. Hayes, personal communication, December 8, 2010).

Results

Descriptive Analyses

Means and standard deviations for all study variables stratified by ethnicity are summarized in Table 1, and correlations among observed variables are reported in Table 2. Missing data were minimal overall, with all variables having less than 1% missing information. Participants with missing data on a particular variable were eliminated from analyses involving that variable. Overall, 1,605 of the 1,620 participants (99%) had no missing data.

Table 1.

Descriptive Statistics for Study Variables Stratified by Women's Ethnicity.

	European American, n = 1,171, M (SD)	African American, n = 143, M (SD)	Latina, n = 144, M (SD)	Asian American, n = 162, M (SD)
Depressive symptoms (CES-D)	16.78a (9.57)	16.97a (9.72)	17.66a (9.50)	17.58a (10.46)
Anxiety symptoms (FDAS)	63.8la (20.12)	61.00a (17.86)	69.35b (22.61)	64.12a,b (23.73)
Hazardous drinking (AUDIT)	4.34a (3.38)	2.43b (2.53)	3.52c (3.76)	2.73b,c (3.12)
Use of sex to regulate affect	8.65a (4.52)	8.90a (4.69)	9.33a (5.36)	8.02a (4.10)

Note. Means in each row that share subscripts do not differ significantly at overall p < .05, Bonferroni adjusted.

Table 2.

Correlations Among Observed Variables.

	1	2	3
1. Depressive symptoms
2. Anxiety symptoms	.72**
3. Hazardous drinking	.12*	.15*
4. Use of sex to regulate affect	.35*	.36*	.23*

Note.

^*p < .01.

^**p < .005.

There were no significant differences in depressive symptoms among women of the four ethnic groups, F(3, 1,610) = 0.61, p = .61. In contrast, there were significant differences in anxiety symptoms among women of the four ethnic groups, F(3, 1,612) = 4.29, p < .01, η² = .008. Follow-up Bonferroni-adjusted pairwise comparisons revealed that Latina women reported significantly more anxiety symptoms than European American (d = .27) and African American (d = .44) women. There were also significant differences among participants in their hazardous alcohol use, F(3,1616) = 23.37, p < .005, η² = .042. Follow-up Bonferroni-adjusted pairwise comparisons revealed that European American women reported significantly higher levels of hazardous drinking than women of the other ethnicities (African American, d = .58; Latina, d = .24; Asian American, d = .48). In addition, African American women reported significantly lower levels of hazardous drinking than Latina women, d = .35. Finally, there were no significant differences among women of the four ethnic groups in their reported use of sexual behavior to reduce negative affect, F(3, 1,616) = 2.14, p = .09.

Information about sexual assault among women of the four ethnic groups is summarized in Table 3. A chi-square analysis indicated that women of the four ethnic groups differed in their likelihood of being a sexual assault victim, χ²(3) = 9.59, p = .02. Follow-up Bonferroni-adjusted pairwise comparisons revealed that Asian American women were less likely to report being a victim of sexual assault than were European American women. Among sexual assault victims, there were differences in the proportion of women in the four ethnic groups who reported that they had engaged in binge drinking prior to the assault, χ²(3) = 24.56, p < .001. Follow-up Bonferroni-adjusted pairwise comparisons revealed that African American women were significantly less likely to report that they were binge drinking prior to the assault than European American or Latina women. There were also significant differences in the proportion of women of the four ethnic groups who reported substance-induced impairment or incapacitation at the time of the assault, χ²(3) = 14.09, p < .005. Follow-up Bonferroni-adjusted pairwise comparisons revealed that African American women were significantly less likely than European American women to report that they experienced substance-induced impairment or incapacitation at the time of the assault.

Table 3.

Sexual Assault Experiences Among Participants Stratified by Ethnicity.

	European American	African American	Latina	Asian American
Sexual assault victim	21.5%a (n = 252)	15.4%a,b (n = 22)	20.8%a,b (n = 30)	12.3%b (n = 20)
Binge drinking prior to assaulta	60.6%a (n = 149)	9.5%b (n = 2)	46.7%a (n = 14)	35.0%a,b (n = 7)
Impaired/incapacitated assaulta	65.9%a (n = 166)	27.3%b (n = 6)	60.0%a,b (n = 18)	50.0%a,b (n = 10)

Note. Means in each row that share subscripts do not differ significantly at overall p < .05, Bonferroni adjusted.

^aProportions refer to proportion of sexual assault victims.

The scores of sexual assault victims on the distress and risk behavior variables stratified by ethnicity are summarized in Table 4. There were no ethnic differences among sexual assault victims in their depressive, F(3, 319) = 0.14, p = .94, or anxiety symptoms, F(3, 320) = 1.06, p = .37, nor in their use of sexual behavior as an affect regulation strategy, F(3, 320) = 0.21, p = .89. In contrast, there were significant ethnic differences in hazardous alcohol use among sexual assault victims, F(3, 320) = 3.02, p = .03, η² = .027. Follow-up Bonferroni-adjusted pairwise comparisons revealed that African American victims reported less hazardous alcohol use than European American victims (d = .68).

Table 4.

Descriptive Statistics for Study Variables Among Sexual Assault Victims Stratified by Ethnicity.

	European American, n = 252, M (SD)	African American, n = 22, M (SD)	Latina, n = 30, M (SD)	Asian American, n = 20, M (SD)
Depressive symptoms (CES-D)	18.24a (10.68)	17.19a (12.31)	17.73a (10.17)	19.25a (14.12)
Anxiety symptoms (FDAS)	68.72a (24.24)	61.68a (20.54)	72.47a (26.89)	73.40a (31.16)
Hazardous drinking (AUDIT)	5.72a (3.98)	3.05b (2.72)	5.90a,b (5.38)	5.50a,b (3.97)
Use of sex to regulate affect	9.94a (4.84)	9.36a (6.08)	10.47a (6.16)	9.80a (3.82)

Note. Means in each row that share subscripts do not differ significantly at overall p < .05, Bonferroni adjusted.

Means and standard deviations of participants on all study variables stratified by sexual assault status are reported in Table 5. Independent samples t tests indicated that victims and nonvictims differed significantly on all study variables. Specifically, victims reported significantly more depressive symptoms, anxiety symptoms, hazardous drinking, and use of sexual behavior as an affect regulation strategy than nonvictims.

Table 5.

Descriptive Statistics for Study Variables Stratified by Sexual Assault Status.

	Sexual Assault Victim, n = 324, M (SD)	Nonvictim, n = 1,296, M (SD)	d
Depressive symptoms (CES-D)	18.19 (10.94)	16.64* (9.30)	.16
Anxiety symptoms (FDAS)	68.89 (24.74)	62.89** (19.27)	.29
Hazardous drinking (AUDIT)	5.54 (4.00)	3.54** (3.07)	.61
Use of sex to regulate affect	9.94 (4.99)	8.32** (4.34)	.36

Note.

^*p < .05.

^**p < .005.

Mediation Analyses

Bias-corrected bootstrap estimates and 95% confidence intervals for the mediation analyses are summarized in Table 6. Results of the analyses focused on hazardous alcohol use indicated significant mediation for both depressive and anxiety symptoms, with higher levels of both types of symptoms mediating the relationship between sexual assault and greater hazardous alcohol use. The analyses focused on sexual behavior as an affect regulation strategy were significant as well, with higher levels of both depressive and anxiety symptoms mediating the relationship between sexual assault and greater use of sexual behavior as an affect regulation strategy. In addition, the path from sexual assault to hazardous drinking and the path from sexual assault to use of sexual behavior as an affect regulation strategy were significant, supporting that depressive and anxiety symptoms were partial mediators of the sexual assault-risk behavior relationship. Both mediation models predicting hazardous drinking explained 7% of the variance in hazardous drinking. The mediation models predicting use of sexual behavior as an affect regulation strategy predicted 14% (anxiety symptoms) and 13% (depressive symptoms) of the variance in this outcome.

Table 6.

Bias-Corrected Bootstrap Estimates and 95% Confidence Intervals (CIs) of the Unstandardized Estimates for Mediation Analyses.

	Estimate	95% CI of Estimate
Sexual victimization/Depression/Hazardous alcohol use
Sexual victimization—depression	1.54	[0.35, 2.93]
Depression—hazardous alcohol use	0.04	[0.02, 0.06]
Sexual victimization—hazardous alcohol use (direct)	1.94	[1.47, 2.43]
Sexual victimization—hazardous alcohol use (indirect)	0.06	[0.02, 0.15]
Sexual victimization/Anxiety/Hazardous alcohol use
Sexual victimization—anxiety	5.99	[3.21, 9.07]
Anxiety—hazardous alcohol use	0.02	[0.01, 0.03]
Sexual victimization—hazardous alcohol use (direct)	1.88	[1.42, 2.38]
Sexual victimization—hazardous alcohol use (indirect)	0.12	[0.06, 0.23]
Sexual victimization/Depression/Sex to regulate affect
Sexual victimization—depression	1.54	[0.36, 2.94]
Depression—sex to regulate affect	0.16	[0.13, 0.18]
Sexual victimization—sex to regulate affect (direct)	1.38	[0.86, 1.96]
Sexual victimization—sex to regulate affect (indirect)	0.24	[0.06, 0.47]
Sexual victimization/Anxiety/Sex to regulate affect
Sexual victimization—anxiety	5.98	[3.20, 9.10]
Anxiety—sex to regulate affect	0.08	[0.06, 0.09]
Sexual victimization—sex to regulate affect (direct)	1.17	[0.65, 1.75]
Sexual victimization—sex to regulate affect (indirect)	0.45	[0.25, 0.69]

Note. Confidence interval (CI) ranges in bold are statistically significant.

Moderated Mediation Analyses

Bias-corrected bootstrap estimates and 95% confidence intervals of the separate mediated paths for women of the four ethnic groups are summarized in Table 7. In addition, bias-corrected bootstrap estimates and 95% confidence intervals for the moderated mediation paths are summarized in Table 7. These second estimates reflect a test of the significance of the estimate for the mediated path for the ethnic group under consideration as compared to the estimate of the mediated path for European American women (the reference group). Results of the moderated mediation analysis for hazardous alcohol use as the outcome and depression as the mediator indicated that the mediated path was significant for European American women but not for women of any of the three ethnic minority groups as indicated by the confidence intervals of the estimates (i.e., the confidence intervals of the estimates for ethnic minority women contained 0). In addition, the size of the mediated path for each group of ethnic minority women differed significantly from that of European Americans as indicated by the confidence intervals of those estimates (i.e., they did not contain 0), supporting the existence of significant moderated mediation. Thus, depression was a significant mediator of the sexual assault-hazardous alcohol use relationship for European American women only. Evaluating the path of the moderated mediation analysis for hazardous alcohol use with anxiety symptoms as the mediator indicated that the mediated path was significant for European American and Asian American women, but not for African American or Latina women. In addition, the size of the mediated path for each group of ethnic minority women differed significantly from that of European Americans, supporting the existence of significant moderated mediation. Thus, anxiety was a stronger mediator of the sexual assault-hazardous alcohol use relationship among European American than among all groups of ethnic minority women (including Asian American women).

Table 7.

Bias-Corrected Bootstrap Estimates for the Mediated and Moderated Paths in Each Ethnic Group.

	Mediation Estimate	95% CI of Mediation Estimate	Moderation Estimate	95% CI of Moderation Estimate
Depression/Hazardous alcohol use
European American	.07	[0.02, 0.17]	—	—
African American	.02	[–0.03, 0.13]	–1.81	[–2.27, –1.32]
Latina	.06	[–0.02, 0.28]	–0.85	[–1.45, –0.20]
Asian American	.01	[–0.07, 0.11]	–1.46	[–1.94, –0.93]
Anxiety/Hazardous alcohol use
European American	.14	[0.07, 0.26]	—	—
African American	–.03	[–0.20, 0.10]	–1.82	[–2.27, –1.33]
Latina	.10	[–0.07, 0.35]	–0.91	[–1.48, –0.21]
Asian American	.11	[0.01, 0.33]	–1.46	[–1.95, –0.93]
Depression/Sex to regulate affect
European American	.24	[0.06, 0.47]	—	—
African American	.28	[0.06, 0.66]	0.02	[–0.75, 0.81]
Latina	.33	[0.09, 0.82]	0.52	[–0.30, 1.29]
Asian American	.17	[0.04, 0.43]	–0.60	[–1.20, 0.11]
Anxiety/Sex to regulate affect
European American	.46	[0.26, 0.70]	—	—
African American	.50	[0.20, 1.00]	0.28	[–0.54, 1.17]
Latina	.55	[0.23, 1.03]	0.19	[–0.58, 0.98]
Asian American	.31	[0.13, 0.59]	–0.55	[–1.17, 0.14]

Note. Confidence interval (CI) ranges in bold are statistically significant. European American women served as the reference group in moderated mediation analyses.

Results of the moderated mediation analysis for use of sexual behavior as an affect regulation strategy with depressive symptoms as the mediator indicated that the mediated path was significant for women of all four ethnic groups. In addition, there were no significant differences in the size of the mediated path from depressive symptoms to use of sexual behavior as an affect regulation strategy for women of the three ethnic minority groups as compared to European American women. Thus, depressive symptoms functioned similarly as a mediator of the sexual assault—sexual behavior as an affect regulation strategy relationship among all groups of women. Results of the moderated mediation analysis for use of sexual behavior as an affect regulation strategy with anxiety symptoms as the mediator indicated that the mediated path was significant for women of all four ethnic groups. In addition, there were no significant differences in the size of the mediated path from anxiety symptoms to use of sexual behavior as an affect regulation strategy for women of the three ethnic minority groups as compared to European American women. Thus, anxiety symptoms functioned similarly as a mediator of the sexual assault-sexual behavior as an affect regulation strategy relationship among all groups of women.

Discussion

Our results documented that sexual assault is common among college women, with 12–21% of college women from the four ethnic groups reporting a history of completed sexual assault in adolescence or adulthood. With regard to differences in rates of reported sexual assault, results were generally consistent with prior studies of college women (e.g., Archambeau et al., 2010; Kalof, 2000). African American, Latina, and European American women reported similar rates of completed sexual assault, and Asian American women reported lower rates of completed sexual assault than European American women. In addition, as hypothesized, African American women were less likely than women of the other ethnicities to report binge drinking prior to the assault and to report that they were impaired as a result of substance use during the assault. Thus, although experiencing completed sexual assault is unfortunately a common experience among college women of different ethnicities, there may be important ethnic differences in the types of assaults experienced by women and in how women conceptualize these experiences. It is also possible that there may be ethnic differences in risk of assault, with Asian American women at lower risk of experiencing a sexual assault than women of other ethnicities. Interestingly, we did not find any evidence for the notion that Asian American women are at lower risk of experiencing alcohol-related assaults, with Asian American women reporting similar levels of binge drinking prior to the assault, and a similar proportion of victims reporting substance-related impairment as European American and Latina women. In addition, Asian American sexual assault victims reported similar levels of hazardous alcohol use as Latina and European American victims.

Also consistent with prior studies (e.g., Clements, 1999; Randolph et al., 2009; Wechsler et al., 2002), there were ethnic differences in hazardous alcohol use, with both African American women and Asian American women reporting lower levels of hazardous alcohol use than European American women, and African American women also reporting lower levels of hazardous alcohol use than Latina women. Thus, our results confirmed that ethnic minority college women, particularly African American and Asian American women, may be at lower risk of hazardous alcohol use than ethnic majority women. These differences in hazardous use are likely due to a combination of factors, including lower levels of peer use and less positive alcohol use expectancies among these groups of women (e.g., Han & Short, 2009; Kahler et al., 2003; Randolph et al., 2009).

In contrast, we found no ethnic differences in women's overall reported use of sexual activity as an affect regulation strategy. Although it is clear that this finding warrants replication, it suggests that there may be few differences in some risky sexual behaviors among college women of different ethnicities. Instead, overall norms regarding casual and other potentially risky sexual behaviors may be more influential in determining college women's sexual risk behavior. In addition, given that immigrant ethnic minority women are likely to experience high levels of acculturative stress, they may adhere to overall actual and perceived norms for these behaviors, and be influenced less by culture-specific norms or attitudes.

Sexual Victimization, Psychological Distress, and Risk Behaviors

Consistent with prior studies of sexual assault among college women (e.g., Zinzow et al., 2010), women in the current study who had experienced sexual assault reported significantly higher levels of both depressive and anxiety symptoms than nonvictims. Similarly, victims reported more hazardous drinking and risky sexual behavior (use of sexual behavior as an affect regulation strategy) than nonvictims (e.g., Brener et al., 1999; McCauley et al., 2009; Najdowski & Ullman, 2009a; Nguyen et al., 2010; Orcutt et al., 2005; Ullman, 2003).

With regard to the mediated relationships, both depressive and anxiety symptoms served as significant partial mediators of the association between having a history of sexual assault and engaging in hazardous drinking among European American women. This pattern supports the possibility that European American victims of sexual assault are engaging in these behaviors in part as a way of coping with symptoms of psychological distress. Indeed, prior studies support that both depression (Carr & Szymanski, 2011; Gonzalez, Reynolds, & Skewes, 2011; Harrell, Slane, & Klump, 2009) and anxiety (Ham, Zamboanga, Bacon, & Garcia, 2009) are associated with risk of hazardous drinking among college women, and that coping motives are important predictors of hazardous drinking among college women (Armeli, Todd, Conner, & Tennen, 2008; Gonzalez et al., 2011; Grant, Stewart, & Mohr, 2009; Ham et al., 2009).

However, depressive symptoms did not significantly mediate the relationship between having a sexual assault history and hazardous alcohol use among any of the groups of ethnic minority women we studied. Similarly, anxiety symptoms did not mediate the relationship between having a sexual assault history and hazardous alcohol use among Latina or African American women. Whereas anxiety symptoms mediated the relationship between sexual assault history and hazardous alcohol use among Asian American women, this relationship was significantly weaker than it was in European American women. Thus, increases in hazardous alcohol use among some ethnic minority women who have experienced sexual assault may not be as clearly related to psychological distress as it is among European American women. In addition, for African Americans, experiencing sexual assault may not be related to a greater risk of engaging in hazardous drinking overall. Thus, it is possible that hazardous alcohol use among ethnic minority women may be more strongly influenced by other factors, such as social motives or peer drinking norms (e.g., Thai et al., 2010; Vaughan et al., 2009), as opposed to being related to coping motives.

In contrast, depressive symptoms and anxiety symptoms significantly mediated the relationship between sexual assault history and use of sexual behavior as an affect regulation strategy among women of all four ethnic groups. This finding is consistent with prior studies supporting that both depression (Swanholm, Vosvick, & Chng, 2009) and anxiety (Kashdan, Collins, & Elhai, 2006) are associated with risky sexual behaviors such as engaging in casual sex, sex with multiple partners, and inconsistent condom use among college women. As stated previously, this finding could potentially reflect the fact that broader social norms may be more influential in affecting college women's sexual behaviors and motives. However, given the limited research on ethnic differences in sexual behaviors, it is clear that further research is needed. In addition, it is also possible that there may be ethnic differences in some sexual behaviors and motives, but not others.

Limitations

Limitations of the current study should be acknowledged. First, although the overall sample size was large and the sample was more ethnically diverse than many prior studies of college women, the number of individuals in each ethnic minority group, particularly the number of sexual assault victims in each group, was small, thus reducing our ability to evaluate differences in some of the variables of interest among women of the various ethnicities. In addition, only two health risk behaviors were examined in the current study. The cross-sectional nature of the current study also limits the extent to which causal inferences can be drawn. For example, it is possible that hazardous alcohol use may have led to elevated depressive and anxiety symptoms among women. As another example, it is possible that sexual assault victims were engaging in the health risk behaviors prior to the assault.

Additionally, several important variables that may explain the relationships found were not examined in the current study, most notably drinking expectancies and motives, immigration stress, and attitudes/beliefs regarding sexual assault. We also should note that women were classified into one of four broad ethnic groups, and thus, within ethnic group differences were not evaluated. Similarly, a broader conceptualization of ethnicity which includes such factors as level of identification with a particular group was not utilized (Helms, Jernigan, & Mascher, 2005; Karlsen, 2004). Finally, the current study involved a college sample and thus the findings may not generalize to other groups of ethnic majority and minority women.

Practice Implications

Bearing these limitations in mind, our findings present a number of possible implications for clinicians and other health care providers working with college women. First, the results support that health risk behaviors can represent responses to psychological distress, or attempts to cope with distress. Thus, it may be necessary to treat the individual's underlying psychological distress prior to, or concurrently with, intervening to reduce health risk behaviors, such as hazardous drinking. In addition, our results support the importance of routinely assessing for sexual violence history, given its association with distress and health risk behaviors, as well as its high prevalence among college women of all ethnicities.

Our findings also highlight the importance of attending to individual, social, and cultural factors when developing health risk prevention and intervention programs, even within specific settings such as universities. For example, some health risk behaviors, such as hazardous drinking, may be more common among some subgroups of college women than others. In addition, it is clear that college women of different ethnic and cultural backgrounds vary in the extent to which various factors, such as peer norms, influence their likelihood of engaging in health risk behaviors, as well as the extent to which their health risk behaviors relate to particular motives (e.g., coping motives). Prevention and intervention programs which are tailored to address risk factors and motives that are relevant to a particular subgroup are thus more likely to be effective than universal prevention programs. Similarly, clinicians and other health care providers should carefully assess risk and protective factors, as well as motives for engaging in risk behaviors, prior to developing intervention programs targeting those behaviors. At the same time, clinicians and health care providers should strive to remain informed regarding the cultural values of their clients/patients as well as factors (e.g., acculturative stress, ethnic identity, social norms) which may influence health risk behaviors.

Research Implications

Our results also have important implications for future research regarding sexual assault among women of different ethnicities. First, our findings indicate that there may be differences in the sexual assault experiences of college women of the four ethnicities we studied, including the frequency with which victims were impaired by substances during the assault as well as the likelihood that the assailant engaged in physical violence. These differences have potential implications for victims’ postassault experiences (e.g., self-blame, disclosure; Littleton, Grills-Taquechel, et al., 2009; Zinzow et al., 2010). Thus, future work should examine the implications of these differences in assault experiences of women of different ethnicities on women's postassault recovery and risk behaviors. As one example, victims of sexual assault who are impaired or incapacitated due to substance use have been found to engage in more self-blame than nonimpaired victims (Littleton, Grills-Taquechel, et al., 2009). Self-blame following sexual assault is associated with a host of negative outcomes including maladaptive coping, lowered self-worth, and PTSD symptomatology (Littleton et al., 2006; Najdowski & Ullman, 2009b). In contrast, victims of sexual assault obtained by force report that the assault resulted in more disruption in their life than victims who were incapacitated due to substance use (Abbey, BeShears, Clinton-Sherrod, & McAuslan, 2004). Our results also support that there may be differences in the frequency of engaging in certain health risk behaviors among women of different ethnicities, perhaps due to differences in peer norms, coping strategies, and motives for these health risk behaviors. Thus, future research should continue to evaluate health risk behaviors among women of different ethnicities, as well as factors that influence these behaviors, such as motives for engaging in these behaviors.

Furthermore, our results indicated that the pathways from sexual assault to some health risk behaviors may vary among women of different ethnicities. Therefore, researchers should be cautious in applying models of risk behaviors following sexual assault developed primarily with European American samples to ethnic minority women. There is also a clear need to identify factors that may affect health risk behaviors among ethnic minority victims of sexual assault, such as peer norms and social motives. Finally, there is a need to more comprehensively examine the sexual assault experiences of ethnic minority women, including exploration of the roles of psychological distress, social norms, immigration stress and discrimination, and sexual norms and attitudes in affecting women's health risk behaviors, conceptualization of their sexual assault experiences, and post-assault recovery.

Conclusion

Sexual assault is a significant problem affecting women of all ethnicities. Although the experience of sexual violence is all too common, there may be important ethnic differences in the sexual violence experiences of women of different ethnicities, including the settings where these incidents occur, the tactics used by the perpetrator, and victims’ behaviors prior to the assault. There may be ethnic differences in women's postassault experiences as well, including the extent to which sexual assault is associated with health risk behaviors and the extent to which these health risk behaviors are linked to efforts to cope with assault-related psychological distress. Thus, it is clear that social and cultural factors are important influences on women's risk for, conceptualization of, and reactions to sexual assault and should be considered when developing models to explain sexual assault risk and recovery. Inclusion of these factors in developing culturally inclusive models of risk and recovery will aid us in understanding the sexual violence experiences of all women and more effectively assisting victims in their recovery.