The Impact of Relationship Violence, HIV, and Ethnicity on

Adjustment in Women

Julie Axelrod
Department of Psychiatry and Biobehavioral Sciences University of California, Los Angeles (UCLA)
Hector F. Myers
Department of Psychology UCLA, and the Charles R. Drew University of Medicine and Science
Ramani S. Durvasula
Department of Psychiatry and Biobehavioral Sciences University of California, Los Angeles (UCLA)
Gail Elizabeth Wyatt
Department of Psychiatry and Biobehavioral Sciences University of California, Los Angeles (UCLA)
Michelle Cheng
Department of Psychiatry and Biobehavioral Sciences University of California, Los Angeles (UCLA)

ABSTRACT

This study examined how relationship violence, HIV, and ethnicity, moderated by social support, social undermining, and relationship satisfaction, influence psychological distress and dysfunction. A community sample of 415 African American, European American, and Latina women (140 HIV negative, 275 HIV positive) participated in the University of California, Los Angeles—Charles Drew Medical Center Women and Family Project. Of the 415, 27% ( n = 112 ; 79% HIV positive, 21% HIV negative) reported a history of relationship violence. Results indicated that HIV-positive women reported significantly more depressive symptoms, slightly more anxiety, but no differences on posttraumatic stress disorder (PTSD) symptoms than HIV-negative women. Women victimized by relationship violence also reported more depressive symptoms and anxiety and evidenced significantly more PTSD symptoms than nonabused women. Indeed, 58% of victimized women evidenced significant PTSD symptoms. Contrary to expectations, however, there were no significant ethnic differences on anxiety, but differences on depressive and PTSD symptoms emerged and were moderated by social undermining. Social support and dyadic satisfaction were not significant moderators of distress or dysfunction. Implications of these findings for clinical intervention and future research are presented.

Relationship violence includes physical battery of women within the context of physical or sexually abusive incidents perpetuated by a husband or partner in a current or most recent relationship. It is the leading cause of injury to American women, exceeding the number of rapes, muggings, auto accidents, and cancer deaths combined ( Dwyer, Smokowski, Bricout, & Wodarski, 1995 ). Some studies have identified the pattern of a high incidence of minority women in violent relationships ( Vasquez, 1998 ; Neff, Holamon, & Schulter, 1995 ; Lockhart, 1987 ), whereas other studies have attributed these differences to socioeconomic factors ( Kantor, Jasinski, & Aldarondo, 1994 ; Lockhart, 1985 ). HIV/AIDS represents another assault on women's health. In 1998, AIDS was the third leading cause of death in American women between the ages of 25—44 years ( Centers for Disease Control and Prevention [CDC], 1998 ). Physically devastating outcomes are common for HIV-positive women, who may also be at increased risk for violence should they (a) disclose their HIV status to their partners ( Gielen, O'Campo, Faden, & Eke, 1997 ) or (b) attempt to reduce their sexual risk.

Relationship Violence and HIV

For both HIV-positive women and women who are survivors of relationship violence, psychological manifestations of distress are common ( van Servellen et al., 1998 ; Hubbard, 1996 ; Campbell, Sullivan, & Davidson, 1995 ). Female victims of relationship violence commonly suffer from symptoms of depression and anxiety ( Russell, Lipov, Phillips, & White, 1989 ), which increase as physically aggressive acts by partners escalate ( Cascardi & O'Leary, 1992 ; Orava, McLeod, & Sharpe, 1996 ). In an attempt to cope, some strategies that women use, such as avoidance, may actually contribute to the severity of their depression ( Mitchell & Hodson, 1983 ). HIV-positive women are also vulnerable to psychiatric problems ( Hayashi & Fukunishi, 1997 ; Kaplan, Marks, & Mertens, 1997 ), with clinically significant levels of mood disorder resulting in serious impairment in physical, social, and role functioning ( McDaniel, Fowlie, Summerville, Farber, & Cohen-Cole, 1995 ).

Chronic abuse or life threatening illnesses can also increase these women's risk for posttraumatic stress disorder (PTSD; Botha, 1996 ; Kelly et al., 1998 ; Vitanza, Vogel, & Marshall, 1995 ). For women in violent relationships, having PTSD was found to severely affect their ability to accurately perceive health dangers, HIV risks, and health needs ( Molina & Basinait-Smith, 1998 ). This research, however, has not yet been conducted with multiethnic samples of women.

Moderators of the Effects of Relationship Violence and HIV

Although vulnerability to emotional anguish in HIV-positive women and victims of violence is magnified, social support can serve to ameliorate the distress ( Astin, Lawrence, & Foy, 1993 ; Boland, 1998 ; Peterson, Folkman, & Bakeman, 1996 ). However, victims of relationship violence tend to receive low levels of social support ( Barnett, Martinez, & Keyson, 1996 ; Forte, Franks, Forte, & Rigsby, 1996 ), and social undermining may further limit the benefits of the meager support that they do receive. Studies have documented the effects of social support for HIV-positive men, with the presence of support moderating the relationship between stress and depressed mood ( Peterson et al., 1996 ). Although this finding has not yet been duplicated for women, the myriad of emotional and instrumental needs of HIV-positive women are considerable ( Armistead & Forehand, 1995 ), increasing the likelihood that social support would serve as an important buffer against psychological distress as well.

The quality of an intimate relationship is also crucial in minimizing the impact of a terminal illness and relationship violence. The limited literature has revealed that marital quality, conflict management style, and thoughts and beliefs about marriage were more negative for both spouses when the husbands were physically aggressive ( Vivian & Malone, 1997 ). We know little, however, about the association between HIV infection and women's relationship quality, because the few existing studies included only gay male couples ( Norman, Kennedy, & Parish, 1998 ; Remien, Carballo-Dieguez, & Wagner, 1995 ). The association between physically abusive relationships and less condom use has been documented ( Molina & Basinait-Smith, 1998 ; Wingood & DiClemente, 1998 ). In fact, attempts to negotiate condom use by women can intensify the violence toward them and actually increase their risks for the unprotected sexual intercourse they were trying to avoid ( Kalichman, Williams, Cherry, & Belcher, 1998 ).

Ethnicity and Adjustment

Current research provides conflicting evidence about the role of ethnicity and culture in psychological adjustment. Some studies report higher rates of psychopathology in ethnic populations, whereas others described the opposite ( Blazer et al., 1998 ; Perl, Bagne, & Gurevich, 1989 ; Siegel, Aneshensel, Taub, Cantwell, & Driscoll, 1998 ). Still others have suggested that most of the observed differences between ethnic groups disappear when socioeconomic status (SES) is statistically controlled ( Biafora, 1995 ; Deal & Holt, 1998 ).

Studies of culture have focused on the role of collectivism, defined as an orientation toward the welfare of the community, as an important value in ethnic communities, particularly those with immigrants ( Gaines, 1997 ). These groups often rely heavily on extended family for emotional and instrumental support that is critical to well-being ( Warren, 1997 ). Marital relationship, another dimension of support, acts as a buffer against distress. Both African American and European American women were reportedly at greater risk for depression if they were separated or formerly married ( Jones-Webb & Snowden, 1993 ). Thus, the role of social support networks in moderating psychological distress may be even more magnified in ethnic communities.

Purpose of the Study

The purpose of this study was to investigate whether (a) HIV serostatus and victimization from relationship violence confer significant risk for psychological distress, when SES and substance abuse were controlled for; (b) these effects were moderated by ethnicity; and (c) availability of social supports, social undermining, and the quality of their primary relationship would moderate the risk for psychological distress attributable to HIV serostatus, relationship victimization, and ethnicity. Consistent with the literature, we expected that HIV-positive women should evidence more psychological distress and dysfunction than an HIV-negative cohort. Relationship victimization should increase risks for psychological distress, so that women who were both HIV-positive and victimized should report more distress than HIV-negative nonvictims. In the absence of any evidence to suggest otherwise, we did not expect that there would be any ethnic group differences in the relationships among HIV serostatus, relationship victimization, and psychological distress and dysfunction. The effects of HIV status and victimization were examined by controlling for demographic factors and substance abuse. The destructive nature of substance abuse potentially affects the outcomes, because it is associated with higher interpersonal risk and lower psychological adjustment ( Kalichman et al., 1998 ). Finally, we hypothesized that availability of social supports and greater relationship satisfaction should moderate the negative psychological effects of HIV seropositivity and relationship victimization, and that social undermining should exacerbate these negative effects.

Method

Procedure

This study examined data derived from the University of California, Los Angeles—Charles Drew Medical Center (UCLA—Drew) Women and Family Project, a longitudinal study examining the impact of HIV on several domains of women's lives, including stress, coping, sexuality, mental health, relationship dynamics, social support, and general health (see Wyatt & Chin, 1999 , for a more detailed description of sample recruitment and study procedures). HIV-positive women of African American, Latina, European American, Asian/Pacific Islander, and American Indian ethnic backgrounds were recruited from a variety of public and private clinics in Los Angeles County. HIV-negative participants were recruited through the Institute for Social Science Research at UCLA, which used stratified probability sampling and random-digit dialing. For the purposes of this study, only baseline data from the European American, African American, and Latina women were used given the limited samples of Asian/Pacific Islander and Native American women. A subsample of 415 women are compared, including 137 European American, 148 African American, and 130 Latinas; 275 were HIV positive and 140 were HIV negative.

Measures

Women were assessed with a comprehensive, semistructured interview, which included both original and standardized open- and closed-ended items designed to obtain both retrospective and current data. For the purposes of this study, a subset of variables was used from the baseline interview.

Relationship Violence.

The seven-item modified Conflict Tactics Scale ( Straus, 1979 ) was used to assess current or previous incidents of relationship violence. This scale identified women as victims of violence if they reported experiencing physical abuse in their most recent relationship (i.e., have been hit, kicked, punched, etc.). Women were identified as frequently victimized if they indicated they were physically abused "most of the time" or " always." Women were identified as infrequently victimized if they were physically abused "sometimes" or "rarely," and women were identified as never having been victimized if they reported no events of being physically abused.

The Wyatt Adult Sexual Abuse Scale, a subfile of the Wyatt Sex History Questionnaire ( Wyatt, 1985 ), is a 13-item scale that records the frequency and severity of incidents of sexual abuse as an adult ( Wyatt, Laurence, Vodounon, & Mickey, 1992 ). This scale identified women as experiencing relationship violence if they indicated that they had been sexually assaulted by a husband, partner, or ex-partner.

Psychological Distress and Dysfunction.

The Center for Epidemiologic Study, Depression scale (CES-D; Radloff, 1977 ), a widely used 20-item scale, was used to assess symptoms of depression. Participants were asked to rate the severity of depressive symptoms on a 4-point Likert-type scale that ranged from 1 ( rarely ) to 4 ( most of the time ). A total CES-D score was calculated with excellent reliability ( a = .92 ).

The Symptom Checklist 90—Anxiety subscale (SCL-90-Ax; Derogatis, Rickels, & Rock, 1976 ), a 15-item subscale of the SCL-90, measures symptoms of anxiety using a 5-point Likert-type scale that ranged from 0 ( not at all ) to 4 ( extremely ). A total SCL Anxiety score was calculated with excellent reliability ( a = .95 ).

Posttraumatic stress symptoms were assessed with a revised, short-form clinical checklist that documented exposure to traumatic events (e.g., violent assault or rape). The symptoms are organized into three clusters: four items that measure distressing reexperiencing (e.g., recurrent nightmares), seven items that assessed avoidance (e.g., avoiding similar situations), and six items that assessed increased arousal (e.g., get upset in similar situations). Women were classified as evidencing significant levels of PTSD symptoms if they reported at least one reexperiencing symptom, three avoidance symptoms, or two symptoms of increased arousal ( a = .98 ). However, a formal PTSD diagnosis could not be given because two additional diagnostic criteria (i.e., presence of symptoms for more than a month and experiencing significant distress or dysfunction because of these symptoms) were not assessed.

Psychosocial Covariates.

Substance abuse was assessed with two modules from the University of Michigan–Revised Version of the Composite International Diagnostic Inventory (UM-CIDI; Kessler et al., 1994 ). The UM-CIDI was developed and administered by trained nonclinicians in the recent National Comorbidity Study on a nationally representative sample of European Americans, African Americans, and Latinos and was reliable and valid ( Kessler et al., 1994 ). The two substance abuse modules used in this study assessed whether patterns of alcohol or illicit-drug use meet Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev., DSM—III—R ; American Psychiatric Association, 1987 ) criteria for abuse or dependence. Participants were classified as abusers of alcohol (yes/no) or illicit drugs (yes/no) if they reported any of the criteria symptoms on each of these subscales. Because very few women met criteria for alcohol abuse, only drug abuse was included in the analyses ( a = .88 ).

The Dyadic Adjustment Scale (DAS; Spanier, 1976 ) assessed relationship quality with a 32-item measure including four dimensions of intimate relationships: Satisfaction, Cohesion, Affectional Expression, and Consensus. This measure was administered only to those women in a committed relationship for 3 months or longer ( N = 229 ). The sum score on the Dyadic Satisfaction subscale was computed ( a = .78 ), which is the most sensitive global index of relationship quality ( Spanier, 1976 ).

Social Support was assessed with a short scale that asked respondents to rate the four most important members of the participant's social network on a 5-point Likert-type scale ( 1 = not at all to 5 = a great deal ). The measure assessed the degree to which each person in their network provided advice and emotional, affectional, and instrumental support, as well as the degree to which they were satisfied with the overall support received. The overall reliability for social support was high ( a = .86 ).

The Social Undermining Scale ( Vinokur & Van Ryn, 1993 ) was used to assess stresses caused by members of the social network using a three-item measure that asked respondents to rate, on a 5-point Likert-type scale, each of the four nominees in the network on the degree to which they "act angry," "criticize," or "make life difficult for them." A total social undermining score was high ( a = . 80 ).

Data Analysis

A total of 112 women (27% of the sample) met criteria as current or past victims of relationship violence and were categorized by the frequency of victimization into three groups: frequently abused (12%), infrequently abused (15%), and never abused (73%). Of the 275 HIV-positive women, 32% were victims of violence, compared with 21% of the HIV-negative women. African American women experienced violence in their relationships most (29%), followed by European Americans (26%) and Latinas (25%).

A series of 3 (ethnicity) × 2 (HIV serostatus) × 3 (frequently, infrequently, never abused) analyses of variance (ANOVAs) and chi-squares were run testing for group differences on demographic characteristics and substance abuse as potential covariates in analyses according to primary hypotheses. In addition, analyses were conducted examining whether social supports, social undermining, and dyadic satisfaction were significant moderators of psychological distress and dysfunction. Means and standard deviations on each of the variables of interest are presented in Table 1 .

Results

Group Differences in Demographic Characteristics

Three-way ANOVAs (HIV serostatus, ethnicity, and level of violence) on income yielded significant ethnicity, HIV, and Ethnicity × HIV effects. European American women reported significantly higher monthly household incomes than both African American and Latina women, F (2, 339) = 22.93 , p < .0001. HIV-infected women reported significantly lower monthly income than their HIV-negative counterparts, F (1, 339) = 61.70 , p < .0001. However, the significant interaction indicated that the income differential between HIV-negative and HIV-positive women was modified by ethnicity, F (2, 339) = 7.13 , p < .0009. The income difference between HIV-positives and HIV-negatives among the African American and European American women was greater than among the Latinas (i.e., a difference of $1,768 and $2,121 vs. $509 per month, respectively).

Three-way ANOVA (HIV serostatus, ethnicity, and level of violence) on education also yielded significant ethnicity, HIV, and Ethnicity × HIV effects, as well as a significant three-way interaction. European American and African American women reported more years of education than Latinas (13.9 and 12.9 years vs. 9.5 years respectively), F (2, 359) = 58.31 , p < .0001, and HIV-negative women reported significantly more education than HIV-positive women (12.8 vs. 11.4 years respectively), F (1, 359) = 17.66 , p < .0001.

However, HIV-positive African American and European American women who had not been victimized in violent relationships had slightly more education than those who had been victims of violence, F (2, 359) = 4.11 , p < .02. On the other hand, and contrary to expectations, HIV-negative African American and European American women who are victims of relationship violence had slightly higher education than those who were nonvictimized. For Latinas, the opposite was true: HIV-positive Latinas who were victims of violence had slightly higher education than nonvictims, whereas among HIV-negative Latinas, those who were victims had lower education than nonvictims.

The results of multiway chi-square tests on marital status also indicated that there was a significant Ethnicity × Violence interaction, χ ² (2, N = 415) = 5.98 , p < .05, as well as an HIV × Violence trend ( p < .08). Ethnic group means indicated that African American and Latina women who had been either frequently or infrequently victimized were significantly more likely to be married than those who had not been victimized. Among European Americans, there were no significant differences. On the other hand, among HIV-negative women, those who were not married were slightly less likely to have experienced relationship violence compared with those who were married. Among the HIV-positive women, however, there were no differences in the experience of relationship violence as a function of marital status.

Because of the observed differences in income, education, and marital status, these variables were treated as covariates in subsequent analyses.

Group Differences in Psychosocial Cofactors

A series of analyses testing for group differences also was conducted on drug abuse, dyadic satisfaction, social support, and social undermining. Multiway chi-square tests indicated significant ethnic group differences on substance abuse, χ ² (2, N = 415) = 5.9, p < .05 , with more African American women meeting criteria for drug abuse (22%) than European American (8%) and Latina (9%) women. As a result, this variable was also treated as a covariate in subsequent analyses.

Results of ANOVAs testing for differences in level of social support yielded significant differences as a function of violence. Women victimized by relationship violence reported receiving less social support than nonvictimized women, F (1, 358) = 12. 85 , p < .0004. Analyses testing for differences on social undermining yielded significant differences as a function of HIV status and violence. The means revealed that HIV-negative women reported experiencing more social undermining from the significant people in their lives than their HIV-positive counterparts, F (1, 356) = 5.6 , p < .02. However, and consistent with expectations, women victimized by relationship violence reported experiencing more social undermining than nonvictimized women, F (1, 356) = 8.76 , p < .003.

Finally, ANOVAs testing for differences in dyadic satisfaction yielded significant ethnic group and relationship violence differences. Women victimized by relationship violence reported less relationship satisfaction than nonvictimized women, F (1, 206) = 1. 21 , p < .002. In addition, African American women reported significantly less satisfaction with their current relationships than European American and Latina women, F (2, 206) = 3.79 , p < .02.

Group Differences in Psychological Distress and Dysfunction Depression.

In testing for the hypothesized differences in symptoms of depression, we ran an analysis of covariance (ANCOVA), testing for differences on the CES-D, with education, income, marital status, and substance abuse controlled. Results yielded significant HIV-status differences on depressive symptoms, with HIV-positive women reporting more symptoms than their HIV-negative counterparts, F (1, 395) = 13.69 , p < .0002. This difference was obtained after controlling for substance abuse ( p < .002), education ( p < .0001), income ( p < .0001), and dyadic satisfaction ( p < .04).

The ANCOVA also revealed that depressive symptoms differed across levels of victimization, F (2, 394) = 2.96 , p < .05, with those who were frequently victimized reporting significantly more symptoms of depression than those who were infrequently victimized or nonvictims. However, this effect was only evident when drug abuse ( p < .002) was controlled.

Furthermore, a series of analyses was conducted to examine the hypothesized moderating effects of social support, social undermining, and dyadic satisfaction with ethnicity, HIV serostatus, and victimization. A significant Ethnicity × Social Undermining effect on depression was obtained, F (2, 387) = 3.03 , p < .05, with African American women who reported more social undermining evidencing more depression. Contrary to expectations, however, neither social support nor dyadic satisfaction modified the relationships among ethnicity, HIV serostatus, relationship victimization, and depressive symptoms.

Anxiety.

An ANCOVA examined differences on the SCL-90 Ax, controlling for education, income, marital status, and substance abuse. In addition, analyses tested the hypothesized moderating effects of social support, social undermining, and dyadic satisfaction with ethnicity, HIV serostatus, and victimization on this outcome.

Results yielded a significant main effect for victimization, F (2, 387) = 3.83 , p < .02, with victims of frequent abuse reporting significantly more anxiety than victims of infrequent abuse and nonabused ( p < .03). However, the main effect of victimization on anxiety was moderated by social undermining, F (2, 387) = 2.99 , p < .05, with the frequently victimized women who also reported more social undermining evidencing more anxiety symptoms than their less-victimized and less-undermined counterparts.

PTSD Symptoms.

In testing for group differences in those who evidenced significant levels of PTSD symptoms, we conducted a series of stepwise logistic regressions. Education, income, and drug abuse treated as categorical variables (i.e., low education, low income, and drug abuse) were entered in the first step as covariates, with victimization (no abuse vs. infrequent abuse vs. frequent abuse), HIV serostatus (i.e., HIV positive vs. HIV negative) and ethnicity (i.e., African American vs. Latino vs. European American) entering in the next step. Results of these analyses indicated that frequent violent victimization was the best predictor of PTSD symptoms (adjusted odds ratio [OR] = .49, p < .01). This effect held true even after controlling for low income ( p < .05). A significant ethnicity effect was also observed ( OR = .361 , p < .03), with African American women at greater risk for PTSD symptoms than Latinas and European American women. However, this effect was only observed when a trend for social undermining ( p < .07) was taken into account.

Discussion

The present study provided an investigation of the impact of relationship violence, HIV serostatus, and ethnicity on symptoms of depression, anxiety, and PTSD in a large multiethnic community sample of women. The moderating effects of social support, social undermining, drug abuse, and relationship quality on psychological distress and dysfunction were also tested. Contrary to expectations, HIV serostatus and history of relationship violence did not exert interactive effects on psychological distress, but rather, each of these factors exerted significant independent effects.

These findings were consistent with the epidemiology of HIV in women in the United States, with HIV-positive African American and Latina women found to have diminished socioeconomic resources. Given that educational and economic resources facilitate access to numerous resources including health care, this pattern appears to place ethnic HIV-positive women at greater risk for poor outcomes. In addition, African American and Latina women who had a history of relationship violence were more likely to be married than European American women with similar histories. The contribution of cultural influences may in part explain this finding, with the cultural valuation of loyalty and solidarity to the family ( Gaines, 1997 ; Oropresa, 1996 ; Vasquez, 1998 ) proving to be a potent influence. This is likely to be pronounced among women who had recently immigrated or were undocumented residents. Among African American women, the reality of mate unavailability may influence their decisions to remain in a violent marriage rather than search for another hard-to-find partner ( James, Tucker, & Mitchell-Kernan, 1996 ). However, more research is needed to specifically address these findings.

As expected, HIV-positive women reported more symptoms of depression, even when low SES and substance abuse were controlled. These effects were not moderated by social support, social undermining, or dyadic satisfaction. Although support, undermining, or dyadic satisfaction did not moderate the HIV—depression relationship, differences between HIV-positive and HIV-negative women on these factors were obtained. HIV-positive women reported less undermining than HIV-negative women, suggesting that people in their networks may strive to be less subversive in consideration of their compromised health status. This may also represent a shift in the composition of the social network that occurred upon learning of HIV seropositivity. In general, however, this finding is inconsistent with other literature that suggests higher levels of psychological distress in HIV-infected men and women ( Kelly et al., 1998 ), although this literature has varied.

As expected, women with recurrent histories of relationship violence reported more symptoms of depression, anxiety, and PTSD. Regardless of whether violence was perpetrated by the current partner, women with these histories of relationship violence reported being less satisfied with their current or recent relationships. This suggests that past or present experience with relationship violence may influence women's willingness to withstand dissatisfaction or mistreatment. It was also noteworthy that the relationship between violence and anxiety was moderated by social undermining. Thus, abuse and undermining represent a double jeopardy for psychiatric risk in women. This has important implications for intervention programs with HIV-positive women and suggests that greater attention to sources of stress and conflict in both intimate and general social relationships is needed to target sources of relationship stress.

Contrary to expectations, symptoms of depression and PTSD were found to vary across ethnic groups, but only when perceived social undermining was also considered. This effect was especially evident among African American women, who reported more symptoms of depression and PTSD than the other groups, but only in the face of greater social undermining. This suggests that African American women in abusive relationships may be exposed to more social undermining or may be more psychologically vulnerable to subversive social networks than Latina and European American women who experience similar abuse. Several hypotheses could also be offered to explain why African American women reported the least satisfaction in their relationships compared with the other ethnic groups. The low-income African American women in this sample were less likely to be married, which may be attributable to low mate availability, greater relationship instability, or both. However, it is not possible to determine from the available data whether relationship dissatisfaction is a cause or a consequence of the lower marriage rate in this group. Nevertheless, more research on African American women and their relationships is needed to better understand these findings and to inform interventions that might improve the social resources of the psychosocially vulnerable women.

Although the findings from the present study are compelling, there are some limitations that must be noted. The Conflict Tactics Scale, used as the primary violence identifier, is a self-report scale and subject to reporting bias. Accounting for the number of violent events within a time frame, rather than relying on an individual's definition of frequency of abuse, would yield more objective data. Some groups (e.g., African American HIV-negative women with infrequent violence) yielded small sample sizes, which potentially limited the power to test for three-way interactions. Nonetheless, the independent and interactive effects obtained here highlight the need for further exploration of these issues with larger groups of women. Also, because the data were derived only from a baseline interview, differences across disease stage were not tested. The degree of the relationships among violence, HIV serostatus, and ethnicity throughout the course of illness bears further exploration and will be examined in future analyses of these data.

References

American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders ((3rd ed., rev). Washington,DC: Author.)
Armistead, L. & Forehand, R. (1995). For whom the bell tolls: Parenting decisions and challenges faced by mothers who are HIV-seropositive. Clinical Psychology: Science andPractice,, 2, 239-250.
Astin, M. C., Lawrence, K. J. & Foy, D. W. (1993). Posttraumatic stress disorder among battered women: Risk and resiliency factors. Violence and Victims,, 8, 17-28.
Barnett, O. W., Martinez, T. E. & Keyson, M. (1996). The relationship between violence, social support, and self-blame in battered women. Journal of Interpersonal Violence,, 11, 221-233.
Biafora, F. (1995). Cross-cultural perspective on illness and wellness: Implications for depression. Journal of Social Distress and theHomeless,, 4, 105-129.
Blazer, D. G., Landerman, L. R., Hays, J. C., Simonsick, E. M. & Saunders, W. B. (1998). Symptoms of depression among community-dwelling elderly African-American and White older adults. Psychological Medicine,, 28, 1311-1320.
Boland, R. J. (1998). Depression in HIV-infected persons. Psychosomatics,, 39, 82-83.
Botha, K. F. H. (1996). Posttraumatic stress disorder and illness behaviour in HIV+ patients. Psychological Reports,, 79 (3, Pt. 1), 843-845.
Campbell, R., Sullivan, C. M. & Davidson, W.S. (1995). Women who use domestic violence shelters: Changes in depression over time. Psychology of Women Quarterly,, 19, 237-255.
Cascardi, M. & O'Leary, K. D. (1992). Depressive symptomatology, self-esteem, and self-blame in battered women. Journal of Family Violence,, 7, 249-259.
Centers for Disease Control Prevention. (1998). Critical need to pay attention to HIV prevention for women: Minority and young women bear greatest burden [Fact sheet of the National Center for HIV, STD, and TB Prevention].(Atlanta, GA: Author)
Deal, L. W. & Holt, V. L. (1998). Young maternal age and depressive symptoms: Results from the 1988 National Maternal and Infant Health Survey. American Journal of Public Health,, 88, 266-270.
Derogatis, L. R., Rickels, K. & Rock, A. F. (1976). The SCL-90 and the MMPI: A step in the validation of a new self-report scale. British Journal of Psychiatry,, 128, 280-289.
Dwyer, D. C., Smokowski, P. R., Bricout, J. C. & Wodarski, J. S. (1995). Domestic violence research: Theoretical and practice implications for social work. Clinical Social Work Journal,, 23, 185-198.
Forte, J. A., Franks, D. D., Forte, J. A. & Rigsby, D. (1996). Asymmetrical role-taking: Comparing battered and nonbattered women. Social Work,, 41, 59-73.
Gaines, S. O. (1997). Culture, ethnicity, and personal relationship processes. (New York: Routledge)
Gielen, A. C., O'Campo, P., Faden, R. R. & Eke, A. (1997). Women's disclosure of HIV status: Experiences of mistreatment and violence in an urban setting. Women and Health,, 25, 19-31.
Hayashi, M. & Fukunishi, I. (1997). Lack of medical support in HIV infection and unstable mood states. Psychological Reports,, 81, 635-639.
Hubbard, M. G. (1996). Depression in abused women.(In C. C. Kroeger & J. R. Beck(Eds.), Women, abuse, and the bible: How scripture can be used to hurt or to heal (pp. 131—147). Grand Rapids, MI: Baker Books.)
James, A. D., Tucker, M. B. & Mitchell-Kernan, C. (1996). Marital attitudes, perceived mate availability, and subjective well-being among partnered African American men and women. Journal of Black Psychology,, 22, 20-36.
Jones-Webb, R. J. & Snowden, L. R. (1993). Symptoms of depression among Blacks and Whites. American Journal of Public Health,, 83, 240-244.
Kalichman, S. C., Williams, E. A., Cherry, C. H. & Belcher, L. (1998). Sexual coercion, domestic violence, and negotiating condom use among low-income African American women. Journal of Women's Health,, 7, 371-378.
Kantor, G. K., Jasinski, J. L. & Aldarondo, E. (1994). Sociocultural status and incidence of marital violence in Hispanic families [Special issue: Violence against women of color]. Violence and Victims,, 9, 207-222.
Kaplan, M. S., Marks, G. & Mertens, S. B. (1997). Distress and coping among women with HIV infection: Preliminary findings from a multiethnic sample. American Journal of Orthopsychiatry,, 67, 80-91.
Kelly, B., Raphael, B., Judd, F., Kernut, G., Burnett, P. & Burrows, G. (1998). Posttraumatic stress disorder in response to HIV infection. General Hospital Psychiatry,, 20, 345-352.
Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C. B., Hughes, M., Eshleman, F., Wittchen, H.U. & Kendler, K. S. (1994). Lifetime and twelve-month prevalence DSM—III—R psychiatric disorders in the United States: Results from the National Co-morbidity Study. Archives of General Psychiatry,, 51, 8-19.
Lockhart, L. L. (1985). Methodological issues in comparative racial analyses: The case of wife abuse. Social Work Research and Abstracts,, 21, 35-41.
Lockhart, L. L. (1987). A reexamination of the effects of race and social class on the incidence of marital violence: A search for reliable differences. Journal of Marriage and the Family,, 49, 603-610.
McDaniel, J. S., Fowlie, E., Summerville, M. B., Farber, E. W. & Cohen-Cole, S. A. (1995). An assessment of rates of psychiatric morbidity and functioning in HIV disease. General Hospital Psychiatry,, 17, 346-352.
Mitchel, R. E. & Hodson, C. A. (1983). Coping with domestic violence: Social support and psychological health among battered women: An ecological perspective. American Journal of CommunityPsychology,, 11, 629-654.
Molina, L. D. & Basinait-Smith, C. (1998). Revisiting the intersection between domestic abuse and HIV risk. American Journal of Public Health,, 88, 1267-1268.
Neff, J. A., Holamon, B. & Schulter, T. D. (1995). Spousal violence among Anglos, Blacks and Mexican Americans: The role of demographic variables, psychosocial predictors, and alcohol consumption. Journal of Family Violence,, 10, 1-21.
Norman, L. R., Kennedy, M. & Parish, K. (1998). Close relationships and safer sex among HIV-infected men with haemophilia. AIDS Care,, 10, 339-354.
Orava, T. A., McLeod, P. J. & Sharpe, D. (1996). Perceptions of control, depressive symptomatology and self-esteem of women in transition from abusive relationships. Journal of Family Violence,, 11, 167-186.
Oropresa, R. S. (1996). Normative beliefs about marriage and cohabitation: A comparison of non-Latino Whites, Mexican Americans, and Puerto Ricans. Journal of Marriage and the Family,, 58, 49-62.
Perl, E., Bagne, C. A. & Gurevich, D. (1989). Prevalence of depressive symptoms in different races. American Journal of Psychiatry,, 146, 408
Peterson, J. L, Folkman, S. & Bakeman, R. R. (1996). Stress, coping, HIV status, psychosocial resources, and depressive mood in African American gay, bisexual, and heterosexual men. American Journal of CommunityPsychology,, 24, 461-487.
Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement,, 1, 385-401.
Remien, R. H., Carballo-Dieguez, A. & Wagner, G. (1995). Intimacy and sexual risk behaviour in serodiscordant male couples [Special section: AIDS in the Americas]. AIDS Care,, 7, 429-438.
Russell, M. N., Lipov, E., Phillips, N. & White, B. (1989). Psychological profiles of violent and nonviolent maritally distressed couples. Psychotherapy,, 26, 81-87.
Siegel, J. M., Aneshensel, C. S., Taub, B., Cantwell, D. P. & Driscoll, A. K. (1998). Adolescent depressed mood in a multiethnic sample. Journal of Youth and Adolescence,, 27, 413-427.
Spanier, G. B. (1976). Measuring dyadic adjustment: New scales for assessing the quality of marriage and similar dyads. Journal of Marriage and the Family,, 38, 15-28.
Straus, M. A. (1979). Measuring intrafamily conflict and violence: The Conflict Tactics (CT) scales. Journal of Marriage and the Family,, 4, 75-88.
van Servellen, G., Sarna, L., Nyamathi, A., Padilla, G., Brecht, M-L. & Jablonski, K. J. (1998). Emotional distress in women with symptomatic HIV disease. Issues in Mental Health Nursing,, 19, 173-189.
Vasquez, M. J. T. (1998). Latinos and violence: Mental health implications and strategies for clinicians. Cultural Diversity and Mental Health,, 4, 319-334.
Vinokur, A. & Van Ryn, M. (1993). Social support and undermining in close relationships: Their independent effects on the mental health of unemployed persons. Journal of Personality and SocialPsychology,, 65, 350-359.
Vitanza, S., Vogel, L. C. M. & Marshall, L. L. (1995). Distress and symptoms of posttraumatic stress disorder in abused women. Violence and Victims,, 10, 23-34.
Vivian, D. & Malone, J. (1997). Relationship factors and depressive symptomatology associated with mild and severe husband-to-wife physical aggression. Violence and Victims,, 12, 3-18.
Warren, B. J. (1997). Depression, stressful life events, social support, and self-esteem in middle class African American women. Archives of Psychiatric Nursing,, 11, 107-117.
Wingood, G. M. & DiClemente, R. J. (1998). Wingood and DiClemente respond: Unanswered questions remain. American Journal of Public Health,, 88, 1267-1268.
Wyatt, G. E. (1985). The Wyatt Sex History Questionnaire. (Los Angeles: University of California, LosAngeles, Department of Psychiatry and Biobehavioral Sciences.)
Wyatt, G. E. & Chin, D. (1999). HIV and ethnic minority women, families, and communities: An overview. Cultural Diversity and Ethnic MinorityPsychology,, 5, 179-182.
Wyatt, G. E., Laurence, J., Vodounon, A. & Mickey, M. R. (1992). The Wyatt Sex History Questionnaire: A structured interview for female sexual history taking. Journal of Child Sexual Abuse,, 1, 51-68.

Preparation of this article was supported by National Institute of Health Grant MH54965. We thank the Women and Family Project staff and Angelika Appleton for manuscript preparation.
Correspondence may be addressed to Julie Axelrod, 300 UCLA Medical Plaza, Room 1512, Los Angeles, California, 90095-1759.

Table 1. Mean (and Standard Deviation) Demographic and Psychosocial Characteristics by HIV Serostatus and Level of Relationship Violence

Email: