Risky sexual behavior in low-income African American women: the impact of sexual health variables.

Within the United States, African Americans are disproportionately infected with HIV and other sexually transmitted diseases (STDs) (Wohl et al., 1998). For example, in the year 2000, African Americans made up 76% of gonorrhea cases, 71% of syphilis cases, 52% of HIV infection cases, and 38% of individuals diagnosed with AIDS, despite constituting only 12% of the U.S. population (Centers for Disease Control and Prevention, 2001). These epidemiologic data highlight a clear, critical need for research aimed at improving the sexual health of this community.

Measures of Risky Sexual Behavior

Multiple concurrent sexual partnerships. Multiple concurrent relationships are a major heterosexual HIV risk indicator for both men and women. Mathematical models demonstrated that concurrent partnerships can amplify nascent HIV epidemics by as much as tenfold (Morris & Kretzschmar, 1995, 1997; Watts & May, 1992), especially in high-prevalence communities (Finer, Darroch, & Singh, 1999) such as Blacks (Adimora & Schoenbach, 2002). Similar to men who have sex with men, there is evidence that multiple partnerships in African American populations (among men and possibly women) are more common than in other racial/ethnic groups (Adimora & Schoenbach; Adimora, Schoenbach, Bonas, et al., 2002; Bakken & Winter, 2002; Catania et al., 1992; Dolcini, Coates, Catania, Kegeles, & Hauck, 1995; Finer, Darroch, & Singh; Ford, Sohn, & Lepkowski, 2002; Peterson et al., 1992; Smith, 1991; Staples & Johnson, 1993; Weinberg & Williams, 1988). Since studies have shown that multiple partnerships are more likely to occur in unmarried relationships (Catania et al.), the higher rates of multiple partnerships in African American samples may be an artifact of the lower marriage rates among African American than in Caucasian and Hispanic samples (Allen & Olson, 2001; Graves & Hines, 1997; Roempke, Graefe, & Lichter, 2002).

In spite of its importance in the HIV-prevention literature, there appears to be little consensus as to terminology for the concept of multiple concurrent sexual partnerships. We found 12 different terms used to describe this concept: non-monogamy, non-monogamous sexual behavior, relative monogamy, non-exclusive relationships, extramarital sex/coitus, concurrent sexual partnerships, multiple partners, multiple sexual partners, overlapping relationships, non-mutually monogamous unions, concurrency/non-exclusivity, and individual concurrency and partnership concurrency (Adimora & Schoenbach, 2002; Bakken & Winter, 2002; Finer et al., 1999; Hines, Snowden, & Graves, 1998; Klitsch, 2002; Manhart, Aral, Holmes, & Foxman, 2002; Norris & Ford, 1999; Santelli, Lowry, Brener, & Robin, 2000; Shain et al., 2002; Short et al., 2003; Smith, 1991).

Consistent condom use. Evidence from HIV-discordant couples shows consistent condom use to be a highly effective HIV/STD prevention strategy, while inconsistent condom use offers insufficient protection (DeVincenzi, 1994; Feldblum, 1991; Miner, Robinson, Hoffman, Albright, & Bockting, 2002; O'Leary & Wingood, 2000; Pulerwitz, Amaro, DeJong, Gortmaker, & Rudd, 2002). Thus, consistent condom use is regarded as the primary outcome measure for HIV risk reduction and is recommended as the primary outcome measure for evaluating HIV-prevention program efficacy (O'Leary & Wingood).

The Sexual Health Model

There is a recognized need to address the sexual context of HIV risk in prevention efforts (Abraham & Sheeran, 1994; Boldero, Moore, & Rosenthal, 1992; Ehrhardt, Yingling, Zawadzki, & Martinez-Ramirez, 1992; Kalichman, 1998). A variety of sexuality variables such as acceptance and comfort with sexuality, sex guilt, and sexual self-esteem have been found to be associated with safer sex behaviors (Abraham & Sheeran; Boldero et al). To provide a conceptual framework for future research in this important area of HIV prevention, we recently developed the Sexual Health Model (Robinson, Bockting, Rosser, Miner, & Coleman, 2002). The model is derived from a sexological approach to education, and as such, is rooted in the belief that if one is more sexually literate, comfortable, and competent, one is also more likely to develop successful, long-term strategies for reducing HIV risk in the context of one's sexual behavior and relationships. This central hypothesis has just begun to be explored empirically in the design and evaluation of several HIV-prevention programs (Robinson, Uhl, et al., 2002; Rosser, Bockting, et al., 2002; Rosser, Ross, et al., 2002).

This study aimed to extend our testing of the Sexual Health Model as it applies to low-income African American women. We specifically focused on understanding the behaviors (e.g., inconsistent condom use, multiple concurrent partner partners) that mediate risk for HIV and other STDs. We explored these behaviors through the following variables in the Sexual Health Model: challenges or barriers to healthy sexuality, sexual anatomy and functioning, positive sexuality, sexual health care and safer sex, and cultural and sexual identity.

Challenges or barriers to healthy sexuality. Several variables included in the Sexual Health Model have been shown to be related to aspects of risky sexual behavior, with the strongest evidence supporting the relationship between challenges/barriers and HIV risk. Sexual and physical abuse, chemical dependency, and prostitution appear most frequently in the literature as being positively related to risky sexual behavior (Goodman & Fallot, 1998; Johnson & Harlow, 1996; Koenig & Clark, 2004; O'Leary & Wingood, 2000; Quina, Morokoff, Harlow, & Zurbriggen, 2004; Ross et al., 2001; Wingood & DiClemente, 1997; Wyatt et al., 2002).

Sexual anatomy and functioning. In contrast to the wealth of literature relating challenges/barrier to HIV risk, we were able to find only one relevant study exploring sexual functioning and risky behaviors. Choi, Catania, and Dolcini (1994) found that African American men and women with sexual problems were more likely to report having extramarital affairs. Given that sexual health assumes a basic knowledge, understanding, and acceptance of one's sexual anatomy, sexual response, and sexual functioning, as well as freedom from sexual dysfunction and other sexual problems (Robinson, Bockting, et al., 2002), we are interested in examining the relationship between this important aspect of sexual health and risky sexual behaviors.

Positive sexuality. Positive sexuality includes appropriate experimentation, affirming sensuality, attaining sexual competence through the ability to get and give sexual pleasure, and setting sexual boundaries based on what one prefers, as well as what one knows is safe and responsible (Robinson, Bockting, et al., 2002). One important measure of positive sexuality, sexual assertiveness, is a concept that has been shown to be associated with safer sex (Quina, Harlow, Morokoff, & Burkholder, 2000).

Sexual health care, pregnancy, and safer sex. As a component of the Sexual Health Model, physical health includes, but is not limited to, practicing safer sex behaviors. Attitudes toward one's health, health care, and safer sex also impact HIV risk. More positive attitudes toward condoms have significantly predicted their use among women (Valdiserri, Arena, Proctor, & Bonati, 1989, as cited in Kalichman, 1998) and a perception that one is at risk for HIV infection makes it more likely that one will take steps to reduce risk behaviors (Hines, Snowden, & Graves, 1998; Kalichman).

Desiring and having children is an important aspect of sexual health care that impacts risky sexual behavior in complicated ways. On the one hand, a desire for pregnancy decreases the likelihood that an individual will use a condom (Cabral, Galavotti, Armstrong, Morrow, & Fogarty, 2001; O'Leary & Wingood, 2000). On the other hand, pregnancy may increase other HIV-prevention behaviors (e.g., reduction in drug injection rates and increased likelihood of entering a drug treatment program; Deren et al., 1995).

Culture and sexual identity. One component of the Sexual Health Model--cultural and sexual identity--deserves a more extended discussion. Our inclusion of this component in the model was driven, in part, by our prior HIV-prevention work in a number of ethnic and sexuality minority communities, including African American men and women (Robinson, Bockting, et al., 2002; Robinson, Harrell, Smith, Bockting, & Malcolm, 1994; Robinson, Miner, et al., 2001); men who have sex with men (Rosser, Bockting, et al., 2002); Latino men who have sex with men (Rosser, Miner, et al., 2002); transgendered individuals (Bockting, 1998); bisexual women (Bockting & Robinson, 1999); and Latino HIV-prevention workers (Rosser, Mazin, & Coleman, 2002). For example, in our HIV-prevention work with African American women, participants talked about their parents' desire for them to remain sexually chaste for as long as possible and the resulting paucity of sex education, including safer sex information, that was given to them. Others have suggested that cultural norms and practices can influence sexual attitudes and behaviors by defining normative behavior (Kline, Kline, & Oken, 1992; Newcomb et al., 1998; Sabogal, Perez-Stable, Otero-Sabogal, & Hiatt, 1995).

While acculturation has been studied in many immigrant groups such as Asian Americans and Latinos, it has only recently been applied to African Americans (Hines et al., 1998; Landrine & Klonoff, 1996). We found only one study that specifically examined the relationship between acculturation and risky sexual behavior in an African American sample (Hines et al.). Although these researchers found no linear relationship between acculturation and any of their four safer sex variables, they did find that highly acculturated African American women who were heavy drinkers were more likely to have multiple partners.

In this study, we applied Landrine and Klonoff's (1996) concept of acculturation in our investigation of the relationship of cultural and sexuality variables to risky sexual behavior. Landrine and Klonoff define acculturation as a continuum (ranging from "traditional" to "acculturated"), measuring the extent to which ethnic-cultural minorities participate in the traditions, values, beliefs, and practices of their own culture versus those of the dominant White society. The effects of cultural attitudes and practices on health behavior may vary depending on the degree of acculturation. Acculturation has been shown to be related to a variety of health attitudes and behaviors, including alcohol consumption (Caetano & Kaskutas, 1995), HIV/STD attitudes (Robinson, Freske, Scheltema, & Chu Yang Heu, 1999), and sexual practices such as oral sex (Newcomb et al., 1998), number of sexual partners (Sabogal et al., 1995), and sexual intercourse and condom use (Ford & Norris, 1993).

Potential moderator variables. Various sociodemographic variables have been reported to be positively associated with multiple concurrent partnerships or condom use, including being unmarried, being younger, and not attending church. Associations have also been reported between these HIV risk factors and education level, but reports are conflicting as to their direction (Adimora et al., 2002; Catania et al., 1992; Choi et al., 1994; Finer et al., 1999; Hines & Graves, 1998; O'Leary & Wingood, 2000; Peterson et al., 1992; Seidman, Mosher, & Aral, 1992; Upchurch, Kusunoki, Simon, & Doty, 2003).

Hypotheses

In this study, we report the results of logistic regression analyses investigating the relationship between measures of five components of the Sexual Health Model (defined above) and five AIDS-related risky sexual behaviors. These dependent measures included three measures assessing multiple concurrent partnerships (self-report, belief about partner, belief about monogamy in relationship), consistent condom use, and a composite measure of risky sexual behavior. We also examined the relationship of several potential moderator variables (age, education, marital status, employment, and religious service attendance) to these same risky behaviors. While not of primary interest, these moderator variables are of importance in explaining variance. Our hypotheses were as follows:

1. Sexual Health Model and risky sexual behavior: Lack of sexual difficulties/dysfunction, positive sexual assertiveness/self efficacy, positive attitudes toward condoms, acknowledgement of HIV risk in one's environment, and a lack of desire or intention for pregnancy are negatively related to risky sexual behaviors. In contrast, potential challenges/barriers to healthy sexuality, including heavy alcohol and drug use, IV drug use, a history of domestic abuse, sexual abuse, prostitution, or STDs, are positively related to risky sexual behaviors.

2. Acculturation and risky sexual behavior: One's level of acculturation is associated with measures of sexual risk for HIV/STD, in that more highly acculturated African American women who are heavy drinkers and who use drugs are more likely to engage in risky sexual behaviors such as having multiple partners (Hines et al., 1998).

3. Potential moderator variables and risky sexual behavior: Higher age and education, marriage, regular employment, and religious service attendance are negatively related to risky sexual behaviors.

METHOD

The institutional review boards at the University of Minnesota (#9608S11642) and the Centers for Disease Control and Prevention (#1742) approved this study.

Sample

The sample was drawn from participants in the Women's Initiative for Sexual Health (WISH), a sexual health HIV-prevention intervention targeting at-risk African American women. For this community-based intervention study, low-income, chemically dependent women were heavily recruited, although all women who wanted to attend were enrolled in the intervention. Most participants were contacted through three collaborating community agencies that served and treated a chemically dependent clientele, community-based service organizations, businesses serving predominantly African Americans, community newspapers, and community activities and celebrations. Individuals volunteered for the study by either indicating to community or agency recruiters or staff that they were willing to participate, completing a tear-out page from a recruitment brochure, or calling a telephone line dedicated to the project. The intervention was designed with a focus on African American women, but women of other races and ethnicities were not excluded from participation. Ethnicity was based on self-report in response to a question asking participants to indicate their cultural or ethnic group identification.

Using a computer-generated random numbers table, participants were randomly assigned to either an intervention ("treatment") group or a control group. Based on a power analysis, 306 primarily African American women were recruited from the metropolitan area of two large midwestern cities, Minneapolis and St. Paul, Minnesota. The substudy reported here was directed at determining predictors of risky sexual behavior in African American women; thus, the sample for these analyses included only African American women who reported having been sexually active with a man within the last 3 months (N = 163).

Demographics (Table 1) illustrate the low-income nature of the sample. Mean age of the 163 participants was 32.2 years (SD = 8.96; range = 15-62 years), with 61% between the ages of 15 and 34. Only 15% of the sample was married, and most (68%) had never been married. Twenty-nine percent had less than a high school education; 53% were employed full- or part-time, while 47% were unemployed or not actively looking for work; thus, 52% had total household incomes at or below the poverty level. Participants were mobile, averaging 2.2 residences in the past year. Fifty percent lived in rental properties or subsidized housing, and 28% resided in a halfway house, group home, treatment center, or shelter; only 20% of the sample owned their own homes. Forty-seven percent of the sample had at least one arrest, and of those, 78% had spent some time in jail or prison. Finally, 30% reported that they were either trying to get pregnant, open to the possibility of pregnancy, or planning to have another baby during the next year.

Procedure

Data for this study were taken from a pretest measure administered prior to any participation in the WISH intervention. Data were collected using in-person structured interviews between July 1997 and August 1998. Trained peers and staff from the collaborating agencies conducted interviews at places mutually agreed upon by the subject and interviewer, usually the participants' home or one of the agency's offices. Participants received monetary reimbursement of up to $75, depending upon their completion of all phases of the study.

Instruments

The two-hour structured interview consisted of 409 questions using mostly fixed-choice and 5-point Likert-type scales, with a limited number of open-ended questions. Whenever possible, the interview included short scales obtained from standardized instruments or from other studies of similar populations. For this study, 17 predictor and 5 dependent variables from the full interview schedule were selected for analysis, as described below.

For all scales, mean scores were obtained by dividing the total score by the number of items in the scale, thus converting the total to original scale metric. Mean substitution (per individual) was done when 10% or less of the items of a scale were missing in order to maximize available sample size. Cronbach's internal consistency reliability analyses were conducted using 218 participants from the original pretest WISH sample who completed the intervention and the post-test measure (Robinson, Uhl, et al., 2002). When initial internal consistency reliability was low, factor analysis (using varimax rotation and minimum factor loadings of 0.3) was conducted to determine appropriate items for scale inclusion.

Predictor Variables of Primary Interest

Acculturation. Adapted from Exner et al. (1995), this 6-item scale measured the degree to which respondents identified with African American culture, through participation in African American customs and holidays; interest in African American issues, interests, and people; and a sense of pride in being African American (e.g., How much pride does being African American give you?; How important is it for you to celebrate African American holidays [e.g., Kwanzaa, Martin Luther King Day, etc.?]; How often/much of the time do you read newspapers, magazines and books concerned with African American issues or interests or watch television shows that are about African American people or issues that concern them?). Items were scored on a 5-point Likert scale, with lower scores indicating greater identification with African American culture (less acculturated). The scale demonstrated marginal internal consistency reliability on the pretest sample ([alpha] = 0.63).

Sexual difficulties. Nine questions based on the National Health and Social Life Survey (Laumann, Gagnon, Michael, & Michaels, 1994) were used to assess respondents' experience with problems in their sex life during the past 12 months (e.g., no interest in sexual intercourse; unable to have orgasms; physical pain during sexual activity, including vaginal or anal intercourse; does not find sex pleasurable; worried about having to have sex; trouble lubricating). One point was given for each identified problem. This scale demonstrated adequate internal consistency reliability on the initial pretest sample (Kuder-Richardson [alpha] = 0.72).

Sexual assertiveness/self efficacy. Adapted from the Dyadic Sexual Regulation Scale (Catania, 1998), this 6-item scale measured respondents' degree of confidence in their ability to express their sexual needs and impact the quality of their sexual relationships (e.g., "I feel that it is difficult to get my main sexual partner to do what makes me feel good during sex," and "If my sexual relations with--are not satisfying, there is little I can do to change things."). Items were scored on a 5-point Likert scale, with higher scores indicating greater confidence. This scale demonstrated marginal internal consistency on the initial pretest sample ([alpha] = 0.62).

Attitudes toward condoms. This 13-item, 5-point scale was adapted from Jemmott, Jemmott, Spears, Hewitt, and Cruz-Collins (1992) and Wingood (1997). Higher scores indicate more generally positive attitudes toward condoms. Factor analysis and Cronbach's alpha demonstrated adequate internal consistency on the initial pretest sample ([alpha] = 0.75).

Assessment of HIV risk. Respondents' perception of HIV risk in their environment was rated on a 5-point Likert-type scale by asking them how many of their close women family and friends had been exposed to the HIV virus; answers ranged from 1 (hardly any) to 5 (most). This variable was selected to replicate as closely as possible a similar variable assessing respondents' perceptions of the level of AIDS-related risk involved in their own sexual activity (Hines et al., 1998).

Desired or intended pregnancy. Research has shown that pregnancy planning is complicated to measure and that such planning does not fit into the way some women, especially women in poverty, see their lives (Bachrach & Newcomer, 1999; Peterson & Mosher, 1999; Trussell, Vaughan, & Stanford, 1999). We followed definitions used by Henshaw (1998) in constructing our measure of desired or intended pregnancy. Participants were asked to indicate if they were trying to get pregnant, open to the possibility of pregnancy, planning to have another baby during the next year, or doing things to keep from getting pregnant. If any of the first three questions was answered "yes" or the last question was answered "no," the respondent was classified as desiring pregnancy for the purposes of this study and coded as 1.

Alcohol and drug use. Respondents were asked how often they used various chemicals (Minnesota Department of Human Services, 1995), as well as how much they typically consumed on a day when they used (El Bassel, 1996). For the current analysis, respondents were characterized as follows: abstainers (no chemical use in the last 3 months and not currently in chemical dependency treatment; coded 0), moderate users (some chemical use in the last 3 months but not enough to be classified as heavy use; coded 1), and heavy users (used alcohol daily, or drink 5 or more drinks per occasion, or used pot more than 3 times per week, or used cocaine or heroin in the last 3 months, or currently in chemical dependency treatment; coded 2). Our measures of alcohol use were calculated to be as similar as possible to procedures outlined by Hines et al. (1998). In addition, participants were asked how often they injected drugs in the last 3 months; responses ranged from "none used" to "daily."

Domestic abuse in current or most recent relationship. Participants were divided into two groups based on their responses to an item asking about their recent history of physical abuse. Domestic abuse was reported by 24% of the women, indicating that they had been hit, kicked, hurt, punched, slapped, thrown, choked, restrained, or beaten by their current or most recent sexual partner (coded 1); the remainder reported no recent history of domestic abuse (coded 0).

Sexual abuse. Participants were divided into two groups based on their responses to relevant items on the Compulsive Sexual Behavior Inventory (Coleman, Miner, Ohlerking, & Raymond, 2001), a 28-item scale that included items originally developed to reflect the behaviors, affects, and developmental histories of individuals seen clinically for compulsive sexual behavior. This scale has good reliability and ability to discriminate between individuals with and without diagnoses of compulsive sexual behavior; its factor structure and validity has been replicated in divergent populations (Coleman et al.). The relevant items for this study questioned respondents on their histories of child and adult sexual abuse. Sexual abuse in childhood or adulthood was reported by 58% of the women, indicating that these participants reported being sexually abused; having sexual encounters with either parent or with a sibling; being forced to have sex with a stranger, casual acquaintance, or friend; or being forced to have sex with their primary partner or lover (coded 1). The rest of the participants reported no history of sexual abuse or assault (coded 0).

Prostitution. Respondents were classified as having engaged in prostitution based on their response of "yes" (coded 1) to two questions: "having sex in exchange for money or drugs in the past 3 months" or ever receiving "money to have sex."

STDs. Respondents were divided into two groups based on their response of "yes" (coded 1) to any one of a series of eight listed STDs. Adapted from the New Haven Women's Health Project (1994) and Quadagno (1994), they were asked if a doctor or other health care provider had ever told them that they had one of a number of STDs (syphilis, gonorrhea, genital herpes, genital warts, hepatitis, chlamydia, pelvic inflammatory disease, or any other STD).

Potential Moderator Variables

Age. Respondents were asked how old they were in years. This was treated as a continuous variable.

Education. Respondents were asked to report the highest year of school they had completed. Responses were grouped into one of 5 categories ranging from "less than high school" to "graduate work or higher."

Marital status. Responses were grouped into two categories: "married" (coded 1) or "not legally married at pretest" (coded 0).

Employment. Responses were recoded as a dichotomous variable: "yes," coded as 1, or "no," coded as 0, indicating whether or not the respondent was employed full- or part-time on a regular basis.

Religious service attendance. Respondents were asked how often they attended religious services during the past year. The 5-point scale ranged from 1 (at least once a day) to 5 (never).

Dependent Variables: AIDS-Related Risky Sexual Behavioral Variables

Based on Hines et al. (1998), we constructed five measures of risky sexual behavior.

Multiple concurrent partners. This complex concept was measured by three variables (self-report, belief about partners, and belief about monogamy in relationship), each assessing a slightly different aspect of the concept.

The self-report measure (called "multiple sexual partners" by Hines et al., 1998) was based on questions that asked the respondent whether she had additional partners (other than her main partner) over the past 3 months. Those who reported having only a main partner were categorized as not having multiple concurrent partners (coded 0); those who reported having other current sexual partners were categorized as having multiple concurrent partners (coded 1).

The belief about partner measure was based on items asking respondents if they thought their main partner had had sex with someone else while they were together (response options included "had no other partners," "I suspect s/he had other partners," "I know s/he had other partners," and "don't know"). In categorizing their partner as having multiple concurrent partners, all ambiguous responses (e.g., suspecting that he had other partners, indicating "don't know" or "other") were resolved by assigning the variable to its riskiest value, multiple concurrent partners. The variable was constructed using two categories: those who believed that their partner did not have other concurrent partners (coded 0) or those who believed or suspected their partner did have other concurrent partners (coded 1).

The belief about monogamy in relationship measure was a composite of the previous two measures assessing multiple concurrent partnerships: self-report and belief about partner. The variable was constructed using two categories: either one or both most likely had other partners (non-monogamous sexual behavior, coded 0) or neither had other partners (monogamous sexual behavior, coded 1).

Consistent/inconsistent condom use. This measure was based on several questions that asked how often the respondent had engaged in vaginal or anal sex without a condom during the past 3 months. Items were presented first with reference to a main sexual partner and again with reference to another sexual partner. This variable was dichotomized to represent inconsistent condom use (condom not used every time, coded 1) versus consistent condom use (condom used every time, coded 0).

Low/high risk sexual behavior. This variable was created by combining the variables for condom use and multiple concurrent partners and was comprised of two categories: high-risk (multiple concurrent partners, inconsistent condom use, coded 1) and low-risk (no multiple concurrent partnerships and consistent condom use; no multiple concurrent partnerships and inconsistent condom use; multiple concurrent partnerships and consistent condom use; coded 0) sexual behavior.

Statistical Analysis

All of the sexual health (predictor) variables were correlated with the five dichotomous (dependent) variables measuring risky sexual behavior. Three of the sexual health variables (sexual difficulties, attitudes toward condoms, sexual assertiveness/self efficacy) were significantly correlated with at least one of the outcome variables and were included as predictor variables in the logistic regressions. Nine variables (acculturation, alcohol and drug use, domestic abuse, sexual abuse, IV drug use, prostitution, STD history, assessment of HIV risk, and desire or intention for pregnancy) were considered important predictors because of their prominence in the literature. Five potential moderator variables (age, education, marital status, employment, and attendance at religious services) were also thought to be predictive based on the literature. These variables constitute a total of 18 degrees of freedom for a sample size of 163. Sample size precluded analysis of interactions. All five regression models included an interaction term between acculturation and alcohol and drug consumption (as was done by Hines et al., 1998), resulting in large standard errors, rendering those models unusable. Therefore, these interactions were dropped from further analysis.

We used forced entry logistic regression to examine the relationships between the five dependent variables measuring risky sexual behavior and the predictor variables described above. The dependent variables were the three multiple concurrent partnerships variables (self-report, belief about partner, and belief about monogamy in relationship), as well as condom use and low/high risk behavior. An alpha level of .05 was used for all analyses.

We checked each analysis for goodness of fit. The Hosmer-Lemeshow goodness-of-fit test was used to determine if the logistic regression model assumptions had been met. Categorical predictor variables that had a standard error greater than 1 for any individual degree of freedom were considered a likely candidate for elimination from the model. In order to assess the extent of multicollinearity, we calculated variance inflation factors (VIF) for all the predictor variables using a linear regression model. This approach was recommended by Allison (1999).

For dichotomous variables, odds ratios less than one and their corresponding reference categories were inverted for ease of interpretation. All continuous predictor variables were assessed for skewness; all had skewness between +/- 1.0, indicating only minor departures from normality. All analyses were conducted using SPSS version 10.1.

RESULTS

Risk Indicators of the Sexually Active African American Female Sample

Table 2 describes several risk indicators of this sexually active sample. Participants reported surprisingly high levels of multiple concurrent partnerships: 25% reported that they had had more than one partner in the past three months, and 59% knew, suspected, or didn't know if their partner had had sex with someone else while they were together. Based on their answers to these two questions, 64% were classified as being in a non-monogamous relationship during the past three months. Additionally, most respondents (86%) used condoms inconsistently. This resulted in 51% being classified as being in a high-risk relationship.

Average age at first sexual intercourse was 15.2 years, with 70% having their first experience with sexual intercourse by the age of 16. Almost one third (30%) had engaged in sex for money, drugs, or food in their lifetimes. Average number of lifetime partners was 25.5, with the usual skewed distribution for this variable (SD = 86.23) and a median of 9 lifetime sexual partners. Several of the respondents were sex workers and estimated their number of lifetime sexual partners at close to 1,000. Not surprisingly then, almost one quarter (24%) had an STD sometime during their lives, and 14% during the last three months. In addition, 87% had taken an HIV test sometime during their lifetime.

Two of the three collaborating agencies dealt with a chemically dependent clientele, and our sample reflected this: 34% percent were in chemical dependency treatment, and 43% had an alcohol or drug problem currently or in the past; only 2% had ever shared a needle with someone. Finally, 58% were sexually abused as a child or an adult, and 24% had experienced domestic abuse in their current or most recent relationship.

Logistic Regression Analysis: Predictors/Correlates of Safer Sex Behaviors

The VIE derived from linear regression, for all the predictor variables used in the five logistic regression models ranged from 1.15 to 2.55 (see Tables 3 & 4). Allison (1999) suggested that VIF > 2.5 (tolerance < .4) is problematic when using linear regression models to measure multicollinearity in logistic regression. When dummy coding categorical variables, the correlations between the individual degrees of freedom can often result in increased multicollinearity in linear regression. Thus, the VIF of 2.55 for one of the dummy variables (alcohol and drug use) is not surprising. Given that the VIF for this dummy-coded variable is barely > 2.5, we do not feel it is a cause for concern. Thus, multicollinearity was not a problem in these logistic regressions models.

Table 3 summarizes the odds ratios and the p-values for the first three logistic regressions examining the relationship between the predictor variables and three different measures of multiple concurrent partnerships. In all three logistic regressions, having sexual difficulties was the only variable significantly and positively related to having multiple partners. A woman who reported having one additional sexual difficulty in the last 12 months was 1.44 times more likely to report having multiple concurrent partners during the last 3 months, 1.52 times more likely to believe that her partner had multiple concurrent partners during the last 3 months, and, inverting the odds ratio, 1.47 times more likely to believe she was in a non-monogamous relationship.

Table 4 summarizes the odds ratios and the p-values for two logistic regressions examining the relationship between the predictor variables and two risky behavioral variables: consistent condom use and a composite high-risk sexual behavior variable. In the first logistic regression, two variables, attitudes towards condoms and a desire or intention for pregnancy, were significantly related to consistent condom use. More favorable attitudes toward condoms were significantly associated with a greater likelihood of consistent condom use. Specifically, a person with a one-point increase toward more positive condom attitudes (scale scores could range from 1.0 to 5.0) would be 13.81 times more likely to use condoms consistently. In addition, desiring pregnancy was significantly related to a lower likelihood of consistent condom use. Someone who desired or intended pregnancy was 11.39 times less likely to use condoms consistently compared to someone who had no desire or intention for pregnancy.

In the last logistic regression, two variables, report of sexual difficulties and a desire or intention for pregnancy, were significantly related to the composite measure of high-risk sexual behavior. A woman who reported having one additional sexual difficulty in the last 12 months was 1.35 times more likely to be engaging in high-risk sexual behavior. Inverting the odds ratio, a woman who desired or intended pregnancy was 3.03 times more likely to engage in high-risk behavior compared to someone who had no desire or intention for pregnancy.

DISCUSSION

In this study, we applied the Sexual Health Model (Robinson, Bockting, et al., 2002) to our investigation of the relationship of sexuality variables to risky sexual behavior. Our hypotheses regarding associations between sexuality variables and sexual risk behaviors were partially supported. Most notably, having sexual difficulties/dysfunctions was associated with four of the five risky sexual behavior variables (all three multiple concurrent partner variables as well as the composite high-risk sexual behavior variable). Women who reported having sexual difficulties were more likely to report that they had multiple concurrent partners, their partners had multiple concurrent partners, and, thus, that they were in a non-monogamous relationship during the past three months. In addition, women with sexual difficulties were more likely to report that they engaged in high-risk behavior overall (the composite measure).

There are many possible explanations for this intriguing finding linking sexual difficulties and risky sexual behavior. It is possible that being in a non-monogamous relationship interferes with sexual quality, making sexual dysfunction more likely. This might occur, for example, if those with multiple partners in non-monogamous relationships are at greater risk for cuts and abrasions that cause pain and thus interfere with sexual pleasure. In support of this explanation, we found that 43% of our respondents with multiple partners (vs. 35% of our respondents without multiple partners) reported having had pain during sexual intercourse. It is also possible that being in a non-monogamous relationship interferes with relational and psychological factors such as sexual intimacy and closeness, which then impact sexual functioning. On the other hand, causality may be reversed; it is also possible that women with sexual dysfunction may blame their primary partner for their lack of sexual pleasure and seek out other partners in an attempt to overcome the dysfunction.

Other feasible explanations focus on the role of monogamy. It is possible that being in a monogamous relationship improves sexual quality, thus making sexual dysfunction less likely. There is evidence that women were more likely to be free of sexual dysfunction (e.g., to have desire, to become aroused, to experience orgasm) when in a monogamous relationship characterized by intimacy and trust (Hurlbert, Apt, & Rabehl, 1993; Stuart, Hammond, & Pett, 1987). New research suggests that women's subjective experience of sexual arousal characteristics is based more on their appraisal of the situation than on merely physiological arousal (Everaerd, Laan, Both, & Van Der Velde, 2000). On the other hand, causality may be reversed; having a satisfying sex life with one's partner may enhance intimacy and satisfaction, thereby encouraging monogamy. Unfortunately, this correlational study cannot elucidate causality and, thus, cannot shed light on which explanation is most accurate. Future research needs to replicate this finding and then investigate which of these, or other, explanations is most likely.

Other important findings included (a) positive attitudes toward condoms were positively associated with consistent condom use and (b) desiring or intending pregnancy was associated with inconsistent condom use and high-risk sexual behavior. Both positive attitudes toward condoms and desiring or intending pregnancy were particularly powerful predictors of consistent condom use, having the highest odds ratios of all of our regression analyses. This latter finding is supported by similar findings that demonstrated that desiring pregnancy was positively related to risky sexual behavior (Adler & Tschann, 1993, as cited in O'Leary & Wingood, 2000; Cabral et al., 2001). In addition, Wyatt (1997) suggested that some African American women may be unable to initiate condom use due to a desire on their partner's part to have a child. It would appear that, quite logically, women interested in becoming pregnant (either because of their own desires or those of their partner) are less likely to use methods of birth control that would interfere with their ability to conceive.

Our hypothesis that certain challenges/barriers to sexual health (heavy alcohol and drug use, IV drug use, prostitution, recent history of physical abuse, histories of child and adult sexual abuse, and history of STDs) were positively related to risky sexual behaviors was not supported. Our measure of prostitution was limited by the fact that respondents reported on whether they had ever engaged in some form of prostitution, a time period that may be too broad to demonstrate this relationship clearly. There is a general belief that alcohol and drugs are unequivocally related to high-risk sexual behavior (O'Leary & Wingood, 2000; Ross et al., 2001). However, the large percentage of women in our sample (34%) who were in inpatient or outpatient chemical dependency treatment and prohibited from engaging in sex while in treatment could have obscured a relationship between chemical use and safer sex. That said, the evidence for a positive relationship between alcohol and drug use and safer sex behaviors among different ethnic groups is still far from clear. Caetano and Hines (1995) reported that for women and men in three mixed-risk ethnic groups (Black, Hispanic, and White), heavier drinking patterns were associated with being nonmonogamous or being with a nonmonogamous partner. Research by Graves and Hines (1997), however, suggests that the relationship between chemical use and high-risk sexual behavior may vary among different ethnic groups. For example, they found that although alcohol use was positively related to the likelihood of having sex with a casual partner in their total sample of White, Hispanic, and Black women, this relationship between alcohol use and casual sex was not significant in their representative sample of Black women. Graves (1995) concluded that the relationship of drinking to sexual activity is complex--an intricate interplay of personality, expectancies, and circumstance.

The situation is similar with regard to the relationship between sexual abuse and sexual risk behavior. We agree with Whitmire, Harlow, Quina, and Morokoff (1999), who noted, "In the search for links between childhood trauma and adult HIV risk, two conclusions seem clear from the data: First, there is a connection, and second, it is not a simple, direct link" (p. 117). They listed seven constructs which receive consistent support in the research literature as outcomes of childhood sexual abuse and as predictors of HIV risky behavior; the dependent variables examined in this study were not on their list. Additionally, although Wingood and DiClemente (1997) found domestic assault was related to less frequent condom use, their measure of condom use was not as stringent as the measure we used in this study (consistent condom use), and that might partially explain why we did not find a similar relationship.

Additionally, our hypothesis about the relationship between sexual assertiveness/self efficacy and assessment of HIV risk to risky sexual behavior was not supported. The sexual assertiveness/self-efficacy measure used here was similar to the Sexual Communication-Preferences Scale, which, unlike the Sexual Communication-Information Scale, was not associated with safer sex (Quina et al., 2000). It would seem that talking about, initiating, or feeling in control of one's sexual pleasure, as the Sexual Communication-Preferences Scale and our measure sexual assertiveness/self-efficacy assess, is not the same as discussing HIV risk more specifically, at least as far as sexual risk behavior is concerned (Morokoff et al., 1997; Quina et al.). Similarly, our measure of participants' assessment of their HIV risk, while selected to replicate as closely as possible a similar variable used by Hines et al. (1998), was measured differently. This may explain why we did not replicate their finding that African American women who perceived their behavior as risky were more likely to use condoms consistently than those who did not.

Our second hypothesis was an attempt to replicate and extend Hines et al.'s (1998) results regarding associations between acculturation, alcohol and drug use, and HIV-related sexuality variables in an African American sample. Like these researchers, we did not find a linear relationship between acculturation and any of our HIV-related sexuality variables. However, because of the small sample size of some cells in our logistic regression model, we were unable to test their finding that highly acculturated African American women who were heavy drinkers were more likely to have multiple partners.

All five regression models including an interaction term between acculturation and alcohol and drug consumption (as was done by Hines et al., 1998) resulted in large standard errors, rendering those models unusable. Consistent condom use was infrequent in this sample; only 14% of respondents reported using condoms consistently in the last 3 months. This lopsided distribution produced five standard errors larger than 1, the largest being 1.45. Results regarding consistent condom use should be replicated with a larger sample, which can compensate for the low percentage of consistent condom users. In the logistic regression predicting multiple partners, using marital status as a predictor produced a standard error of 3.64. This was because only 25 respondents (15% of the sample) were married and only 3 of those reported having another partner. Therefore, in this logistic regression model, marital status was dropped as a predictor variable. Researchers wanting to understand the relationship between marital status and multiple partners in a similar sample, while controlling for other risk variables, will need to use a much larger sample in order to compensate for the low percentage of married respondents.

Although acculturation has been repeatedly shown to modify significantly the effect of many health factors among Latino populations (O'Leary & Wingood, 2000), little is known about its relationship to health factors in African American populations. In addition, the application of the concept of acculturation is controversial in African American populations. Certainly, the concept of acculturation is more complex and difficult to measure when applied to African American culture, as opposed to immigrant or refugee cultures for which measures such as language ability and years in the U.S. provide concrete indicators of acculturation (Robinson, Freske, et al., 1999; Strunin, 1991). Given the state of the knowledge base, we encourage future researchers to continue to explore the relationship between acculturation and health factors, including safer sex, in African American populations.

Our final hypothesis was that several potential moderator variables (age, education, marriage, regular employment, and religious service attendance) were negatively related to risky sexual behaviors. Our data indicate no relationship between the four risky sexual behaviors and age, education, or marital or employment status, despite evidence to the contrary reported previously (Newcomb et al., 1998; O'Leary & Wingood, 2000; Sabogal et al., 1995). These discrepancies might be explained, in part, by sample differences; earlier investigations focused on Latina women (Newcomb et al.; Sabogal et al.), or used a national probability sample of African American women (Hines et al., 1998). For example, in our sample of low-income African American women, only 15% were married, which may have obscured any relationship between marital status and safer sex that existed.

Finally, religious service attendance was not related to any of our five dependent variables in our sample. Given that other authors have found relationships between religious affiliation and HIV risk in representative samples (Finer et al., 1999), this lends some support to hypotheses and findings (Douglas, 1999; Seidman et al., 1992; Staples, 1972, 1973) that religion or religious orientation (Allen & Olson, 2001) affects Black sexual codes differently than White sexual codes. Weatherford and Weatherford (1999) noted striking cultural differences in the religious patterns and beliefs of African Americans and Caucasians. For example, even though African Americans are more likely to attend a weekly religious service than Whites (40% versus 29%), this attendance may not impact Blacks in the same way. Additionally, the types of religious institutions that Blacks and Whites belong to are likely to be very different as well (McCubbin, Thompson, Thompson, & Futrell, 1998).

Limitations

Our sample was neither random nor representative, with participants reporting high frequencies of risky behaviors (e.g., only 14% reported always using a condom, 36% believed they were in a monogamous relationship, 43% were heavy users of alcohol or drugs or currently in chemical dependency treatment). This limits the generalizability of the study findings to similar low-income, high risk populations of African American women. At the same time, this prevalence of high-risk behaviors did provide more statistically powerful opportunities for prediction. Another concern is that our study relied completely on participant self-reports of behavior to an interviewer, conditions under which adults typically under-report undesirable sexual attitudes and activities and over-report desirable ones (Turner, Miller, & Rogers, 1997; Wagstaff, Abramson, & Pinkerton, 2000). Options for remedying this problem have been described previously (Robinson, Uhl, et al., 2002). Thirdly, our study would have benefited from using a more extensive measure of acculturation with higher reliability and validity, such as Landrine and Klonoff's (1996) 74-item, 8-subscale African American Acculturation Scale. Given our need to shorten an already long interview, our measure of acculturation consisted of a 6-item scale. Lastly, our measure of relationship monogamy (belief about monogamy in relationship), while conservative, may be affected by the inaccuracy of the data on partners. Nonetheless, a surprisingly high percentage of women reported that their partner was having or could have had multiple concurrent partners (59%). This suggests that study participants were being realistic, or at least not naively optimistic, about their partners' risky behavior.

Conclusion

This study of low-income, high-risk, African American women adds to the relatively new body of literature examining the relationship between acculturation in African American populations and their health-related behaviors and attitudes. This is the first study to demonstrate relationships between sexual difficulties, multiple concurrent partnerships, and high-risk sexual behavior overall. It confirms previous findings demonstrating a relationship between desire and intentions for pregnancy and inconsistent condom use and high-risk sexual behavior in general. It also confirms previous findings demonstrating the importance of condom attitudes in one's willingness to use condoms consistently. These findings are congruent with growing empirical support for an explicit focus on sexuality and relationships in HIV prevention (Abraham & Sheeran, 1994; Boldero et al., 1992; Ehrhardt et al., 1992; Kalichman, 1998; Kraft, Rosser, Robinson, & Bockting, 2001; Rosser, Bockting, et al., 2002).

Future studies should continue to investigate these complex relationships in diverse populations and pay attention to measurement issues which make it more difficult to find consistent findings across studies. First, the relationship of acculturation to health behaviors and risky sexual behaviors in African American populations needs to be elucidated. Is there an important relationship as Landrine and Klonoff (1996) hypothesized, and does it operate similarly in African American populations and immigrant populations? Second, since it is not possible to prove causation or specify the direction of these relationships in this one correlational study, the exact nature of the relationship between sexual problems/dysfunction and risky sexual behaviors needs to be studied and replicated in different and more heterogeneous populations. Third, our finding that a desire for pregnancy leads to risky sexual behaviors reiterates the need for the new HIV-prevention techniques: specifically, ones which allow for procreation without significant risk of disease transmission. Finally, the HIV-prevention literature is extensive and sophisticated enough that attention should be paid to developing standards or recommended measurement strategies for common concepts and variables. If researchers could agree on consistent ways to measure some of the major variables, we could be more confident that contradictory findings in the literature were not due primarily to measurement error.

Note. The Women's Initiative for Sexual Health (WISH) was a collaborative effort between faculty and staff at the Program in Human Sexuality (University of Minnesota Medical School, Department of Family Medicine & Community Health) and three community-based organizations: Turning Point, Inc., African American Family Services, and Minneapolis Urban League. The WISH intervention was funded by the Minnesota Department of Health (AIDS/STD Prevention Services Section, #1742-634-9027), and its evaluation by the Centers for Disease Control and Prevention (AIDS/STD Prevention Services Section, Program Evaluation Research Branch, #U62/CCU513219).

Manuscript accepted November 9, 2004

REFERENCES

Abraham, C. S., & Sheeran, P. (1994). Modeling and modifying young heterosexuals' HIV-preventive behaviour: A review of theories, findings and educational implications. Patient Education and Counseling, 23, 173-186.

Adimora, A. A., & Schoenbach, V. J. (2002). Contextual factors and the black-white disparity in heterosexual HIV transmission. Epidemiology, 13(6), 707-12.

Adimora, A. A., Schoenbach, V. J., Bonas, D. M., Martinson, F. E. A., Donaldson, K. H., & Stancil, T. R. R. (2002). Concurrent sexual partnerships among women in the United States. Epidemiology, 13(3), 320-327.

Allen, W. D., & Olson, D. H. (2001). Five types of African-American marriages. Journal of Marriage and the Family, 27(3), 301-314.

Allison, P. D. (1999). Logistic regression using the SAS system: Theory and application. Cary, North Carolina: SAS Institute.

Bachrach, C. A., & Newcomer, S. (1999). Contraceptive failure and unintended pregnancy. Intended pregnancies and unintended pregnancies: Distinct categories or opposite ends of a continuum? Family Planning Perspectives, 31(5), 1-3.

Bakken, R. J., & Winter, M. (2002). Family characteristics and sexual risk behaviors among black men in the United States. Perspectives on Sexual and Reproductive Health, 34(5), 252-257.

Bockting, W. O. (1998). Transgender HIV Prevention: A Minnesota response to a global health concern. Minneapolis, MN: Program in Human Sexuality.

Bockting, W. O., & Robinson, B. E. (1999). Beyond." Sexual health seminars for women who have sex with women and men manual. Minneapolis, MN: Program in Human Sexuality.

Boldero, J., Moore, S., & Rosenthal, D. (1992). Intention, context, and safe sex: Australian adolescents' responses to AIDS. Journal of Applied Social Psychology, 22(17), 1,374-1,396.

Cabral, R. J., Galavotti, C., Armstrong, K., Morrow, B., & Fogarty, L. (2001). Reproductive and contraceptive attitudes as predictors of condom use among women in an HIV prevention intervention. Women and Health, 33, 117-132.

Caetano, R., & Hines, A. M. (1995). Alcohol, sexual practices, and risk of AIDS among Blacks, Hispanics, and Whites. Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology, 10, 554-561.

Caetano, R., & Kaskutas, L. A. (1995). Changes in drinking patterns among Whites, Blacks and Hispanics, 1984-1992. Journal of Studies on Alcohol, 56, 558-565.

Catania, J. A. (1998). Dyadic Sexual Communication Scale and Dyadic Sexual Regulation Scale. In C. M. Davis, W. L. Yarber, R. Bauserman, G. Schreer, & S. L. Davis (Eds.). Handbook of sexuality-related measures. Thousand Oaks, CA: Sage Publications.

Catania, J. A., Coates, T. J., Stall, R., Turner, H., Peterson, J., Hearst, N., et al. (1992). Prevalence of AIDS-related risk factors and condom use in the United States. Science, 13(258), 1,101-1,106. Centers for Disease Control and Prevention. (2001). 2000 HIV/AIDS and Sexually Transmitted Disease Surveillance Report. Retrieved April 16, 2002 from http://www.cdc.gov.

Choi, K., Catania, J. A., & Dolcini, M. M. (1994). Extramarital sex and HIV risk behaviors among U.S. adults: Results from the National AIDS behavioral survey. American Journal of Public Health, 84(12), 2,003-2,007.

Coleman, E., Miner, M, Ohlerking, E, & Raymond, N. (2001). Compulsive Sexual Behavior Inventory: A preliminary study of reliability and validity. Journal of Sex & Marital Therapy, 27, 325-332.

Deren, S., Davis, W. R., Tortu, S., Beardsley, M., Ahluwalia, I., & National AIDS Research Consortium. (1995). Women at high risk for HIV: Pregnancy and risk behaviors. The Journal of Drug Issues, 25(1), 57-71.

DeVincenzi, I. (1994). A longitudinal study of human immunodeficiency virus transmission by heterosexual partners. New England Journal of Medicine, 331, 341-347.

Dolcini, M. M., Coates, T. J., Catania, J. A., Kegeles, S. M., & Hauck, W. W. (1995). Multiple sexual partners and their psychosocial correlates: The population-based AIDS in multiethnic neighborhoods (AMEN) study. Health Psychology, 14(1), 22-31.

Douglas, K. B. (1999). Sexuality and the Black Church: A womanist perspective. Maryknoll, NY: Orbis Books.

Ehrhardt, A. A., Yingling, S., Zawadzki, R., & Martinez-Ramirez, M. (1992). Prevention of heterosexual transmission of HIV: Barriers for women. Journal of Psychology and Human Sexuality, 5(1/2), 37-67.

El Bassel, N. (1996). Project SUPPORT, Baseline Questionnaire, Main. Unpublished manuscript. Columbia University School of Social Work, New York.

Everaerd, W., Laan, E. T. M., Both, S., & Van Der Velde, J. (2000). Female sexuality. In L.T. Szuchman & F. Muscarella (Eds.), Psychological perspectives on human sexuality. New York: John Wiley & Sons, Inc.

Exner, T., Meyer-Bahlburg, H., Yingling, S., Hoffman, S., Ortiz-Torres, B., & Ehrhardt, A. A. (1995). Baseline interview for Project FIO (The future is ours). Unpublished manuscript, Columbia University, Department of Psychiatry, New York.

Feldblum, P. (1991). Results from prospective studies of HIV-discordant couples. AIDS, 5, 1265-1277.

Finer, L. B., Darroch, J. E., & Singh, S. (1999). Sexual partnership patterns as a behavioral risk factor for sexually transmitted diseases. Family Planning Perspectives, 31(5), 228-236.

Ford, K., & Norris, A. E. (1993). Urban Hispanic adolescents and young adults: Relationship of acculturation to sexual behavior. The Journal of Sex Research, 30(4), 316-323.

Ford, K., Sohn, W., & Lepkowski, J. (2002). American adolescents: Sexual mixing patterns, bridge partners, and concurrency. Sexually Transmitted Diseases, 29(1), 13-19.

Goodman, L. A., & Fallot, R. D. (1998). HIV risk-behavior in poor urban women with serious mental disorders: Association with childhood physical and sexual abuse. American Journal of Orthopsychiatry, 68(1), 73-83.

Graves, K. L. (1995). Risky sexual behavior and alcohol use among young adults: Results from a national survey. American Journal of Health Promotion, 10(1), 27-36.

Graves, K. L., & Hines, A. M. (1997). Ethnic differences in the association between alcohol and risky sexual behavior with a new partner: An event-based analysis. AIDS Education and Prevention, 9(3), 219-237.

Henshaw, S. K. (1998). Unintended pregnancy in the United States. Family Planning Perspectives, 30(1), 24-30.

Hines, A. M., & Graves, K. L. (1998). AIDS protection and contraception among African American, Hispanic, and White women. Health & Social Work, 23(3), 186-194.

Hines, A. M., Snowden, L. R., & Graves, K. L. (1998). Acculturation, alcohol consumption and AIDS-Related risky sexual behavior among African American women. Women & Health, 27(3), 17-35.

Hurlbert, D. E, Apt, C., & Rabehl, S. M. (1993). Key variables to understanding female sexual satisfaction: An examination of women in non-distressed marriages. Journal of Sex and Marital Therapy, 19(2), 154-165.

Ickovics, J., Morril, A. C., Beren, S. E., Walsh, U., & Rodin, J. (1994). New Haven Women's Project Questionnaire. Unpublished document. New Haven.

Jemmott, J. B., Jemmott, L. S., Spears, H., Hewitt, N., & Cruz-Collins, M. (1992). Self-efficacy, hedonistic expectancies, and condom-use intentions among inner city black adolescent women: A social cognitive approach to AIDS risk behavior. Journal of Adolescent Health, 512-519.

Johnson, L. W., & Harlow, L. (1996). Childhood sexual abuse linked with adult substance use, victimization, and AIDS-risk. AIDS Education and Prevention, 8(1), 44-57.

Kalichman, S. (1998). Preventing AIDS: A sourcebook for behavioral interventions. Mahwah, NJ: Lawrence Erlbaum Associates.

Kline, A., Kline, E., & Oken, E. (1992). Minority women and sexual choice in the age of AIDS. Social Science Medicine, 34(4), 447-457.

Klitsch, M. (2002). Predictors of overlapping relationships are different for men and women. Perspectives on Sexual and Reproductive Health, 34(5), 267-272.

Koenig, L. J., & Clark, H. (2004). Sexual abuse of girls and HIV infection among women: Are they related? In L. J. Koenig, L. S. Doll, A. O'Leary, and W. Pequegnat (Eds.), From Child Sexual Abuse to Adult Sexual Risk. Washington, DC: American Psychological Association.

Kraft, C., Rosser, B. R. S., Robinson, B. E., & Bockting, W. O. (2001). Body Mass Index, Body Image, and Unsafe Sex in Men who have Sex With Men. Manuscript submitted for publication.

Landrine, H., & Klonoff, E. A. (1996). African American acculturation: Deconstructing race and reviving culture. Thousand Oaks, CA: Sage Publications.

Laumann, E. O., Gagnon, J. H., Michael, R. T., & Michaels, S. (1994). The Social Organization of Sexuality: Sexual practices in the United States. Chicago: University of Chicago Press.

Manhart, L. E., Aral, S. O., Holmes, K. K., & Foxman, B. (2002). Measurement, prevalence, and correlates among urban 18-39-year-olds. Sexually Transmitted Diseases, 29(3), 134-143.

McCubbin, H. I, Thompson, E. A., Thompson, A. I., & Futrell, J. A. (1998). Resiliency in African-American families. Thousand Oaks, CA: Sage Publications.

Miner, M., Robinson, B. E., Hoffman, L., Albright, C., & Bockting, W. O. (2002). Improving safer sex measures through the inclusion of relationship and partner characteristics. AIDS Care, 14(6), 827-837.

Minnesota Department of Human Services. (1995). State of Minnesota. DAANES (Drug and Alcohol Normative Evaluation System). 444 Lafayette Rd, St. Paul, Minnesota, 55101.

Morokoff, P. J., Quina, K., Harlow, L. L., Whitmire, L., Grimley, D. M., Gibson, P. R., & Burkholder, G. J. (1997). Sexual Assertiveness Scale (SAS) for Women: Development and validation. Journal of Personality and Social Psychology, 73(4), 790-804.

Morris, M., & Kretzschmar, M. (1995). Concurrent partnerships and transmission dynamics in networks. Social Networks, 17, 299-318.

Morris, M., & Kretzschmar, M. (1997). Concurrent partnerships and the spread of HIV. AIDS, 11, 641-648.

Newcomb, M. D., Romero, G. J., Wayment, H. A., Wyatt, G. E., Tucker, M. B., Carmona, J. V., et al. (1998). Acculturation, sexual risk taking, and HIV health promotion among Latinas. Journal of Consulting Psychology, 45(4), 454-467.

Norris, A. E., & Ford, K. (1999). Sexual experiences and condom use of heterosexual, low-income African American and Hispanic youth practicing relative monogamy, serial monogamy, and nonmonogamy. Sexually Transmitted Diseases, 26(1), 17-25.

O'Leary, A., & Wingood, M. (2000). Interventions for sexually active heterosexual women. In J. L. Peterson & R. J. DiClemente (Eds.), Handbook of HIV prevention. New York: Kluwer Academic/Plenum Publishers.

Peterson, J. L., Grinstead, O. A., Golden, E., Catania, J. A., Kegeles, S., & Coates, T. J. (1992). Correlates of HIV risk behaviors in Black and White San Francisco heterosexuals: The population-based AIDS in multiethnic neighborhood (AMEN) study. Ethnicity & Disease, 2, 361-370.

Peterson, L. S., & Mosher, W. D. (1999). Contraceptive failure and unintended pregnancy: Options for measuring unintended pregnancy in Cycle 6 of the National Survey of Family Growth. Family Planning Perspectives, 31(5), 1-4.

Pulerwitz, J., Amaro, H., DeJong, W., Gortmaker, S. L., & Rudd, R. (2002). Relationship power, condom use and HIV risk among women in the USA. AIDS Care, 14(6), 789-800.

Quadagno, D. (1994). Women empowered against risks interview schedule. Unpublished document. Florida State University.

Quina, K., Harlow, L. L., Morokoff, P. J., & Burkholder, G. J. (2000). Sexual communication in relationships: When words speak louder than actions. Sex Roles, 42(7/8), 523-549.

Quina, K., Morokoff, P. J., Harlow, L. L., & Zurbriggen, E. L. (2004). Cognitive and attitudinal paths from childhood trauma to adult HIV risk. In L. J. Koenig, L. S. Doll, A. O'Leary, and W. Pequegnat (Eds.), From child sexual abuse to adult sexual risk. Washington, DC: American Psychological Association.

Robinson, B. E., Bockting, W. O., Rosser, B. R. S., Miner, M. & Coleman, E. (2002). The Sexual Health Model: Application of a sexological approach to HIV prevention. Health Education Research: Theory & Practice, 17(1), 43-57.

Robinson, B. E., Freske, P., Scheltema, K., & Chu Yang Heu, H. L. (1999). HIV/STD knowledge, attitudes, and risk behaviors in Hmong American adolescents: An unstudied population. Journal of Sex Education and Therapy, 2(4), 37-46.

Robinson, B. E., Harrell, T., Smith, P., Bockting, W. O., & Malcolm, J. (1994). WISH (Women's Initiative for Sexual Health) Seminars Manual. University of Minnesota Medical School, Minneapolis, Minnesota: Program in Human Sexuality.

Robinson, B. E., Miner, M., Bockting, W. O., Scheltema, K. E., Albright, C., Rosser, S., et al. (2001). Final report: Evaluation of WISH. Sexual Health Seminars for women. Minneapolis, MN: University of Minnesota Medical School, Program in Human Sexuality and Atlanta, GA: Centers for Disease Control & Prevention, Program Evaluation Research Branch.

Robinson, B. E., Uhl, G., Miner, M., Bockting, W. O., Scheltema, K., Rosser, B. R. S., et al. (2002). Evaluation of a sexual health approach to prevent HIV among low income, urban, primarily African-American women: Results of a randomized controlled trial. AIDS: Education & Prevention, 14, Supplement A, 81-96.

Roempke Graefe, D., & Lichter, D. T. (2002). Marriage among unwed mothers: Whites, Blacks, and Hispanics compared. Perspectives on Sexual and Reproductive Health, 34(6), 286-293.

Ross, M. W., Rosser, B. R. S., Bauer, G. R., Bockting, W. O., Robinson, B. E., Coleman, E., et al. (2001). Drug use, unsafe sexual behavior and internalized homonegativity in men who have sex with men. AIDS and Behavior, 5(1), 97-103.

Rosser, B. R. S., Bockting, W. O., Rugg, D. L., Robinson, B. E., Ross, M. W., Bauer, G. R., et. al. (2002). A randomized controlled intervention trial of a sexual health approach to long-term HIV risk reduction: I. Results of the intervention on unsafe sexual behavior among men who have sex with men. AIDS: Education & Prevention, 14, Supplement A, 59-71.

Rosser, B. R. S., Mazin, R., & Coleman, E. (2002) Cara a Cara/Face to Face Manual. University of Minnesota Medical School, Minneapolis, Minnesota: Program in Human Sexuality, Center for HIV/STI Intervention and Prevention Studies.

Rosser, B. R. S., Miner M., Gurak, L., Bockting, W. O., Merino-Espinoza, C., et al. Researching Internet-mediated sex between men: the Men's INTernet Study (MINTS). Abstracts of the XIVth International Conference on AIDS, July 7-12, 2002, TuPeF5289.

Rosser, B. R. S., Ross, M. W., Bockting, W. O., Bauer, G. R., Kraft, C., Robinson, B. E., et al. (2002). A randomized controlled intervention trial of a sexual health approach to long-term HIV risk reduction: II. Effects of the intervention on modifying internalized homonegativity in men who have sex with men. Manuscript submitted for publication.

Sabogal F., Perez-Stable, E. J., Otero-Sabogal, R., & Hiatt, R. A. (1995). Gender, ethnic, and acculturation differences in sexual behaviors: Hispanic and non-Hispanic white adults. Hispanic Journal of Behavioral Sciences, 17(2), 139-159.

Santelli, J. S., Lowry, R., Brener, N. D., & Robin, L. (2000). The association of sexual behaviors with socioeconomic status, family structure, and race/ethnicity among US adolescents. American Journal of Public Health, 90(10), 1,582-1,588.

Seidman, S., Mosher, W., & Aral, S. (1992). Women with multiple sexual partners: United States, 1988. American Journal of Public Health, 82(10), 1,388-1,394.

Shain, R. N., Perdue, S. T., Piper, J. M., Holden, A. E., Champion, J. D., Newton, E. R., et al. (2002). Behaviors changed by intervention are associated with reduced STD recurrence: The importance of context in measurement. Sexually Transmitted Diseases, 29(9), 530-532.

Short, M. B., Succop, P. A., Mills, L., Stanberry, L. R., Biro, P. M., & Rosenthal, S. L. (2003). Non-exclusivity in adolescent girls' romantic relationships. Sexually Transmitted Diseases, 30(10), 752-755.

Smith, T. W. (1991). Adult sexual behavior in 1989: Number of partners, frequency of intercourse and risk of AIDS. Family Planning Perspectives, 23(3), 102-107.

Staples, R. (1972). Research on black sexuality: Its implication for family life, sex education, and public policy. The Family Coordinator, 21(2), 183-188.

Staples, R. (1973). The sexuality of black women. New York: Interpersonal Publications.

Staples, R., & Johnson, L. B. (1993). Black families at the crossroads: Challenges and prospects. San Francisco: Jossey-Bass Publishers.

Strunin, L. (1991). Adolescents' perceptions of risk for HIV infection: Implications for future research. Social Science and Medicine, 32(2), 221-228.

Stuart, F. M., Hammond, D. C., & Pett, M. A. (1987). Inhibited sexual desire in women. Archives of Sexual Behavior, 16(2), 91-106.

Trussell, J., Vaughan, B., & Stanford, J. (1999). Are all contraceptive failures unintended pregnancies? Evidence from the 1995 National Survey of Family Growth. Family Planning Perspectives, 31(5), 246-260.

Turner, C. F., Miller, H. G., & Rogers, S. M. (1997). Survey measurement of sexual behavior. In J. Bancroft (Ed.), Researching sexual behavior. Indiana University Press: Bloomington.

Upchurch, D. M., Kusunoki, Y., Simon, E, & Dory, M. M. (2003). Sexual behavior and condom practices among Los Angeles women. Women Health Issues, 13(1), 8-15.

Wagstaff, D. A., Abramson, P. R., & Pinkerton, S. D. (2000). Research in human sexuality. In L. T. Szuchman & F. Muscarella (Eds.), Psychological perspectives on human sexuality. New York: John Wiley & Sons, Inc.

Watts, C. H., & May, R. M. (1992). The influence of concurrent partnerships on the dynamics of HIV/AIDS. Math Biosci, 108(1), 89-104.

Weatherford, R. J., & Weatherford, C. B. (1999). Somebody's knocking at your door: AIDS and the African-American church. Binghamton, New York: Haworth Pastoral Press.

Weinberg, M. S., & Williams, C. J. (1988). Black sexuality: A test of two theories. The Journal of Sex Research, 25(2), 197-218.

Whitmire, L. E., Harlow, L. L., Quina, K., & Morokoff, P.J. (1999). Childhood trauma and HIV. Women at risk. Philadelphia: Brunner/Mazel, Inc.

Wingood, G. (1997). SISTA (Sisters Informing Sisters about Topics on AIDS) Questionnaire. Unpublished manuscript. Birmingham, AL: Center for AIDS Research, University of Alabama.

Wingood, G. M., & DiClemente, R. J. (1997). The effects of an abusive primary partner on the condom use and sexual negotiation practices of African-American women. American Journal of Public Health, 87(6), 1,016-1,018.

Wohl, A., Lu, S., Odem, S., Sorvillo, F, Fegues, C. E, & Kerndt, P. R. (1998). Sociodemographic and Behavioral Characteristics of African-American Women with HIV and AIDS in Los Angeles County, 1990-1997. Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology, 19, 413-420.

Wyatt, G. E. (1997). Stolen women: Reclaiming our sexuality, taking back our lives. New York: John Wiley & Sons, Inc.

Wyatt, G. E., Myers, H. E, Williams, J. K., Kitchen, C. R., Loeb, T., Carmona, J. V., et al. (2002). Does a history of trauma contribute to HIV risk for women of color? Implications for prevention and policy. American Journal of Public Health, 92(4), 660-665.

Beatrice "Bean" E. Robinson, Karen Scheltema, and Tonya Cherry

Program in Human Sexuality, Department of Family Practice and Community Health, University of Minnesota Medical School

Address correspondence to Beatrice "Bean" Robinson, Ph.D., Program in Human Sexuality, University of Minnesota Medical School, Department of Family Practice and Community Health, 1300 S. 2nd Street, Suite 180, Minneapolis, MN, 55454; e-mail: [e-mail].

Table 1. Sociodemographic Description of the Sexually
Active African American Female Sample (N = 163)

Sociodemographic
   Variables                                n     %     Mean     SD

Age (years)                                             32.2    8.96
    15-17 (1)                                5     3
    18-24                                   29    18
    25-34                                   65    40
    35-44                                   52    32
    45-54                                    8     5
    55-62                                    4     3
Marital Status
    Never married                          110    68
    Married                                 25    15
    Separated or divorced                   25    15
    Widowed                                  3     2
Education
    < High school                           47    29
    High school                             41    25
    Some college                            50    31
    College degree                          14     9
    Graduate school                         10     6
Regularly employed,
  full- or part-time                        83    53
Total yearly income at or
  below poverty                             68    52
Number of residences in past year                       2.2     2.26
Living Situation
    Rent house or apartment                 73    48
    Treatment setting (group, halfway
      or treatment center)                  42    28
    Own house or apartment                  30    20
    Government or subsidized housing         3     2
    Other (e.g., shelter, street)            5     3
Ever arrested                               77    47
    Jailed, prison, or on work release      60    78
Open to getting pregnant                    49    30

(1) Inclusion criteria were that all participants were emancipated
minors (as defined under Minnesota law) or greater than 18 years of
age.

Table 2. Risk Indicators of the Sexually Active African American Female
Sample (N = 163)

Risk Variables                        N   %   Mean   SD    Median  Mode

Multiple concurrent partners
  Self-report                         40  25
  Belief about partner                93  59
  Belief About Monogamy in
    Relationship                     101  64
Inconsistent condom use              137  86
Composite Variable: High-Risk Sexual
    Behavior                          82  51
Age at first sexual intercourse               15.2    2.7    15     14
  6-12                                21  12
  13-16                               89  58
  17-18                               27  18
  19-25                               16  11
Had sex for money, drugs or food
  In lifetime                         49  30
  In last 3 months                    13   8
Lifetime number of sexual partners            25.5  86.23    9      4
  1-5                                 42  28
  6-10                                52  34
  11-20                               28  18
  21-999                              30  20
Had an STD
  Lifetime                            89  55
  In last 3 months                    23  14
Ever had an HIV test                 141  87
In chemical dependency treatment
  currently                           55  34
Ever used a needle with someone        3   2
Sexually abused as a child or adult   94  58
Domestic abuse in current or most
  current relationship                39  24

Table 3. Three Multiple Logistic Regression Models Exploring the
Relationship Between Sexual Health, Moderator Variables and Multiple
Concurrent Partnerships

Sexual Health
Component                   Variable(s)

                                                  VIF (a)

Cultural & sexual     Acculturation (b)            1.16
  identity:

  Sexual anatomy &    Sexual
functioning           difficulties (b)             1.50

Positive sexuality     Sexual assertiveness/
  sexuality           self efficacy (b)            1.15

Sexual health care &  Attitudes toward
  safer sex           condoms (b)                  1.22
                      Assessment of HIV risk (b)   1.25
                      Desire for pregnancy (c)     1.31

Barriers to healthy   Alcohol and drug use (d)
  sexuality            --moderate user             1.74
                       --heavy user/
                      in treatment                 2.55
                      Domestic abuse
                      in current or most
                      recent relationship (e)      1.55
                      Ever sexually abused? (f)    1.35
                      IV drugs (g)                 1.18
                      Prostitution (h)             1.59
                      Ever had an STD? (i)         1.27

Potential             Age (b)                      1.43

Moderator             Education (b)                1.65

Variables             Marital status (j)           1.29
                      Employment status (l)        1.65
                      Attendance at
                      religious services (b)       1.29

                                                      Self-Report
Sexual Health                                          (N = 121)
Component                   Variable(s)
                                                  Odds     95% CI for
                                                  Ratio    Odds Ratio

Cultural & sexual     Acculturation (b)           1.43     0.47, 4.37
  identity:

  Sexual anatomy &    Sexual
functioning           difficulties (b)            1.44 *   1.03, 2.01

Positive sexuality     Sexual assertiveness/
  sexuality           self efficacy (b)           1.39     0.52, 3.75

Sexual health care &  Attitudes toward
  safer sex           condoms (b)                 2.38     0.81, 7.02
                      Assessment of HIV risk (b)  0.98     0.63, 1.53
                      Desire for pregnancy (c)    0.59     0.17, 2.00

Barriers to healthy   Alcohol and drug use (d)
  sexuality            --moderate user            4.16     0.81, 21.50
                       --heavy user/
                      in treatment                3.62     0.66, 19.99
                      Domestic abuse
                      in current or most
                      recent relationship (e)     0.22     0.04, 1.35
                      Ever sexually abused? (f)   1.99     0.56, 7.09
                      IV drugs (g)                2.77     0.68, 11.33
                      Prostitution (h)            1.77     0.39, 8.08
                      Ever had an STD? (i)        3.07     0.79, 11.92

Potential             Age (b)                     0.93     0.85, 1.01

Moderator             Education (b)               1.26     0.71, 2.24

Variables             Marital status (j)          N/A          N/A
                      Employment status (l)       1.23     0.31, 4.90
                      Attendance at
                      religious services (b)      0.95     0.57, 1.58

                                                   Multiple concurrent
                                                        Partners

                                                  Belief About Partner
Sexual Health                                          (N = 121)
Component                   Variable(s)
                                                  Odds     95% CI for
                                                  Ratio    Odds Ratio

Cultural & sexual     Acculturation (b)           0.82     0.36, 1.89
  identity:

  Sexual anatomy &    Sexual
functioning           difficulties (b)            1.52 **  1.15, 1.99

Positive sexuality     Sexual assertiveness/
  sexuality           self efficacy (b)           0.70     0.36, 1.37

Sexual health care &  Attitudes toward
  safer sex           condoms (b)                 2.20     0.91, 5.31
                      Assessment of HIV risk (b)  0.86     0.61, 1.20
                      Desire for pregnancy (c)    1.14     0.41, 3.12

Barriers to healthy   Alcohol and drug use (d)
  sexuality            --moderate user            1.11     0.36, 3.43
                       --heavy user/
                      in treatment                1.11     0.29, 4.32
                      Domestic abuse
                      in current or most
                      recent relationship (e)     1.06     0.31, 3.58
                      Ever sexually abused? (f)   0.94     0.36, 2.51
                      IV drugs (g)                1.60     0.33, 7.73
                      Prostitution (h)            0.80     0.25, 2.62
                      Ever had an STD? (i)        1.02     0.41, 2.54

Potential             Age (b)                     1.05     1.00, 1.11

Moderator             Education (b)               1.26     0.80, 1.98

Variables             Marital status (j)          0.51     0.14, 1.93
                      Employment status (l)       0.51     0.18, 1.43
                      Attendance at
                      religious services (b)      0.84     0.56, 1.27

                                                   Composite Variable:
                                                  Belief About Monogamy
                                                     in Relationship
Sexual Health                                           (N = 121)
Component             Variable(s)
                                                  Odds     95% CI for
                                                  Ratio    Odds Ratio

Cultural & sexual     Acculturation (b)           1.19     0.51, 2.76
  identity:

  Sexual anatomy &    Sexual
functioning           difficulties (b)            0.68 **  0.51, 0.89

Positive sexuality     Sexual assertiveness/
  sexuality           self efficacy (b)           1.41     0.72, 2.76

Sexual health care &  Attitudes toward
  safer sex           condoms (b)                 0.59     0.25, 1.41
                      Assessment of HIV risk (b)  1.12     0.80, 1.57
                      Desire for pregnancy (c)    1.30     0.46, 3.66

Barriers to healthy   Alcohol and drug use (d)
  sexuality            --moderate user            0.84     0.26, 2.64
                       --heavy user/
                      in treatment                0.98     0.25, 3.82
                      Domestic abuse
                      in current or most
                      recent relationship (e)     1.08     0.32, 3.70
                      Ever sexually abused? (f)   0.96     0.36, 2.59
                      IV drugs (g)                0.64     0.17, 2.41
                      Prostitution (h)            0.93     0.28, 3.07
                      Ever had an STD? (i)        0.94     0.37, 2.38

Potential             Age (b)                     0.96     0.91, 1.02

Moderator             Education (b)               0.79     0.50, 1.26

Variables             Marital status (j)          2.33     0.63, 8.63
                      Employment status (l)       1.04     0.37, 2.91
                      Attendance at
                      religious services (b)      1.35     0.89, 2.03

(a) Derived from linear regression. (b) Continuous variable.
(c) Reference category is "no desire for pregnancy." (d) Reference
category is "not used in last 3 months and not currently in chemical
dependency treatment." (e) Reference category is "no domestic abuse in
recent or most current relationship." (f) Reference category is "never
was sexually abused." (g) Reference category is "no IV drug use within
the last 3 months." (h) Reference category is "no indication of
prostitution." (i) Reference category is "never had an STD."
(j) Reference category is "not currently married." (k) Due to an
unacceptably large standard error, the variable "marital status" was
dropped from this model. (l) Reference category is "not currently
employed regularly."

* p [less than or equal to] .05 ** p [less than or equal to] .01
*** p [less than or equal to] .001

Table 4. Two Multiple Logistic Regression Models Exploring the
Relationship among Sexual Health, Moderator Variables and High Risk
Sexual Behaviors (Consistent Condom Use and Composite High Risk
Variable)

Sexual Health
Component           Variable(s)

                                                VIF (a)

Cultural & sexual   Acculturation (b)           1.16
  identity
Sexual anatomy &    Sexual difficulties (b)     1.50
  functioning
Positive sexuality  Sexual assertiveness/
                      self efficacy (b)         1.15
Sexual health care  Attitudes toward
  & safer sex         condoms (b)               1.22
                    Assessment of HIV risk (b)  1.25
                    Desire for pregnancy (c)    1.31
Barriers to         Alcohol and drug use (d)
  healthy            --moderate user            1.74
  sexuality          --heavy user/in treatment  2.55
                    Domestic abuse in current
                      or most recent
                      relationship (e)          1.55
                    Ever been sexually
                      abused? (f)               1.35
                    IV drugs (g)                1.18
                    Prostitution (h)            1.59
                    Ever had an STD? (i)        1.27
Potential           Age (b)                     1.43
  Moderator         Education (b)               1.65
  Variables         Marital status (j)          1.29
                    Employment status (k)       1.65
                    Attendance at religious
                      service (b)               1.29

                                                Consistent Condom Use
Sexual Health                                         (N = 120)
Component           Variable(s)
                                                Odds       95% CI for
                                                Ratio      Odds Ratio

Cultural & sexual   Acculturation (b)            0.87      0.25, 3.09
  identity
Sexual anatomy &    Sexual difficulties (b)      1.13      0.68, 1.86
  functioning
Positive sexuality  Sexual assertiveness/
                      self efficacy (b)          0.87      0.35, 2.17
Sexual health care  Attitudes toward
  & safer sex         condoms (b)               13.81 ***  2.86, 66.58
                    Assessment of HIV risk (b)   0.87      0.49, 1.55
                    Desire for pregnancy (c)    11.39 *    1.02, 127.65
Barriers to         Alcohol and drug use (d)
  healthy            --moderate user             0.67      0.14, 3.29
  sexuality          --heavy user/in treatment   1.05      0.15, 7.59
                    Domestic abuse in current
                      or most recent
                      relationship (e)           0.28      0.02, 4.82
                    Ever been sexually
                      abused? (f)                0.48      0.11, 2.06
                    IV drugs (g)                 0.70      0.16, 2.98
                    Prostitution (h)             0.49      0.06, 4.15
                    Ever had an STD? (i)         0.42      0.09, 1.99
Potential           Age (b)                      1.00      0.91, 1.10
  Moderator         Education (b)                1.05      0.53, 2.07
  Variables         Marital status (j)           0.28      0.02, 3.65
                    Employment status (k)        0.50      0.10, 2.48
                    Attendance at religious
                      service (b)                0.61      0.31, 1.22

                                                  Composite Variable:
                                                   High Risk Sexual
                                                       Behavior
Sexual Health                                          (N = 120)
Component           Variable(s)
                                                Odds       95% CI for
                                                Ratio      Odds Ratio

Cultural & sexual   Acculturation (b)            0.97      0.42, 2.21
  identity
Sexual anatomy &    Sexual difficulties (b)      1.35 *    1.04, 1.74
  functioning
Positive sexuality  Sexual assertiveness/
                      self efficacy (b)          0.88      0.47, 1.64
Sexual health care  Attitudes toward
  & safer sex         condoms (b)                0.59      0.25, 1.36
                    Assessment of HIV risk (b)   0.94      0.68, 1.30
                    Desire for pregnancy (c)     0.33 *    0.12, 0.92
Barriers to         Alcohol and drug use (d)
  healthy            --moderate user             0.75      0.25, 2.30
  sexuality          --heavy user/in treatment   0.83      0.23, 2.94
                    Domestic abuse in current
                      or most recent
                      relationship (e)           1.02      0.31, 3.30
                    Ever been sexually
                      abused? (f)                1.27      0.50, 3.22
                    IV drugs (g)                 1.94      0.48, 7.81
                    Prostitution (h)             1.17      0.39, 3.55
                    Ever had an STD? (i)         1.42      0.59, 3.43
Potential           Age (b)                      1.03      0.98, 1.09
  Moderator         Education (b)                1.11      0.72, 1.73
  Variables         Marital status (j)           0.56      0.14, 2.14
                    Employment status (k)        1.48      0.54, 4.07
                    Attendance at religious
                      service (b)                0.96      0.65, 1.43

(a) Derived from linear regression. (b) Continuous variable.
(c) Reference category is no desire for pregnancy." (d) Reference
category is "not used in last 3 months and not currently in chemical
dependency treatment." (e) Reference category is "no domestic abuse in
recent or most current relationship." (f) Reference category is "never
was sexually abused." (g) Reference category is "no IV drug use within
the last 3 months." (h) Reference category is "no indication of
prostitution." (i) Reference category is "never had an STD."
(j) Reference category is "not currently married." (k) Reference
category is "not currently employed regularly."

* p [less than or equal to] 5.05 ** p [less than or equal to] .01
*** p [less than or equal to] .001