|Home | About | Journals | Submit | Contact Us | Français|
We explored links among economically motivated relationships, transactional sex, and risk behavior for human immunodeficiency virus (HIV) and sexually transmitted infections (STIs) among unmarried African American and white women.
We drew on data from 1,371 unmarried African American and white women aged 20 to 45 years that we collected via a random-digit-dial telephone survey in the U.S.
Of all respondents, 33.3% (95% confidence interval 28.8, 38.0) reported staying in a relationship longer than they wanted to because of economic considerations. African American women were more likely than white women to report starting a relationship due to economic considerations (21.6% vs. 10.5%) and having transactional sex with someone who was not a regular partner (13.1% vs. 2.9%). These behaviors were all associated with lack of education, experience of economic hardship, need to care for dependents, and increased levels of HIV/STI risk. All three behaviors were associated with having more sexual partners. Staying in a sexual relationship because of economic considerations was also associated with anal sex, reduced condom use, and concurrent sexual partnerships. Transactional sex with non-regular partners was associated with concurrent sexual partnerships, binge drinking, drug use, perceived concurrency by main partner, and having high-risk sexual partners.
HIV/STI risk-reduction policies and programs in the U.S. need to explicitly address overall economic disempowerment among women, as well as racial disparities in poverty. These economic disparities likely contribute both to increasing rates of HIV among women in the U.S. and to the extraordinary racial disparities in HIV/STI risk among American women.
The proportion of U.S. cases of acquired immunodeficiency syndrome (AIDS) that occurs among women has risen nearly fourfold during the past 20 years, from 7% in the mid-1980s to 27% in 2007. Among women in the 34 states with confidential name-based reporting, 83% of human immunodeficiency virus (HIV)/AIDS cases newly diagnosed in 2007 were attributed to high-risk heterosexual contact. African American women are by far at greatest risk: they comprise 14% of the total female population of these states, but account for 66% of HIV/AIDS cases among women.1 Understanding the reasons for the increasing rates of HIV/AIDS among women, as well as the accompanying racial disparities, is crucial to informing interventions and policy to reduce women's HIV risk in the U.S.
Recent research from settings with hyperendemic HIV, and in particular sub-Saharan Africa, has shown that economically motivated sexual relationships and encounters, often referred to as “transactional sex,” significantly contribute to women's risk of acquiring HIV through heterosexual transmission.2–14 This work has shown that economic motivations and transactions are inherent in a wide range of relationships and sexual encounters, rather than being confined to commercial sex work.2–5,7–13,15 The emerging concept of transactional sex thus occupies the broad conceptual space between sex work, on the one hand, and gifts or material support freely offered and accepted as part of courtship or ongoing expressions of love, on the other.
An increasing body of research, largely from sub-Saharan Africa, shows that while economically motivated relationships and transactional sex are tied to women's relative lack of socioeconomic resources, these behaviors are also associated with a range of partnership dynamics and sexual behaviors that may increase risk of HIV and sexually transmitted infection (STI) transmission. These behaviors include multiple sexual partners, sexual concurrency, increased coital frequency, reduced condom use, and sex while intoxicated.2,3,6,16–26 Several qualitative studies have shown that women are more likely to have sex in situations where they might otherwise refrain when financial or material gain is at stake,2,16,17,19–21,27 and that women may be fearful of refusing sex with partners on whom they rely for material support.18 Transactional sex is also often associated with alcohol use, and can lead to sex while intoxicated,22–26 which can also increase HIV risk.28 In South Africa, two recent studies have shown self-reported participation in transactional sex to be associated with HIV seropositivity,4,14 affirming the risk suggested by behavioral data and underscoring the importance to HIV/STI prevention of better understanding economically motivated relationships and transactional sex.
While extensive international data affirm the importance of economically motivated relationships and transactional sex to women's HIV risk, far less is known about the contribution of economic considerations to sexual risk among women in the U.S. and whether such considerations may contribute to racial disparities in HIV risk, as African American women experience considerably higher rates of poverty than white women.29
To address these issues, we drew on data from a nationally representative random-digit-dial (RDD) telephone household survey of unmarried African American and white women conducted from October 2006 to May 2007. Potential respondents who self-reported being female, African American or white, aged 20 to 45 years, and unmarried or not currently in any relationship equivalent to marriage were eligible for inclusion. In total, we interviewed 1,453 women (1,042 African American and 411 white women).
Sampling employed a dual-frame design, incorporating two selection stages without stratification in each frame. The larger frame was designed to provide national coverage of the eligible population (both white and African American), and was defined as all counties with any eligible household prevalence of 10% or greater as estimated from U.S. Census data. To ensure adequate representation of African American women, we sampled from a second frame targeting counties containing a high density of African American women, with an estimated eligible household prevalence of 7% or greater. Within each residential household contacted, we enumerated any female adults in the target age range, selected one via simple random sampling, and then screened for the remaining eligibility criteria. Women agreeing to participate were compensated $50 for completing the assessment.
We conducted interviews using computer-assisted telephone interviewing technology, and interview times averaged about 50 minutes. We placed calls to 81,280 residential households to identify 2,947 eligible potential respondents. Of these women, 1,453 (49%) agreed to participate after a lengthy informed consent process. This rate is comparable to similar RDD telephone surveys on sexual risk.30,31 We computed an analysis weight for each observation to reflect the dual-frame sampling design, unequal selection probabilities, and calibration to census-based estimates of the population of eligible women within the continental U.S. The Institutional Review Boards at Emory University and Macro International approved all study procedures.
Economically motivated relationships and transactional sex with casual partners. We developed an initial draft assessment of women's participation in economically motivated relationships and transactional sex with casual partners through an iterative process of consultation with researchers and health educators who specialize in transactional sex and/or women's HIV-risk behavior. We tested the resulting set of draft questions through cognitive interviews32 with 19 unmarried women (nine African American and 10 white) using 10 face-to-face interviews followed by nine telephone interviews of women recruited by RDD sampling. The cognitive interviews involved administering the questions and then asking participants about their perceived meaning of the items, their process of formulating answers, and any emotions evoked, with a particular focus on stigma and potential nonresponse. The questions were then refined and pilot-tested with 42 women from the target population in a national RDD sample. We chose the final set of questions after examining item response and correlations from the pilot, as well as debriefing interview staff regarding participants' reactions to the items.
The final set of questions addressed lifetime history of three different behaviors: staying in a relationship longer than desired (“staying in”), starting a new relationship (“starting”), and having sex with someone who was not a regular partner (“transactional sex”) for any of three reasons: (1) concern about things the participant couldn't afford; (2) concern about/desire for help with expenses such as housing, groceries, utilities, or other bills; or (3) concern about/desire for help with expenses related to children or other dependents. This resulted in nine total items (Table 1). To facilitate comparisons with other surveys, a separate item, “I have had sex with someone who was not a main partner because I needed or wanted money or drugs,” was placed in a separate, preceding section of the survey. For all items, the respondent was offered four answer choices: true, somewhat true, somewhat false, or false. These categories were chosen to allow participants space to acknowledge the complexities in motivation that often underlie choices about sex and relationships; they were also intended—and confirmed in cognitive interviews—to mitigate the stigma potentially attached to disclosure. For the purposes of these analyses, a respondent was considered to have responded positively to “staying in” or “starting” an economically motivated relationship or engaging in “transactional sex” if she answered “true” or “somewhat true” to one or more of the three available reasons.
These variables included race, marital history, current relationship, age, education, employment, household income, number of dependents, and past year economic hardship. We assessed economic hardship as a score of 0 to 5, and this variable covered the inability to afford (1) food, clothing, or shelter; (2) utilities; (3) telephone; (4) childcare; or (5) leisure activities. For this analysis, due to limited variance, the score was dichotomized, comparing a score of 0 to 2 vs. a score of 3 to 5.
These variables included lifetime history of vaginal and anal sex with men; lifetime and past year numbers of male and/or female sexual partners; and lifetime and past year concurrent sexual partnerships, assessed by asking women whether they ever or during the past year “had sex with one person while involved in any kind of sexual relationship with another person.” Substance use variables included binge drinking (past two weeks), marijuana use (past year), and injection drug use (past year). Measures of consistent condom use for vaginal sex during the past 90 days were derived by asking women how many acts of vaginal intercourse they engaged in and how many of those acts were condom protected; consistent use was defined as 100% condom coverage.
Perceived partner concurrency was defined as an affirmative response to the question, “Do you think your main partner has had sex with someone else since the beginning of your relationship?” Other perceived partner risks were assessed using the question stem, “Thinking about his whole life, do you think your main partner has ever: (a) injected any illegal drug; (b) spent more than 24 hours in jail, prison, or a detention center; (c) had sex with a man; or (d) had a sexually transmitted disease.” All items were scored on a four-point scale—definitely yes, probably yes, probably no, and definitely no—to allow women space for lack of certainty. “Definitely yes” and “probably yes” were considered affirmative responses.
We conducted analyses with STATA 10.1,33 using appropriate estimation commands for complex survey data. We first estimated the prevalence, with 95% confidence intervals (CIs), of each type of economically motivated relationship/transactional sex we queried.
We next explored two-way associations between each type of economically motivated relationship/transactional sex and potential sociodemographic predictors. Sociodemographic variables significant at p≤0.20 in two-way analyses were then tested in multivariable logistic regression models. We included race and age in all sociodemographic models; other variables were retained only if they were significant at p≤0.05. All potential interaction terms were tested; no significant interactions were observed.
We then examined associations between economically motivated relationships/transactional sex and HIV/STI risks. Sexual risk and substance use variables were treated as potential sequelae of economically motivated sex and were, therefore, modeled as dependent variables in multivariable logistic regression models, with each economically motivated sex variable as predictor. These models controlled for all significantly associated sociodemographic measures from the previous analyses and other theoretically important variables.
A total of 1,453 women completed the survey; 80 participants who had never had sex and two who provided inadequate information about economically motivated sex were dropped, leaving 1,371 participants for this analysis (995 were African American and 376 were white).
Overall, a weighted 33.3% (95% CI 28.8, 38.0) of respondents reported staying in a relationship longer than they wanted because of economic pressures, with no significant differences by race. In contrast, African American women were more likely than white women to report both starting a relationship (21.5% vs. 10.5%) and having transactional sex with someone who was not a regular partner (13.1% vs. 2.9%) (Table 1).
Only 4.4% (95% CI 3.1, 6.3) of all participants responded “true” or “somewhat true” when asked about sex for money or drugs. Overall, “sex for money or drugs” was reported by only 42.1% of women who met our definition of transactional sex with casual partners, 10.1% of women who reported staying in relationships, and 15.9% who started relationships for economic reasons. In contrast, 83.3% (95% CI 61.5, 94.0) of women who reported “sex for money or drugs” reported at least one other type of economically motivated or transactional sex. The majority of women who reported transactional sex with a casual partner also reported staying in (76.3%) or starting (55.4%) a relationship for economic reasons. Among women who reported starting a relationship for economic reasons, 67.4% also reported staying in a relationship for economic reasons, while 26.5% reported transactional sex. Among women who reported staying in a relationship for economic reasons, 29.3% also reported starting a relationship for economic reasons, and 15.7% reported transactional sex. Overall, 39.7% of participants reported at least one of the three economically motivated or transactional sex behaviors we assessed, with 28.0% reporting one, 8.1% reporting two, and 3.6% reporting all three behaviors.
Women who had previously been married were more likely to report staying in a relationship for financial reasons, while those with a current main sexual partner were more likely to report starting relationships or having transactional sex with non-regular partners for economic reasons (Table 2). Older women were more likely to report ever having engaged in all three economically motivated relationships/transactional sex behaviors we assessed, with significant differences for staying in relationships and transactional sex. Women with a high school education or less, with yearly household incomes less than $25,000, or with higher economic hardship scores were significantly more likely to report all three types of economically motivated relationships/transactional sex. An increasing number of financial dependents was correlated with all three behaviors and was significant for both staying in relationships and transactional sex.
In multivariable logistic regression modeling (Table 3), race remained a significant predictor of starting relationships for economic reasons and transactional sex with non-regular partners, while the effects of age were largely tempered. Women who had ever been married were more likely to report staying in relationships for economic reasons, while those with current partners were more likely to report starting them. Economic hardship remained strongly predictive of starting and staying in relationships for economic reasons, as well as transactional sex. Transactional sex was also associated with low education. In the final models, the number of dependents only predicted staying in relationships for economic reasons.
All three measures of economically motivated relationships and transactional sex were associated with increased partner numbers (Table 4). Staying in relationships for economic reasons and transactional sex predicted an increased number of lifetime partners, while starting relationships for economic reasons and transactional sex predicted having two or more partners in the past year. A history of staying in relationships for economic reasons was also associated with sexual concurrency (lifetime), anal sex (lifetime), and significantly reduced condom use with both main partners and all partners during the last 90 days. Transactional sex with a non-regular partner was associated with both lifetime and past year concurrency, binge drinking in the past two weeks, past year marijuana use, and past year injection drug use. For women with current partners, it was also associated with perceived partner concurrency, and all other partner risks assessed: jail, drug use, sex with other men, and STI history.
We assessed prevalence of economically motivated relationships and transactional sex among a national telephone sample of unmarried African American and white women aged 20 to 45 years in the U.S. One-third of respondents reported staying in a relationship longer than they wanted to for economic reasons, with no differences by race. In addition, 22% of African American women and 11% of white women reported starting a relationship in response to economic concerns, while 13% of African American and 3% of white women reported transactional sex with a non-regular partner. All three of these behaviors were linked to increased partner numbers, while staying in relationships for economic reasons and engaging in transactional sex with casual partners were both associated with a wide range of other HIV/STI risks. These findings thus suggest some important mechanisms through which poverty and women's economic disempowerment may contribute to both increasing rates of HIV among women and racial disparities in the U.S. HIV epidemic.
While economically motivated and transactional sex are well-established contributors to women's HIV risk in other cultural contexts, consideration of economically motivated sex in U.S. populations has largely been confined to enquiries on “sex for money or drugs,” often with a focus on women who use substances.34–37 Our findings affirm that economically motivated relationships and transactional sex among U.S. women cover a much broader range of behaviors. While most women in our sample who reported sex for money or drugs also reported engaging in economically motivated relationships or transactional sex, less than one-tenth of women responding affirmatively to at least one of our question sets reported sex for money or drugs. This finding suggests that the questions presented in our survey captured an overlapping, but considerably broader, set of behaviors linking economic pressure to sexual decision-making and HIV/STI risk among U.S. women.
All three types of economically motivated and transactional sex assessed in this study were associated with increased numbers of sexual partners. While women in many settings proactively seek multiple partners to meet their needs or desires for material support,2,9,13,38,39 the power dynamics inherent in economically asymmetric relationships may further increase partner numbers. Women in qualitative research from sub-Saharan Africa reported that they are more likely to have sex in situations where they might otherwise refrain when material gain is at stake,2,16,17,19–21,27 and can be fearful of refusing sex with men on whom they rely for material support.18 Recent work in the U.S. has highlighted how women living in low-income housing, who are more likely to be African American, are vulnerable to demands for sex from landlords. Such economically exploitive sexual harassment may also be common in other spheres.40
Multiple sexual partnerships are often concurrent. In our sample, women who had stayed in relationships for economic reasons and women who reported transactional sex with non-regular partners were more likely to report engaging in sexual concurrency. This concurrency may represent women seeking emotional connection from another partner while retaining material support from a primary partner. It may also represent seeking partners for material support while having a primary emotional and sexual relationship with a primary partner who lacks resources, or maintaining multiple partners who provide different types and levels of material support.2,13,20 Sexual concurrency can increase rates of HIV transmission as larger numbers of people are sexually connected at any given time.41
Staying in relationships for economic reasons was also associated with significantly lower rates of condom use among women who had had sex in the past 90 days. This finding is consistent with data from qualitative research in sub-Saharan African, where women reported that they are less likely to be able to successfully request condom use with transactional partners,2,3,16–8,27,21 and with data from a recent study in Kenya, which found that the likelihood of a man reporting condom use at last sex was inversely associated with the amount of money he provided to his female partner.6
Among women with current main partners, transactional sex with non-regular partners was also associated with perceived concurrency by the partner, as well as other partner HIV/STI risks. Research in Kenya and South Africa has found that men who give resources to female partners are more likely to report multiple and concurrent partnerships,6,42 while studies in other African countries affirm that men often use transactional sex to secure multiple partners.2,10,17,43 Research in South Africa has further shown that men who participate in transactional sex are more likely to report alcohol use problems, as well as perpetration of intimate partner violence and sexual assault.42
Transactional sex with non-regular partners was associated with binge drinking and drug use among women in our sample. Data from other settings suggest that transactional sex is often associated with alcohol use, and can be associated with having sex while intoxicated.22–26 Although our study modeled substances as possible sequelae of transactional sex, we acknowledge that the relationship is likely to be more complex and to involve reciprocal causality. Substance use is known to place women at risk for engaging in sex for money or drugs as it contributes both to women's poverty and to addiction.44,45 However, substance use may also be a strategy for coping with stress associated with such sex work encounters.44 Understanding the extent to which either or both of these dynamics are at play with regard to transactional sex as defined in this article will require further research.
A crucial question is whether economically motivated relationships and transactional sex may contribute to the increasing proportion of HIV infections occurring among women in the U.S. and the dramatic disparity in HIV prevalence between African American and white women. One in three women in this study reported staying with partners longer than desired for economic reasons, with no differences by race. Staying with partners for economic reasons was, however, linked to past year economic hardship and number of dependents. Given both the high reported prevalence of staying with partners for economic reasons and the link between this behavior and reduced condom use, lifetime partner numbers, concurrency, and anal sex, it seems quite possible that economic dependence on partners contributes to the increasing rates of HIV among women in the U.S.
In contrast to staying in established relationships, transactional sex with non-regular partners was approximately four times more likely to be reported by African American than by white women, a difference that persisted after controlling for socioeconomic variables, including significant associations with education and recent economic hardship. It is quite possible that African American identity serves as a marker for other poverty or wealth indicators not measured. For example, African Americans overall have less access to resources through intergenerational transmission of wealth than white Americans,46 and are more likely to live in segregated neighborhoods with higher poverty levels.47 African American women may, therefore, have less access to financial and material resources through their familial and social networks than white women, fostering increased reliance on sexual partners. It is also possible that cultural norms within some African American women's social and sexual networks may be more accepting of economically motivated and transactional sexual relationships. It is equally possible that such norms (if they exist) are a function of social class in addition to, or instead of, race,48 and that African American women are simply more likely than white women to be poor or working class.29 Further research with racially and economically diverse U.S. women is required to understand these issues.
The most significant limitation of this study was the relatively low response rate, which may have introduced response bias into our findings. Low response rates are a known problem with telephone surveys in the U.S., especially as intercept devices become more common and more households drop landlines in favor of cellular phones.49 Our ability to recruit participants was further restricted by our stringent eligibility criteria, the 50-minute length of our assessment, and the sensitive subject matter (fully disclosed during our informed consent process). Given these constraints, we believe that our response rates compare favorably to other recent telephone surveys on sexuality.30,31
Our relatively small sample size limited both precision and statistical power in analyses. It is possible that women with an interest in the study topic, or with a stronger desire for the $50 incentive, are overrepresented in our sample; it is impossible to know whether or how such potential overrepresentation might have biased the results of this analysis.
Other limitations include potential underreporting of stigmatized behaviors such as sexual risk-taking, although we attempted to mitigate this through our extensive cognitive testing process and careful training of field staff. Our cross-sectional data also limit our ability to draw causal inferences about the relationship between economically motivated/transactional sex and other variables studied. We acknowledge that research on economically motivated relationships and transactional sex is very much an evolving field, and that other scholars may prefer different terminology than we have chosen to describe these behaviors. We have primarily contextualized our findings with reference to research from settings with hyperendemic HIV, notably sub-Saharan Africa, where the majority of existing research on heterosexual transactional sex has been conducted. While we believe that scholarship on the dynamics of heterosexual HIV risk is often valid cross-culturally, we acknowledge that additional research in the U.S. will be required to fully understand the dynamics of economically motivated relationships and transactional sex, and their links to HIV/STI risk in this country.
Staying in sexual relationships for economic reasons and having transactional sex with non-regular partners are common among U.S. women and extend broadly beyond commercial sex work and sex for money or drugs. These behaviors are associated with lack of education, economic hardship, need to care for dependents, and increased levels of HIV/STI risk. HIV/STI risk-reduction policies and programs in the U.S. need to explicitly address overall economic disempowerment among women, as well as racial disparities in poverty. These economic disparities may well contribute to both the increasing rates of HIV among women in the U.S. and to the extraordinary racial disparities in HIV/STI risk among American women.
This project was funded by the National Institute of Child Health and Human Development grant R01 HD041716-01A1. The authors thank Bill Sribney for providing statistical support and performing data weighting, and Elizabeth Nesoff for support with preparing the article. Additional support was offered by the Emory Center for AIDS Research (P30 AI050409).