|Home | About | Journals | Submit | Contact Us | Français|
The objectives of this paper are to (1) understand the nature of men’s extramarital sexuality in three low income communities in Mumbai, India; (2) explore the associations between marital relationships and extramarital sex; and (3) assess the implications of the research results for intervention. Results are based on survey data collected from 2,408 randomly selected men from the three study communities and a matched subset of 260 randomly selected men and their wives who responded to a female version of the men's survey. These surveys produced a unique data set, which allows sociodemographic, attitudinal and behavioral variables from husband and wife and variables that are the product of husband and wife interaction to be utilized to predict men's extramarital sex through multiple sequential logistic regression analysis. Results indicate that men's extramarital sex is significantly associated with husband's and wife's age, wife's perception of domestic violence, husband's education and place of birth, husband's alcohol use, wife's willingness to engage in marital sex, and types of marital sexual acts. These results confirm the need to move from the individual to the couple as the unit of research and the need for intervention to reduce the risk of HIV/STI transmission within marriage both in India and internationally.
Unprotected extramarital sex has become a great concern for those involved in prevention of HIV/AIDS. First, unsafe extramarital sexual liaisons put the involved individual at risk for HIV/STI transmission; secondly, they put the other member of the couple at risk as well; and thirdly extramarital sex often impacts on the quality of the couple relationship, creating difficulties in communication, sexuality, and violence that may further exacerbate extramarital sex and risk of infection. This paper assesses extramarital sex, marital relationships and sexual risk in three urban poor communities in Mumbai (Bombay), India. The objectives of this paper are to: (1) understand the nature of men's extramarital sexuality in these communities; (2) explore the associations between marital relationships and extramarital sex; and (3) assess the implications of the research results for intervention.
One indicator of the impact of unprotected extramarital sex on the HIV epidemic is the growing rate of infection among women. Over the past decade there has been a dramatic shift in gender ratio so that currently almost half of the 40 million persons worldwide with AIDS are women.1,2 This global epidemiological picture is being replicated in India; as the epidemic accelerated in the early 1990s, gender distribution estimates for 1994 indicated a male to female ratio of 5:1, with female cases mostly among sex workers.3 By the end of the decade, the epidemic had crossed over to the general population and gender ratios had reduced to a 3:1 ratio with current estimates at 1.7:1.4–7 Women currently represent an estimated 40% of the 5.1 million people aged 15–49 living with HIV/AIDS in India.4,7
Men's extramarital sexual behavior is the most important factor for introducing HIV/STI into marital life.6,8–12 Husbands' extramarital sex has been seen as closely linked with the marital relationship and marital satisfaction for both husband and wife and male sexual satisfaction within marriage.11,13–17 Married men who report marital sexual dissatisfaction refer to their need for sexual excitement, sexual curiosity, novelty or variety, and sexual enjoyment as justifications for extramarital sex.16 Such justifications are exacerbated by husbands' sense of hyper-masculinity or “real manhood” that argues for their perceived “natural ability” to have continued access to multiple sexual partners and their perceived “natural need” for frequent sexual satisfaction.18,19
Although marital relationship and sexual satisfaction are important correlates of extramarital sex, gender-based violence also seems to be both a cause and a consequence of extramarital sex. Xu and associates found that a significantly higher number of abusive husbands had extramarital affairs and such affairs significantly predicted their wives being physically or sexually abused, or both, in their lifetime and in the previous year.20 Fear of domestic violence among women is a major barrier to control over their own sexuality and their husbands' sexual activity outside of marriage.5,21 As a result, wives are exposed to unsafe marital sex22,23 poor marital communication about sexual risk and sexuality, limited capacity for wives to refuse a husband's demand for sex, and limited use of condoms in marital sex as protection against disease transmission.5,24,25
Much of the collection of data on extramarital sexuality has been done from the perspective of one or the other member of the marital dyad. Our approach in both our methodology and the analyses presented in this paper was to examine extramarital sex in the context of and in interaction with the marital relationship. This approach involves inclusion of both the individual characteristics, attitudes and behaviors of each of the spouses and the junction of those characteristics to create a set of spousal variables that only exist in interaction. We hypothesize in this paper that husband, wife and interactive variables will each contribute to understanding the correlates of risky extramarital sex and the ways to reduce that risk.
The data presented in this paper is drawn from projects involving the collaboration of the University of Connecticut School of Medicine, the Institute for Community Research in Connecticut and the International Institute of Population Sciences (IIPS) in Mumbai. This collaboration led to the establishment of the program Research & Intervention in Sexual Health: Theory to Action (RISHTA, an acronym meaning “relationship” in Hindi and Urdu) at IIPS. The RISHTA program consists of two projects: (1) a 5-year research, intervention and evaluation project designed to involve men in urban poor communities in prevention and early treatment of HIV/STI (Men's Project); and (2) a pilot project to examine women's risk of HIV/STI within marriage (Women's Project). This paper reports on data drawn from the formative research phase of both the men's and women's projects.
The three study communities represent the typical economically marginal or as they are locally referred to, “slum” communities in Mumbai, a city with an estimated population of 17 million. The three communities have grown rapidly over the last two decades and now have an estimated population of 700,000; a majority of the population having migrated from poor, rural communities in the north and south of India. Residents of these communities are mixed Muslim and Hindu.
Formative research in these projects involved the collection of both qualitative in-depth interviews and quantitative survey data. This paper will focus primarily on the quantitative survey data. As a part of the Men's Project, a baseline survey was conducted from June to September 2003 with married men ages 21–40. Sample selection involved identification of all “lanes” or “plots” in the community, a random selection of these units for household sampling and a systematic random sample of households. If the household did not contain an eligible man then the next household was selected; if the household contained more than one eligible man then the individual was randomly selected. All surveys were conducted in Mumbaia Hindi in a face-to-face interview format by male RISHTA staff and were carried out in the household or in nearby areas if privacy could not be maintained in the household. The final total sample was 2,408 married men (approximately 800 men in each of the three communities). All men interviewed provided written consent, with 8% of men refusing the interview and/or the consent form.
As a part of the Women's Project, married women were chosen utilizing a component of the survey sample of males. A random sub-sample of men (n = 311) whose wives were living in the household was selected and men were asked for their verbal consent to have their wives interviewed for the women's project. When these men agreed, their wives were recruited into the sample after they gave their written consent. Of the 311 women who were contacted, 9.3% refused to be interviewed and 7.1% had family members who refused to allow them to be interviewed. The women's project survey instrument (with many overlapping items with the men's survey) was administered from February–June 2004 by female RISHTA staff to a final sample of 260 married women. Women were interviewed in Mumbaia Hindi in their homes at times when husband and children were not present; when privacy could not be maintained, a second visit was scheduled to complete the instrument.
In this paper, univariate analyses (frequencies, means) describing the socio-demographic and behavioral characteristics of men and the household and men's extramarital sexual behavior will be derived from the men's survey data (N = 2,408). A comparison of husbands' and wives' and women's characteristics will be derived from the dyadic data (N = 260). Multiple sequential logistic regression analyses focus on the dyadic data (N = 260) with extramarital sex as the dependent variable. The sequential procedure allowed each variable entered to capture any variance in the dependent measure, thereby addressing issues of multi-collinearity. A set of paired variables was compiled, each pair containing both the husband's as well as the wife's response to a particular item (e.g., age of husband and wife). Each set of paired variables was entered in sequential analysis. The second step in each analysis was to assess the additional effect of the interaction of the respective dyadic variables. A regression was run separately for each pair of variables plus the pair's interaction.
From this set of candidate variables, a number were selected for inclusion into a single regression. If the interaction of a particular pair of predictors showed at least a trend toward significance, then the two main effects as well as the interaction variable were entered into the regression equation. If the interaction did not meet criteria for a given pair, then each of the two main effects were retained, respectively, if they showed at least a trend toward significance. The final model was fit allowing for inclusion for interactions and main effects if they showed a trend toward significance. However, the analyses required the inclusion of main effects, regardless of their significance, if the interaction of the pair met criteria. In these analyses, the χ2 for the inclusion of the interaction step was used to judge its contribution to the model, while the two Wald tests from the initial two-variable model step were used to judge the contribution of the two main effects. All quantitative analyses were conducted using SPSS 12.0.26
Table 1 shows that Muslims are in the slight majority over Hindus in the study communities with a predominance of migrants from Uttar Pradesh and other northern states. Household income is low making poverty and financial shortage as major elements in the lives of residents. The predominance of marriages in these communities are arranged, which means that most wives and husbands come together as virtual strangers. This emotional distance can be exacerbated by the fact that a husband who migrates to Mumbai for work, most frequently leaves his wife in the native village with his parents until he finally settles in Mumbai, seeing his wife only when he returns periodically. At the time of the survey 16% of married men were living without their wives for a mean of 8.6 months. Once women join their husbands, the opportunities for increased marital intimacy is limited by other household members (siblings, children) and extremely limited residential space. Households consist primarily of one crowded room with little opportunity for intimacy. Nuclear households are most common, followed by joint and extended and finally by households consisting of either husband only or husband plus same age male relatives, friends and co-workers sharing a single residence.
Slightly over a fifth of the men reported that they had ever had extramarital sex and slightly over a tenth reported that they had extramarital sex in the last year. The great majority of these men reported having extramarital sex at least one time with a woman who was not a sex worker, about one third with a female sex worker and a small number with a man (see Table 2).
Women whose husbands reported involvement in pre- and extramarital sex showed little knowledge of these sexual liaisons (see Table 3).
The dependent variable chosen for the multivariate analyses in this paper is men's extramarital sex within the 12 months previous to the time of survey administration. This dichotomous variable (presence/absence) is a composite of: (1) a report of extramarital sex in the last 12 months, (2) number of partners who are sex workers, (3) number of female and male non-sex workers, and (4) oral sex, anal sex and vaginal sex with any other person than wife; a positive on any of these items, indicated the presence of extramarital sex.
Key predictor variables for both men and women included: socio-demographic characteristics (age, migration, education, age at marriage and the structure of the household); husband–wife interaction (love v. arranged marriage, division of household labor, communication and domestic violence); self-perception (as a spouse, a sex partner, and a global assessment of the quality of one's life); and the nature of marital sex (frequency of marital sex in the last 30 days, the number and types of sexual acts in the last marital sex, degree of satisfaction in last marital sex, wife's willingness to be involved in last marital sex). In addition, scales of masculinity and women's empowerment were constructed that were based on in-depth interviews on these topics. The scales covered comparable gender equitable and dis-equitable items with specific questions for men on peer-related avtivities and for women on the degree of control over their sexuality.
These variables were hypothesized as candidate predictors for men's extramarital sex. The results of the sequential multiple logistic regression in Table 4 identified a series of demographic factors, husband perceptions, wife's perceptions and husbands behavior as reported by both wife and husband as associated with men's extramarital sex.
In terms of demographic variables, the younger age of husbands and wives (a trend) are associated with the greater likelihood of men's extramarital sex. Greater age discordance between husbands and wives is associated with men's extramarital sex. Men born in Mumbai are more likely to engage in extramarital sex than those who migrated to Mumbai either with their families or on their own. The final demographic variable involves men's education; those with lower education are more likely to have engaged in extramarital sex.
Men who consumed alcohol on a daily or weekly basis were more likely to have extramarital sex. Husbands' reports of lower numbers of intimate pre-penetration sexual acts in last marital sex show a relationship to having extramarital sex. Husbands' reports that their wives were less willing (requiring coercion) to participate in their last marital sex shows a trend toward significance in association with the likelihood of extramarital sex.
Wives' perception of themselves as less adequate wives (according to local social norms) is associated with men's extramarital sex. Finally, wives' reports of husband's domestic violence are significantly associated with the greater likelihood of men's extramarital sex.
Although it is difficult to distinguish the causes from the consequences of extramarital sex, there is clear cross-cultural evidence that it has a negative impact on the marital dynamic.13 Extramarital sex increases the risk of HIV and STI transmission for husbands and wives and makes the marital unit a location for conflict, propelling men to seek further sexual liaisons outside marriage and thereby creating a marital environment of further risk.
The frequency of ever having extramarital sex in the study communities puts it at the lower to medium point among the 14 countries assessed in a WHO study.27 The frequency drops when men were asked about extramarital sex in the last year, with the majority having a sexual liaison with women who were not sex workers, while a significant minority is having sex with a sex worker. Results show irregular use of condoms with sex workers, that most men perceive low risk with women who are not sex workers, and little or no risk in male to male sex. In summary, extramarital sexuality in these communities is risky sex. Wives seem to know little about husbands' pre- and extramarital liaisons, reducing their ability and opportunity to influence husbands or advocate for safer marital sex.
The multivariate analysis that was conducted examined the demographic characteristics, perceptions and behavior of wives and husbands as they related to extramarital sex. The results support the hypothesis that it is necessary to have data on both members of a marital dyad to fully characterize couple relationships and assess associations with phenomena such as men's extramarital sex.28–30 Five husband variables and four wife variables were found to be significantly related to men's extramarital sex in the multiple logistic regression. Of the demographic variables, younger couples were associated with men's extramarital sex. The greater discordance between husband's and wife's age was also associated with men's extramarital sex; men who were significantly older than their wives (22.6% are older by 6 years or more) were more likely to have extramarital sex. Age discordance primarily reflects the marriage of older men to younger women (23.2% were 15 years of age at marriage or younger) who may not be prepared or mature enough for a marital and/or a sexual relationship.
Those men who consume alcohol daily or weekly are more likely to be involved in extramarital sex. Alcohol is intimately bound-up with extramarital sex. Places to drink including “beer bars” and illegal brewers of country liquor (desi daaru) are also locations to find women available for sex. Many illegal brewers and/or sellers are widowed or divorced women who provide sexual favors in association with providing alcohol. Alcohol with peers is seen as a “dis-inhibitor” and a preferred step toward having extramarital sex.
Two additional men's variables showed a significantly different direction than has been reported in the literature. Men who had less education and men born in Mumbai and were not migrants were more likely to be involved in extramarital sex. Lower education may be associated with men who have low status occupations that provide opportunities for sexual liaisons such as construction, day laborers and truck loaders. Migrants to Mumbai, including those who have left their wives and families in the rural area and are living with other single men do not show a greater degree of extramarital sexual liaisons. The results suggest that men born in Mumbai are more familiar with the environment and have a peer network that facilitates relationships with women in and out of their communities.
Husbands' reports of the nature of marital sexuality are also associated with extramarital sex. Those husbands who report that their wives were less willing to participate in the last marital sex (indicating coercion) showed a trend in which they were more likely to be involved in extramarital sex. In addition, husbands who reported less pre-coital acts in the last marital sex (another indicator of coercion) were more likely to be involved in extramarital sex. In-depth interviews with a subset of women focused on the fact that their husbands did not approach them with love (pyaar karma), did not “get them ready” for sex and coerced them, both physically and/or mentally if they indicated a lack of interest. Fifty-seven percent of men indicated that their wives could not refuse sex when they demanded it.
The final set of variables focuses on wives reporting on their husband's level of violence, with those women reporting greater domestic violence being associated with their husbands' extramarital sex. It is interesting to note that men's reports of violence toward their wives were less forthcoming than women's reports and as a result, men's reports did not show a significant relationship to extramarital sex.
Finally, those women that had a more negative assessment of themselves as spouses had husbands who were more significantly involved in extramarital sex. Our qualitative data provides long narratives of couples who have had difficult relationships that started with the first night after marriage (frequently the day they first meet). As a result, a subset of women describes their husbands as critical and emotionally abusive, leading to a poor perception of themselves as wives.
Unsafe extramarital sex brings with it the risk of HIV and other sexually transmitted infections for men in these communities. It also increases the risk that they will bring these infections to their wives. In addition extramarital sex impacts on and is an indicator of difficulties in marriage that are manifested in social, economic, psychological and sexual interaction between spouses. It is insufficient to simply address men and warn them of the risks of extramarital sex. This issue must be addressed within the context of the marital relationship: first to make women aware that their monogamy does not necessarily protect them from risk and second to improve the marital relationship so that it can serve to prevent husbands' unsafe extramarital sex in the future.
There is a need to target the couple as a unit of both research and intervention to promote the reduction of extramarital sex and its associated HIV/STI risk and to assist couples whose marital relationships had been severely impacted by extramarital sex. Couple-based interventions need to be conducted by skilled male and/or female facilitators (as appropriate), and should involve both partners for several sessions that address sexual relationships and intimacy, gender-based inequities and discrimination, problem-solving skills and strategies, negotiation and communication through couple-communication exercises, and development of trust and confidence within the couple.31–33 The advantage of the couple, rather than the individual, as the unit of intervention is that it allows the dyad to jointly address sensitive issues, which protects women against the risk of violence, stigmatization and separation that is often caused by behavior change introduced by women unilaterally.34,35 Instead of considering men and women as separate targets, research and intervention programs need to address the relationship dynamics, sexual health needs and concerns of the couple, both in India and other parts of the world.
The research on which this paper is based was funded by the National Institute of Mental Health, “Male Sexual Concerns and Prevention of HIV/STD in India” (RO1 MH 64875) and a supplement, “Assessing Women’s HIV/STD Risk in Marriage in India” funded by the Office of AIDS Research of the National Institutes of Health. We thank the field and office staff of the RISHTA program at the International Institute for Population Sciences for their hard work and commitment to the collection of quality data under difficult circumstances.
Schensul, Mekki-Berrada, Burleson, and Bojko are with the Center for International Community Health Studies, Department of Community Medicine, University of Connecticut School of Medicine, 263 Farmington Ave., Farmington, CT 06030-6325, USA; Nastasi is with the Institute for Community Research, Hartford, CT, USA; Singh is with the International Institute for Population Studies, Mumbai, India.