«Shelley Clark* McGill University Caroline Kabiru African Population and Health Research Center Eliya Zulu African Population and Health Research ...»
Differences in Men’s and Women’s Sexual and Romantic Partnerships in Kenya:
How Big is the Gender Gap?
African Population and Health Research Center
African Population and Health Research Center
*Correspondence Author: Shelley Clark, Institute for Health and Social Policy, McGill
University, 1130 Pine Ave, Montreal, QC, H3A 1A3. Phone: 514-398-8822. E-mail:
Young men and women in Kisumu, Kenya experience strikingly different patterns of risk of acquiring HIV/AIDS. These differences are undoubtedly related to differences in partnership formation, disruption, and sexual behaviors within these partnerships.
Previous research on men and women have found enormous gaps in young men’s and women’s reported levels of sexual contact, sexual frequency, and condom use; yet they have been unable to identify the source of these differences primarily due to severe data limitations. This paper draws on newly collected and exceptionally rich data containing the ten-year relationship histories of 1629 young men and women interviewed in Kisumu, Kenya during the summer of 2007. These data also contain 311 matched married and unmarried couples. Using these data, we are able to explore differences in men’s and women’s relationship history experiences, their aspirations and perceptions with respect to these relationships, and their differences in reporting within these relationships.
In Kisumu, Kenya, like in many other severely AIDS afflicted regions of Africa, a clear pattern is exhibited whereby women are becoming infected with HIV/AIDS at a much younger age than are men. HIV prevalence peaks at ages 20 to 24 for women, reaching nearly 40 percent, while for men the peak prevalence does not occur until their early thirties (Buvé et al. 2001; Clark 2004). With an estimated 85% or more of HIV infections occurring via heterosexual intercourse, differences in sexual partnerships, sexual behaviors, and biological factors are likely explain these radically different patterns. In particular, large age disparities between young women and their partners as well as potentially greater biological susceptibility among women are often cited as important contributing factors. The evidence, however, is mixed and often rather weak primarily because detailed quantitative studies of men’s and women’s sexual partnerships and behaviors between the pivotal ages of 15 and 24 are surprisingly rare and the data generated by these studies is quite thin. As a result, our ability to understand the causes of these different patterns of HIV rates and to design effective HIV interventions appropriate for both young men and women has been critically hampered.
Previous studies have been limited in at least three important ways. First, the questions asked about previous sexual partnerships are scant. In some surveys, respondents are only asked to report their total number of previous sexual partners. While some surveys gather more detailed information on up to three previous partnerships (often their current, previous and first sexual partners), they still tend to leave large holes in many respondents’ relationship histories. Second, nearly all previous studies examine sexual partnerships only. While HIV is unlikely to be transmitted via a purely romantic (i.e.
non-sexual) relationship, many studies (mainly among adolescents and young adults in the United States) demonstrate that sexual and romantic partnerships are closely entwined, and it is possible that exclusively focusing on sexual partnerships may further exacerbate reporting errors (Giordano 2003; Giordano, Longmore and Manning 2006;
Thornton 1990). Finally, studies of unmarried men’s and women’s sexual partnerships and behaviors almost universally compare reported sexual behaviors drawn from two independent samples of men and women. These studies consistently show wide discrepancies between men’s and women’s reported sexual behaviors with respect to their number of sexual partners, frequency of sexual activity and condom use—with men typically reporting much higher levels of all three measures. However, because the sexual partners of the men in the sample are not necessarily included in the sample of women in the study (or vice versa), researchers cannot determine whether these differences are due to sample selection or levels of (mis-) reporting by men and women.
Gaining a more complete and accurate picture of men’s and women’s different relationship histories and sexual behaviors is important, not only because it may shed light on the difference in the timing of exposure to HIV, because it will help illuminate gender differences in their transitions to adulthood. The primary aim of this paper is to explore differences in men’s and women’s actual relationship experiences, their aspirations and perceptions with respect to these relationships, and their differences in reporting the characteristics and sexual behaviors within their relationships. To investigate these differences we will draw on data from an innovative survey conducted in the summer of 2007 in Kisumu, Kenya. This survey gathered detailed information about all romantic and sexual partnerships of 1318 men and women aged 18 to 24 for the past ten years. In addition, it located and interviewed the recent marital and non-marital partners of 311 of the respondents.
Data and Methods:
In the summer of 2007, a team of researchers from the African Population and Health Research Center (APHRC), Brown University and McGill University interviewed a total of 1318 index respondents, which consisted of men and women aged 18 to 24 in Kisumu, Kenya. Kisumu, the third largest city in Kenya, is located in Nyanza Province which has the highest HIV/AIDS prevalence rate (13%)(National AIDS Control Council 2006).
These index respondents were randomly assigned to receive either a standard questionnaire about their previous sexual relationship in the last year or an experimental “Relationship Histories Calendar” questionnaire, which provided monthly information about all romantic and sexual partnerships occurring in the last ten years. Detailed assessments of these two methods is provided elsewhere (Luke, Clark and Zulu 2007).
The same questions were asked about each of the respondent’s previous relationships in the last year, including when the couple met, the characteristics of their partner, and the sexual as well as reproductive behaviors within each reported partnership.
In addition, all index respondents were given the opportunity to nominate one or more of their recent partners to be interviewed for the study. Respondents could only nominate partners they had been with during the course of the last year and only those who were currently living in Kisumu at the time of the study. Interviewers then sought to contact and interview all nominated partners with the assistance of the respondent. Ultimately, 311 matched partnerships were identified-- approximately half of which were unmarried.
184 of these couples received the standard questionnaire and 127 received the Relationship Histories Calendar questionnaire.
Unfortunately, as the data are in the final stages of entry and cleaning (we expect this process to be finalized in early October, 2007), we are unable to present preliminary results in this abstract. Below, however, we describe the three types of analyses we can and will pursue in the full paper.
First, using the full sample of 1629 male and female respondents, we will compare aggregate differences in men’s and women’s reported relationship histories over the previous ten years. Table 1 will examine differences by sex along the following
Table 1. Aggregate differences in men’s and women’s relationship histories.
Patterns in romantic and sexual partnerships Total number of romantic partnerships Total number of sexual partnerships Level of concurrency or serial monogamy Length of relationships Timing of relationship relative to other life events (education, employment, migration, etc) Partner characteristics Age of partner Age difference between partner and respondent Ethnicity of partner Economic status of partner Educational attainment of partner Residence of partner Marital status of partner (if not spouse) Relationship aspirations and perceptions Reasons for entering relationships Reasons for ending relationships Familiarity with partner prior to relationship Type of relationship (casual, boyfriend/girlfriend, fiancé(e), spouse) Marital aspirations (at the beginning and end of relationship) Sexual behaviors Ever had sex in relationship Interval between the beginning of relationship and first sex Frequency of sex within relationship Duration sexually active within relationship Condom use within relationship Perceptions of partners’ sexual exclusivity In the initial comparison between men’s and women’s reports about their relationships, the unit of analysis for partner characteristics, relationship aspirations, and sexual behaviors will be each relationship. Additional analyses will control for multiple relationships per respondent and examine variation in relationship characteristics and sexual behaviors by partnership type within respondents.
Our second set of analyses will assess which of the recent partnerships were nominated by the respondent and among those nominated, which were interviewed. Following each interview, the respondent was asked whether he or she would be willing to nominate one or more of his or her recent partners to be interviewed for our study. We will first determine the percentage of eligible partners who were nominated by the respondent to be interviewed and then estimate the proportion of those nominated who were ultimately interviewed. Table 2 will carefully assess biases with respect to both the nomination and interviewing of partners. For example, we will examine whether respondents are more likely to nominate their more “serious” relationships (i.e. are they more likely to nominate their fiancée than their girlfriend). We will also evaluate whether respondents involved in sexually exclusive relationships are more likely to nominate their partners.
Finally, we will examine who is more likely to be interviewed. In other words, did the probability of interviewing a particular partner depend on the type or duration of the relationship or other characteristics of the partner, like their age, marital status, and ethnicity?
After addressing these issues of selection, we will then examine the relationship histories of matched couples more closely. Below are the relationship histories of three sets of matched couples. Figure 1 shows the life-histories and relationship histories of the first matched-couple, which consists of a 27 year old Luo man and a 19 year old Luo woman.
He completed his university degree in February 2006 and is currently working as a salesman, while she is currently enrolled in a nearby university. While the man has had three previous sexual partnerships, ranging from a one-night-stand, to a dating partner, to a serious relationship, this is the first partnership of any type for the woman. The woman believes that her male partner does not have any other partners, though she admits that she really cannot be sure. The man, however, reported having an on-going relationship with a “serious girlfriend,” although the relationship is currently plutonic. For the majority of their relationship (all but the first month), the man is certain that the woman has not other partners and this is consistent with her own reports. Both members of the dyad agree that it is a non-sexual dating relationship and that the relationship is still ongoing. While the man began the relationship because he was physically attracted to the woman, she began dating him because she liked his personality. Neither member appears to be seriously considering marriage. Somewhat encouragingly, both the man and the woman report having had an HIV test shortly before their relationship began.
(insert Figure 1: Matched Couple #1 about here) In the second couple, both the Luhya man (aged 21) and the Luo woman (aged 19) report having had a total of five sexual partners each, all within the last three years and many of them concurrent. In their current matched partnership, the female considers the male to be a serious partner and she is sexually exclusive, while the male refers to the female as a causal partner and has another concurrent partner, whom he says he is dating. The woman said that she did not know whether he had other partners, while the man stated that he was certain she had no other partners. Despite multiple concurrent partners, the man has never been tested for HIV, while the woman was tested within the last six months. The man reports having slightly more regular sex and claims that condoms are always used, while the woman indicates that condoms are “mostly” used. She claims that she is in love with him, while he stated that the primary reason he is with her is because she is physically attractive. Neither member, however, reports wishing to marry this particular partner. Both the man and the woman left school before completing their secondary education and neither of them is currently working or holding a formal job.
(insert Figure 2: Matched Couple #2 about here) The final matched partnership (Figure 3) is a sexual dating relationship between a 22 year old male butcher and a 22 year old female, who works as househelp. Both members of this couple are Kikuyu. In this relationship the man reports a slightly shorter duration of sexual activity, yet both agree that they have sex about one to four times a month and always use a condom. While the butcher would have wanted to marry both of his previous partners, he does not wish to marry this particular current partner. In contrast, she would like to marry him. Finally, we find that while both the man and the woman have had previous sexual relationship, they are currently mutually sexually exclusive.
Interestingly, the woman reports that she is certain that he has no other partners, while he states that he simply doesn’t know whether she is being monogamous.