Advancing Sexual and Reproductive Health and Rights
 
International Family Planning Perspectives
Volume 24, Number 1, March 1998

Sexual Activity and Condom Use in Lusaka, Zambia

By Sohail Agha

Context: Condom social marketing is among the AIDS prevention strategies being attempted in Lusaka, Zambia, a country where women generally are of low socioeconomic status relative to men, and where the prevalence of sexually transmitted diseases is high.

Methods: The 1996 Lusaka Sexual Behavior and Condom Use Survey gathered data on sexual activity in Lusaka from 806 respondents; multiple regression analysis was performed to identify factors that predicted men's and women's condom use.

Results: Most respondents reported that their most recent intercourse was with their marital partner (62% of women and 43% of men) or with a regular partner (20% of women and 23% of men); almost one-quarter of men (24%), however, reported having last had intercourse with a casual partner. Overall, 17% of women and 24% of men had used a condom at last intercourse. Multivariate logistic regression analysis found that women whose last intercourse was with a regular or casual partner rather than a marital partner were significantly more likely (odds ratios, 2.8 and 3.1, respectively) to have used a condom, as were those who recalled hearing a social marketing advertising message for condoms (2.8). For men, use of condoms was elevated among those who were younger than 30 (odds ratios, 3.3-3.8), who had an education beyond the secondary level (2.2) and who had easy access to condoms (1.9).

Conclusions: Because of gender inequity, programs directed at men are more likely to succeed in encouraging condom use than are those aimed at developing women's skills in negotiating condom use. =paragraph

The spread of HIV infection has been rapid in Zambia, a country of about 9.2 million people. It has been estimated that between 40,000 and 50,000 people died of AIDS-related causes in Zambia in 1993; this number was expected to have doubled by 1997.1 Together with concerns about population growth, the AIDS epidemic has led to significant changes in Zambian population and health policies. Indeed, the social and public policy environment is markedly different from that of the 1960s and 1970s, when attempts to promote contraception in Zambia were met with substantial opposition.2

Since 1986, a wide range of approaches has been used to raise awareness of AIDS in Zambia. These efforts include the formation of the National AIDS Prevention and Control Program in 1988 and the launching of a number of projects to educate the public about the threat of AIDS.3 In 1989, the government adopted a strong position in favor of contraceptive use. In 1992, Population Services International, in collaboration with the Pharmaceutical Society of Zambia, began condom social marketing activities. The project marketed a packet of four condoms under the brand name "Maximum." These efforts were funded by the U.S. Agency for International Development and the Government of Japan as part of their assistance to the Zambian Ministry of Health. Making inexpensive condoms widely available and promoting their use through mass media advertising and peer education campaigns has been the project's major emphasis.

National educational efforts have raised the public's awareness of AIDS: Surveys show that knowledge of the syndrome's existence was widespread in Zambia by the early 1990s.4 There is some evidence of a shift in attitudes as well. After campaigns to promote preventive behaviors led to a public debate, religious organizations ceased their criticism of condom promotion.5 In addition, news media reports have noted that significant changes in sexual behavior and lifestyles have occurred as a consequence of the AIDS epidemic.6

However, the extent and nature of these changes remain unknown. In particular, how these changes have been accommodated within the context of male and female gender roles remains unexplored. Indeed, whether women can enforce condom use in their sexual relationships is a subject of considerable importance to HIV prevention programs.

In any society, sexual relations occur in the context of social relations, structural factors (such as poverty, women's status and economic underdevelopment) and social factors (such as the sex ratio, migration and the availability of social services).7 Gender and power relations between men and women influence sexual behavior within the overall structural and social conditions.8

It has been argued that economically vulnerable women who have a low sense of empowerment are unable to negotiate safer sex practices such as condom use.9 When women are likely to be stigmatized or viewed negatively by their partners for suggesting condom use and when insistence on condom use may lead to instability or dissolution of relationships,10 it is unlikely that women will be able to exert direct influence on their partner's condom use. However, in contexts where women are not economically vulnerable and have greater ability to negotiate safer sex, women are able to insist on condom use: Studies in the United States have shown that women's individual characteristics such as age and education exert a direct influence on condom use.11

In Lusaka, the basic amenities of life, such as housing, have traditionally been available to women only through their husbands or fathers. Even in present-day Zambia, employed women who are single heads of households have no claim to housing in their own right.12 The persistence of these regulations after independence in 196413 suggests that gender inequity in Zambia has remained fundamentally unaltered.

Women continue to have far fewer formal-sector employment opportunities available to them than do men. Only 15% of Zambian women are employed in the formal sector, compared with 85% of men.14 Urban women remain at a particular disadvantage. Compared with women in rural areas and with men in both rural and urban areas, urban women have higher unemployment rates and lower labor force participation rates.15

While the basic social organization of urban Zambia has not changed radically, the AIDS epidemic itself has changed relations between men and women. Assessing the extent to which sexual behavior and condom use have changed in response to the epidemic, in light of the underlying gender issues that affect this behavior, is crucial to refining the strategies of ongoing and future HIV prevention efforts.

Using data from the 1996 Lusaka Sexual Behavior and Condom Use Survey, we explore the effects of heightened AIDS awareness, greater societal openness towards condom use and the widespread availability of condoms on the sexual behavior and safer sex practices of Zambian men and women. The survey was conducted by Population Services International as part of a midterm assessment of the performance of the Zambia Social Marketing Project.

DATA AND METHODS

The 1996 survey data were collected using a two-stage probability sample of 806 men and women aged 15-49. The sample was drawn by the Central Statistical Office of Zambia. One hundred standard enumeration areas were randomly selected. In each area, a male or female interviewer administered the survey to four respondents of the same sex. The survey questionnaire was adapted from the Partner Relations Survey and the Knowledge Attitude Behavior Practices Survey, which were conducted by World Health Organization (WHO) Global Programme on AIDS. These surveys were conducted in 18 developing countries from 1989 to 1993 as part of WHO's efforts to document patterns of sexual behavior in the developing world.

The questionnaire included items on demographic characteristics, relationship with last sexual partner (marital, regular or casual), the number of sexual partners in the last 12 months, condom use at last intercourse, the condom brand used at last intercourse, recall of condom brand advertising message and access to condoms (as defined by travel time to a source). As with the WHO surveys, respondents who reported having had sex with someone other than their marital or regular partner in the last 12 months were asked how many persons they had sex with during this period. The question took into account polygynous unions among men. We compared our results with data from the 1990 WHO survey of Lusaka.16

Statistical analysis was performed using the SPSS 6.1 software program. Chi-square tests of independence were conducted at the bivariate level, and differences were determined to be statistically significant at p<.05. At the multivariate level, logistic regression analysis was used to measure the net effect of independent variables on the likelihood of condom use at last intercourse.17 (Because the sample size was small, we also report differences at p<.10 for the multivariate analyses.)

RESULTS

There were no significant differences in the age distribution of men and women in the sample (Table 1). Nor did men and women differ significantly in their levels of schooling. A significantly greater proportion of women were married than were men, a difference attributable to women's younger age at first marriage in Zambia. The respondent's relationship with his or her last sexual partner also varied significantly by sex: Men were almost one-third less likely than women to have had last intercourse with their marital partner. Men were more likely than women to recall the Maximum brand advertising message ("Strong for Maximum protection, sensitive for Maximum pleasure"). Men were also more likely than women to report being within a 10-minute walk of a source of condoms.

Table 1. Percentage distribution of respondents, and percentage who reported condom use at last intercourse, by sex and background characteristics, 1996 Lusaka Sexual Behavior and Condom Use Survey
Characteristic % distribution % using condom at last intercourse
Women
(N=391)
Men
(N=415)
Women
(N=341)
Men
(N=370)
Total 100.0 100.0 17.1 23.9
Age
15-19 29.4 26.9 27.0 34.7
20-24 23.3 18.3 20.7 35.7
25-29 15.1 17.1 11.9 34.3
30-34 12.3 13.5 8.3 7.3
35-39 10.5 9.9 12.2 12.2
40-49 9.4 14.5 16.2 9.8
Partnership status
Married 57.2 45.7 na na
Unmarried, sexually active
No current partner 10.5 21.9 na na
Regular partner 19.5 21.4 na na
Has never had sex 12.8 11.1 na na
Schooling
None/any primary 45.6 40.4 13.6 18.2
Any secondary 34.9 34.9 17.5 25.4
>secondary 19.5 24.8 25.0 29.9
Relationship with last partner
Marital 61.8 42.5 10.8 11.8
Regular partner 19.7 22.6 32.5 37.2
Casual partner 5.6 23.8 31.8 33.3
Has never had sex 12.8 11.1 na na
Recall of Maximum ad message
Recalls message 8.2 14.9 40.0 34.5
Has seen/heard other advertising 78.0 76.4 15.7 22.5
Has not seen/heard any advertising 13.8 8.7 12.5 16.1
Travel time to source of condoms
¾10 minutes walk 46.9 65.9 17.2 28.3
>10 minutes walk 53.1 34.1 16.9 15.9
Notes: The percentages and the numbers of cases presented in this and subsequent tables are weighted. na=not applicable.

CURRENT PARTNERSHIP STATUS

Current partnership status has important implications for sexual behavior that places individuals at risk of contracting sexually transmitted diseases. Consistent with women tending to marry at earlier ages than men, there were significant differences in the current partnership status of younger men and younger women (Table 2): Four times as many women as men aged 15-24 were married (40% vs. 9%). There was no significant difference in the current partnership status of men and women aged 25-49: Seventy-six percent of each were currently married.

Table 2. Percentage distribution of respondents, by partnership status, according to age and sex
Partnership status Total

(N=806)
15-24 25-49
Women
(N=206)
Men
(N=188)
Women
(N=185)
Men
(N=227)
Unmarried, sexually active
No regular partner 16.4 8.7 33.5 12.4 11.9
Regular partner 20.5 26.7 33.0 11.4 11.9
Married 51.3 40.3 9.3 76.2 75.9
Total 100.0 100.0 100.0 100.0 100.0

LAST INTERCOURSE WITH A CASUAL PARTNER

All respondents were asked to describe their relationship with their last sexual partner. Their responses were coded by interviewers as last sex with a marital, a regular, a casual or a commercial sex partner. Five male respondents reported having a commercial sex partner (not shown). For this analysis, these cases were recoded as last intercourse with a casual partner.

Overall, men were almost four times as likely as women to report having had last sex with a casual partner (27% vs. 7%). Men were approximately 5-6 times as likely as women to have last had sex with a casual partner, whether they were married (26% vs. 4%) or in a nonmarried regular partnership (5% vs. 1%). In contrast, unmarried men who did not have a regular partner were less than twice as likely as comparable women to report having had sex with a casual partner at last intercourse (73% vs. 41%).

CHANGES BETWEEN 1990 AND 1996

Number of partners. We compared our findings with the 1990 WHO data on the number of sexual partners that respondents reported having had outside of marital or regular partnerships in the last 12 months. In 1990, 33% of men and 9% of women reported having had at least one sexual partner outside of marriage or a regular partnership in the last 12 months. In 1996, these percentages were almost identical (32% and 9%, respectively). Thus, over a six-year period, among both men and women who were married or had a regular partner, there appears to have been virtually no change in sexual partnerships outside these marital or regular relationships.

Access to condoms. A comparison of the 1996 data with the 1990 WHO data shows that the percentage of respondents who knew of a condom source increased from 54% in 199018 to 79% in 1996. Moreover, among those who knew of a condom source, the percentage of respondents reporting less than 15 minutes traveling time to a source of condoms more than doubled, from 32% to 72%.

CONDOM USE AT LAST INTERCOURSE

Bivariate analyses. Condom use was significantly higher among women aged 15-24 than among those aged 25 and older (Table 1). Among men, condom use was significantly higher for men 15-29 than for those 30 and older. There were no significant differences in condom use between men and women aged 15-19 and those older than 30. Condom use at ages 20-24 and 25-29, however, was significantly higher for men than for women.

Condom use varied significantly by relationship with last partner. For both women and men, condom use was almost three times as high for sex with a regular or a casual partner as for sex with a marital partner. (Of the five men in the survey who reported having had sex with a commercial sex worker, three used condoms.)

There was no significant association between schooling and condom use at last intercourse for women or men, although for men the *2 test was marginally significant at p<.10. The absence of a significant association with schooling is surprising, and suggests that factors other than education may have a more important role in determining condom use in Lusaka. Nevertheless, it is possible that some effect of schooling on condom use is captured by the respondent's age or his or her relationship with the last partner.

Recall of the Maximum brand advertising message showed a significant association with condom use at last intercourse for both men and women. Forty percent of women and 35% of men who had heard the advertising message reported having used a condom at last sex. This finding suggests that information about condoms may increase use.

The relationship between the accessibility of condoms and condom use was significant only for men, who were more likely to report having used a condom at last intercourse if they also reported being within a 10-minute walk of a source of a condoms (28%) than if they lived further away (16%).

Multivariate analyses. We conducted multivariate logistic regression analyses to assess the net effects of age, relationship with last partner, schooling, recall of the advertising message and the accessibility of condoms on men's and women's reports of condom use at last intercourse. Before adding any other variable, however, we found that age was a significant predictor of women's report of condom use at last sex (Table 3): Women aged 15-19 were more likely to report use of a condom than were women older than 20. When we added women's relationship with their last partner, this variable became significant, while age ceased to be. This is because younger women are more likely to be unmarried than older women and younger women are more likely than older women to engage in sex with a casual or regular partner.

Table 3. Odds ratios from logistic regression analysis showing the likelihood that a woman's partner used a condom at last intercourse, by set of variables included in analysis, according to characteristic
Characteristic Age only Plus relationship Plus schooling Plus recall Plus distance
Age
15-19 2.82*** 1.56 2.02 2.12 2.11
20-24 2.05* 1.73 1.90 1.93 1.93
25-29 1.12 1.09 1.20 1.21 1.20
30-49 1.00 1.00 1.00 1.00 1.00
Relationship with last partner
Marital na 1.00 1.00 1.00 1.00
Regular na 3.49*** 2.98*** 2.81*** 2.83***
Casual na 3.58** 3.29** 3.07** 3.06**
Schooling
None/any primary na na 1.00 1.00 1.00
Any secondary na na 1.05 1.00 0.99
>secondary na na 2.08* 1.77 1.76
Recall of Maximum ad message
Recalls message na na na 2.76** 2.78**
Has seen/heard other advertising na na na 0.76 0.75
Has not seen/heard any advertising na na na 1.00 1.00
Distance to source
¾10 minutes na na na na 0.91
>10 minutes/no source known na na na na 1.00
Model *2 9.04 14.87 3.49 5.00 0.09
df 3 2 2 2 1
*p<.10. **p<.05. ***p<.01. Note: na=not applicable.

Women's relationship with their last partner remained significant when schooling was added, as well as when advertising recall was included. Upon controlling for all of these factors, we found that women who had last had sex with a casual or a regular partner were about three times as likely as women whose last intercourse had been with their marital partner to report condom use. In addition, women who recalled the advertising message were almost three times as likely as other women to report condom use at last intercourse. Finally, when access to condoms was included, there was no change in the effect of women's recall of the advertising message on condom use.

For men, age showed a powerful effect on condom use at last intercourse (Table 4). Even after their relationship with their last partner, schooling and recall of advertising were controlled for, age remained significant: Men aged 15-29 were more than three times as likely as men older than 30 to report having used a condom at last sex.

Table 4. Odds ratios from logistic regression analysis showing the likelihood that a man used a condom at last intercourse, by set of variables included in analysis, according to characteristic
Characteristic Age only Plus relationship Plus schooling Plus recall Plus distance
Age
15-19 5.28*** 2.80** 3.55*** 3.54*** 3.49***
20-24 5.53*** 3.22*** 3.46*** 3.38*** 3.31***
25-29 5.20*** 3.70*** 3.88*** 3.93*** 3.81***
30-49 1.00 1.00 1.00 1.00 1.00
Relationship with last partner
Marital na 1.00 1.00 1.00 1.00
Regular na 2.50** 2.19** 2.14* 2.06*
Casual na 2.13** 1.99* 1.96* 1.89
Schooling
None/any primary na na 1.00 1.00 1.00
Any secondary na na 1.41 1.39 1.42
>secondary na na 2.21** 2.06** 2.16**
Recall of Maximum ad message
Recalls message na na na 1.31 1.42
Has seen/heard other advertising na na na 0.85 0.99
Has not seen/heard any advertising na na na 1.00 1.00
Distance to source
¾10 minutes na na na na 1.93**
>10 minutes/no source known na na na na 1.00
Model *2 35.88 6.13 5.37 0.67 4.95
df 3 2 2 2 1
*p<.10. **p<.05. ***p<.01. Note: na=not applicable.

Men's relationship with the last sexual partner showed at least marginal significance (p<.10) through all stages of the multivariate analysis. We also conducted tests for both men and women to measure differences in condom use by last sex with regular or casual partners, using sex with regular partner as the reference category (not shown). We found no difference in the likelihood of condom use between casual and regular partners for men or women. Men's schooling, however, showed a significant association with men's reported condom use at last intercourse in all analyses: Men with schooling beyond the secondary level were twice as likely as men with primary or no schooling to report using a condom at last intercourse.

Being within 10 minutes of a condom source significantly increased a man's likelihood of condom use at last intercourse: A man with a nearby source of condoms was about twice as likely as other men to have used a condom at last sex, even after the effects of age, relationship, schooling and advertising recall were taken into account.

DISCUSSION

The prevalence of casual sex was substantial among men in Lusaka, with 27% of all sexually experienced men reporting last sex with a casual partner. Casual sex was highest among unmarried men who did not have a regular partner. Among women, casual sex was also high among those who were unmarried and did not have a regular partner, although the proportion was still lower than that for comparable men. Overall, men were more likely than women to report having had last intercourse with a nonmarital partner. In part, this is a consequence of the younger ages at marriage for women. For both men and women, marriage substantially lowered the likelihood of casual sex.

Some of the differential between men and women in casual sex may also be attributed to the role of gender ideology and the sexual double standard that not only permits men to have multiple partnerships,19 but treats multiple sexual partnerships as a reflection of a healthy lifestyle for men,20 even though Zambian husbands take strong actions against their wives for similar behavior. (Such actions may include taking their wives to court for committing adultery.21)

Women in regular partnerships may view their relationship as a precursor to marriage, and may be hesitant to jeopardize their marital prospects by engaging in casual sex. Because of the sexual double standard and because sex outside of marriage places them in a vulnerable position, women may also be less likely to report having sex with someone other than their marital or regular partner.

The lower differential in casual sex between men and women who are not in a regular or marital partnership may also be explained by social status of unmarried women in Zambia. In a society where marriage and childbearing is the norm for women22 and where access to societal resources is through men,23 women may be extremely vulnerable if they are not in a relationship with a man. The social vulnerability of unattached women is underscored by the fact that married Zambian women look upon unattached Zambian women with distrust, as potential competitors.24 Women who are neither married nor in a regular relationship with a man may have casual relationships in order to form emotional ties leading to marital partnership and greater financial and emotional security.

We found no evidence that men and women decreased the number of sexual partners they had outside of marriage or regular partnership between 1990 and 1996. Given that patterns of sexual behavior reflect fundamental aspects of social organization, and that these patterns are influenced by gender ideologies that prescribe roles for men and women, the lack of change in sexual partnerships outside of regular or marital relationships is not surprising. Our findings are in keeping with results of surveys conducted in the Copperbelt and Northern regions of Zambia, which also showed no change in the percentage of men and women reporting nonregular sexual partnerships.25 However, that study did show an increase in condom use during this period. Because information on the level of condom use at last sex was not available in the 1990 survey, we were unable to make a comparison of changes in condom use in Lusaka over time. We did examine levels and correlates of condom use at last sex in 1996, however.

Our findings are consistent with the hypothesis that Zambian women, given their low socioeconomic status, are limited in their ability to negotiate condom use in sexual relationships. For women, relational factors were very important in determining whether their partners used condoms; women's individual characteristics, such as age or education, were not associated with condom use in the full multivariate analysis. Nevertheless, the strong association between women's recall of the brand advertising message and condom use suggests that informed women may be able to encourage their partners to use condoms by relaying information about condoms to their partners. Research has shown that women are often the source of information on family planning methods for their partners.

For men, age and education were significant predictors of condom use. Younger and more educated men were more likely to use a condom at last intercourse. Younger men may have greater motivation for condom use because of having become sexually active during the height of the AIDS epidemic. Men aged 15-29 in 1996 were in their adolescence or early adulthood when awareness about the AIDS epidemic was becoming widespread in Lusaka, and were more than three times as likely as men older than 30 to have used a condom in last sex. These findings suggest that younger cohorts of men in Lusaka have adopted condom use as a preventive strategy in response to the AIDS epidemic. More educated men may be similarly motivated to use condoms because of their greater awareness of AIDS.

Condom use among men aged 20-24 and 25-29 was significantly higher than among women of similar age. Part of this difference may be explained by the fact that men have sex with younger women. Some of this difference may also be explained by higher levels of commercial sexual partnerships among men aged 20-29.

Access to condoms was a significant predictor of condom use at last intercourse for men, but not for women. This is not surprising, given that 79% of female respondents reported that their partner obtained a condom the last time they had intercourse. A woman's character may be thought questionable if she obtains condoms. Because of cultural acceptance of multiple sexual partnerships for men, it may not be considered unusual for a man to obtain condoms: Eighty-six percent of male respondents reported that they obtained condoms themselves the last time they had sex.

Overall, the motivation to use condoms appears to be a more powerful determinant of condom use for men than relational factors. These findings are consistent with the interpretation that given their higher status relative to women and their greater power within relationships, men are able to enforce condom use, and suggest that campaigns that encourage men to use condoms may be more successful in promoting condom use than campaigns aimed at women.

Condom use was significantly associated with nonmarital sexual activity. About one-third of men and women reported condom use at last intercourse with a casual or a regular partner. Women's reports of higher condom use in casual or regular sex than in marital sex is in keeping with the finding that men are more likely to use condoms outside of marriage than within marriage. It may also be easier for women to negotiate condom use outside marriage than within marriage. The fact that condom use with a casual or regular partner is at similar levels for both men and women suggests that risk assessment criteria for sex outside of marriage differ from those based on epidemiolgic notions of risk.

Men and women do not evaluate their risk of contracting HIV simply on the basis of whether the relationship is a casual or regular partnership. Both men and women likely realize that regular nonmarital partnership does not necessarily offer them protection against sexually transmitted diseases. Men are probably aware that since the level of financial support that they are able to offer is very limited, their regular nonmarital partners may have other sexual relationships,26 and women know that men feel free to have multiple sexual partnerships.27

Other factors, such as whether a person perceives a partner to be at risk of HIV on the basis of their appearance and the comfort or emotional closeness that a person feels with their partner, may be more important to individuals in determining condom use.28 Women sometimes believe that certain sexual positions do not place them at risk of HIV infection. Many people also believe that one-night stands, rapid intercourse or intercourse followed by a shower are risk-free.29

The lower level of condom use in marriage is not surprising, given the issues of trust raised by the use of condoms in marital sex. Married Zambian women feel helpless in raising the subject of condom use with their husbands and are afraid of denying men sex because they fear reprisal.30 In addition, married men may not want to use condoms with their wives, since condom use may be construed as an admission of sexual relationships outside marriage: While extramarital sex may be culturally accepted31 for men, it can still be a cause of dispute within marriage. Men may also avoid using condoms with their marital partners if they feel it decreases sexual pleasure. In addition, marital partners may be using more effective contraceptive methods to prevent pregnancy or they may want more children. Thus, the relevance of condom use within marriage may be limited.32

CONCLUSION

Our findings suggest that AIDS prevention approaches that target men directly should receive the greatest attention in Zambia. Programs that focus on teaching women how to negotiate safer sex may be less effective, as they are likely to work only when women have control over
decision-making.33

There is also a clear need for a better understanding of male sexuality. Men's fears about compromising their masculinity, for example, may be very important in determining whether they take precautions such as condom use.

Programs to relay gender-specific information to men in a manner acceptable to them are urgently needed. Changes within Zambian society to increase the status of women and foster the development of more egalitarian gender ideologies are no less important.

REFERENCES

1. Fylkesnes K, Brunborg H and Msiska R, Zambia: The Current HIV/AIDS Situation and Future Demographic Impact, Background Paper No. 1, Lusaka, Zambia: Ministry of Health, 1994; and van den Borne F, Tweedie IA and Morgan WB, Family Planning and Reproductive Health in Zambia Today, IEC Field Report Number 2, Baltimore, MD, USA: Johns Hopkins Center for Communication Programs, 1996.

2. van den Borne F, Tweedie IA and Morgan WB, 1996, op. cit. (see reference 1).

3. Yoder PS, Hornik R and Chirwa BC, Evaluating the program effects of a radio drama about AIDS in Zambia, Studies in Family Planning, 1996, 27(4):188-203.

4. Ingham R, AIDS: knowledge, awareness and attitudes, in: Cleland J and Ferry B, eds., Sexual Behavior and AIDS in the Developing World, London: Taylor and Francis, 1995, pp. 43-74; Gaisie K, Cross AR and Nsemukila G, Zambia Demographic and Health Survey 1992, Columbia, MD, USA: Macro International, 1993; and Kapilikisha M, AIDS: Zambians seem to be getting the message, New African, 1990, No. 268, p. 39.

5. Mouli V, Bridges crossed yesterday, peaks to be conquered tomorrow: AIDS and the condom, African Health, 1992, 14(5):12-14; and van den Borne F, Tweedie IA and Morgan WB, 1996, op. cit. (see reference 1).

6. Kapilikisha M, 1990, op. cit. (see reference 4).

7. Sweat MD and Denison JA, Reducing HIV incidence in developing countries with structural and environmental interventions, AIDS, 1995, 9(Supplement A): S251-S257.

8. Dixon-Mueller R, The sexuality connection in reproductive health, Studies in Family Planning, 1993, 24(5):269-282.

9. Worth D, Sexual decision-making and AIDS: why condom promotion among vulnerable women is likely to fail, Studies in Family Planning, 1989, 20(6):297-307; and Ulin PR, African women and AIDS: negotiating behavioral change, Social Science in Medicine, 1992, 34(1):63-73.

10. Soskolne V et al., Condom use with regular and casual partners among women attending family planning clinics, Family Planning Perspectives, 1991, 23(5):222-225; Mwale G and Burnard P, Women and AIDS in Rural Africa, Brookfield, VT, USA: Ashgate Publishing Company, 1992; and Le Franc E et al., Working women's sexual risk taking in Jamaica, Social Science in Medicine, 1996, 42(10):1411-1417.

11. Potter LB and Anderson JE, Patterns of condom use and sexual behavior among never-married women, Sexually Transmitted Diseases, 1993, 20(4):201-208; and Soskolne V et al., 1991, op. cit. (see reference 10).

12. Hansen KT, Negotiating sex and gender in urban Zambia, Journal of Southern African Studies, 1984, 10(2):219-238; and Hansen KT, Keeping House in Lusaka, New York: Columbia University Press, 1997.

13. Hansen KT, 1997, op. cit. (see reference 12).

14. Campbell T and Kelly M, Women and AIDS in Zambia: a review of the psychosocial factors implicated in the transmission of HIV, AIDS Care, 1995, 7(3):365-373.

15. Ibid.

16. Caraël M et al., Sexual behavior in developing countries: implications for HIV control, AIDS, 1995, 9(10):1171-1175.

17. Armitage P and Berry G, Statistical Methods in Medical Research, London: Blackwell Scientific Publications, 1987; and Norusis MJ, SPSS Base System User's Guide, Chicago: SPSS Inc., 1990.

18. Mehryar A, Condoms: awareness, attitudes and use, in: Cleland J and Ferry B, 1995, op. cit. (see reference 4), pp. 43-74.

19. Caraël M, Sexual behavior, in: Cleland J and Ferry B, 1995, op. cit. (see reference 4), pp. 75-123.

20. Hawkins K, Male Participation in Family Planning, A Review of Programme Approaches in Africa, London: International Planned Parenthood Federation, 1992.

21. Hansen KT, 1997, op. cit. (see reference 12).

22. Campbell T and Kelly M, 1995, op. cit. (see reference 14).

23. Hansen KT, 1997, op. cit. (see reference 12).

24. Mwale G and Burnard P, 1992, op. cit. (see reference 10).

25. Yoder PS, Hornik R and Chirwa BC, Impact of a radio drama about AIDS in Zambia: a program called Nshilakamona, final evaluation report, Working Paper, Philadelphia, PA, USA: Annenberg School for Communication, 1993, No. 1013.

26. Schoepf BG, Women and AIDS and economic crisis in Central Africa, Canadian Journal of African Studies, 1988, 22(3):625-644.

27. Mwale G and Burnard P, 1992, op. cit. (see reference 10).

28. Sobo EJ, Finance, romance, social support, and condom use among impoverished inner-city women, Human Organization, 1995, 54(2):115-128; and Lear D, Sexual communication in the age of AIDS: the construction of risk and trust among young adults, Social Science in Medicine, 1995, 41(9):1311-1323.

29. Nzioka C, Lay perceptions of risk of HIV infection and the social construction of safer sex: some experiences from Kenya, AIDS Care, 1996, 8(5):565-579; and Schoepf BG, 1988, op. cit. (see reference 26).

30. Mwale G and Burnard P, 1992, op. cit. (see reference 10).

31. Ibid.

32. De Zoysa I, Sweat MD and Denison JA, Faithful but fearful: reducing HIV transmission in stable relationships, AIDS, 1996, 10(Supplement A):S197-S203.

33. Campbell CA, Male gender roles and sexuality: implications for women's AIDS risk and prevention, Social Science in Medicine, 1995, 41(2):197-210.

Sohail Agha is research officer at Population Services International, Washington, DC. The research on which this article is based was made possible through financial support from the U.S. Agency for International Development. This paper has benefited from comments and suggestions made by Dominique Meekers and Guy Stallworthy. Support of the research project by Brad Lucas, Sanjay Chaganti and Chris Mukkuli is greatly appreciated by the author.