Secondary Logo

Journal Logo


A Study of Sexual Behavior Change in Rural Senegal

Lagarde, Emmanuel; Pison, Gilles; Enel, Catherine

Author Information
Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology: March 1, 1996 - Volume 11 - Issue 3 - p 282-287
  • Free


WHO estimates indicate that globally there were 16 million persons with the HIV virus in mid-1994. Of these, 10 million live in sub-Saharan Africa, where sexual contact among heterosexuals is the major mode of transmission. In the absence of efficacious treatments and vaccines, the only weapon to prevent infection is convincing people to use condoms and to modify their behavior. Hence, studies of sexual behavior aimed at identifying risky behaviors with a view to formulating effective prevention programs are urgently needed. Several studies have investigated the distribution and determinants of sexual behavior in various African cities where the prevalence of HIV is high (1-5). But there is little information about sexual behavior in rural areas, where the vast majority of the African population lives and where recent reports indicate that HIV prevalence is beginning to rise (6-12).

A pilot sociodemographic study was undertaken in 1984 in a rural zone of Senegal. The population has been followed-up to describe the demographic features and their evolution over a 10-year period. The study provided a propitious background to undertake socioepidemiological investigations regarding AIDS in this area. A study of sexual practices that has been conducted from 1989 to 1994 has had two purposes: first, to describe sexual activities and, second, to shed light on determinants of risk behavior with respect to HIV transmission. We present here results showing large changes in marriage and sexual debut and in multipartnership's determinants in this area over 30 years.



The study area is located in the region of Ziguinchor, in the extreme southwest corner of Senegal. It is located 50 km west of the city of Ziguinchor and 25 km north of the border with Guinea-Bissau. This community has been under observation by the authors for the past 10 years. In 1984, the authors conducted a complete enumeration of the study zone and have returned every year for the past 10 years to record all new vital events. In 1993, there were 7,125 individuals living in the study zone.


The population belongs mainly to the Diola ethnic group and is of animist or Catholic religion. Polygamy is rare (2% of the married women are in polygamous marriages). The principal means of subsistence in the region is rice farming. Most of the adult population spend half of the year or more away from the area as seasonal migrant workers (13). Both men and women migrate before they are married, but, once married, women no longer migrate. Men continue their seasonal migrations up to the age of ≈60 even after they are married. Eighty percent of women aged 15-24 leave to work as maids in the major cities of Senegal and the Gambia. Eighty-two percent of men aged 20-39 migrate to other regions of Senegal during the dry season, mostly to harvest palm wine. There is also systematic, but sometimes temporary, emigration to urban areas for divorced women.

In 1990, a serological survey of the population was made of individuals aged ≥20 identifying 27 individuals out of a possible 3,230 as HIV positive, a prevalence of 0.8% (14). Prevalence was very low between the age of 20 and 24. After age 25, it was rather constant with age. A case-control study showed that infection was associated with several factors among which temporary emigration played a crucial role (12).

Marriage Survey

Couples in the community who marry in the same year pass through some of the marriage steps together. This creates ties between couples who naturally identify themselves as members of a cohort. We took advantage of the existence of these marriage cohorts to determine retrospectively the date of all the marriages performed over the past 50 years. During the 1984 census, each once-married adult was systematically asked to list all the people who married for the first time in the same year as him/her. Once the marriage cohorts were constructed, the year of marriage of each cohort was determined by using a combination of direct and indirect methods. For the younger cohorts, marriage certificates were often available from some of the couples in the cohort. For the older cohorts, for which no official documentation was available, we matched the date of the marriage ceremony with the calendar of local historical events. In other cases, we used the dates sometimes carved in walls above front doors, as one step in the marriage process is for a newly married man to build a new house. Consequently, we derived the date of first marriage for every once-married person in the population, which we subsequently combined with information on their present age to derive ages at first marriage.

Survey of Sexual Practices

Our original goal was to interview a random sample of the adult population of the area on sexual practices. But this proved unrealistic: as explained earlier, during the dry season, a large part of the population is engaged in temporary migration and is scattered in places located between one and hundreds of kilometers from the study area. Consequently, they cannot be reached easily. On the other hand, during the rainy season, most people are present, but they are farming the entire day and therefore cannot be easily interviewed. Furthermore, the time spent in the village is only a part of their life, while time is also spent outside the village. This is true for most of the population. It would have been misleading to restrain the survey only to the periods spent in the villages. Therefore, instead of selecting one sample of the whole population, we chose to select several samples: one for every subgroup of the population. Hence, between 1989 and 1994, several surveys were carried out among those subpopulations: 79 currently married sedentary men and 78 currently married sedentary women, 51 male seasonal migrants, 48 young never-married female seasonal migrants, and a sample of 74 men and women interviewed in the villages at the end of the rainy season. A sample was randomly developed for every subpopulation. Respondents were interviewed partly in the place of seasonal migration and partly at home in order to take into account both life in the village and life away from home. Our sample is therefore the concatenation of several surveys conducted of the same population among several homogeneous subpopulations.

Initially, 332 persons were selected. Two men refused to participate and two other men refused to reply to some of our questions. Finally, 330 persons were included in the analysis: 165 men and 165 women. This represents 9% of the adult population. A comparison of the 330 subjects in the subsample with the general population showed significant differences in age, marital status, and migration status (Table 1). In particular, never-married people, migrants, and the aged were underrepresented because of the sampling process.

Following informed consent, information on matrimonial history, sexual practices (age at first intercourse; premarital, and extramarital sexual activity; and frequency of sexual intercourse), and knowledge and attitude toward contraception and AIDS, was obtained from each participant during a personal interview with a modified version of the WHO Partner Relation Survey questionnaire (14). Local interviewers chosen among those who were accustomed to the ongoing demographic survey underwent a 3-day training program designed to standardize the use of the questionnaire.


A χ2 test was used to compare proportions and a t test to compare means. Stepwise multiple logistic regression was performed to evaluate the association between predictor variables and declared risk behavior. Knowledge of AIDS was estimated by using a summary score developed from seven questions about routes of AIDS transmission. One point was assigned for each correct response and 0 for each incorrect or unsure response. Persons who had not heard of AIDS were given the score of 0. Thus, the score ranged from 0 (no knowledge of AIDS) to 7 (very knowledgeable).


Age at First Marriage and Age at First Sexual Intercourse

The evolution of age at first marriage is shown in Fig. 1. From 1955-64 to 1985-92, age at first marriage increased from 25 to 32 years for men and from 21 to 24 years for women. The course of age at first sexual intercourse is shown in Fig. 2. Women declared the same age on average—19 years at first intercourse—whatever their date of birth, yet age at first intercourse declared by men markedly decreased from ≈27 years for those born before 1940 to 18 years for those born after 1957.

Table 2 shows the proportion of married persons who declared having had premarital sex and the mean number of premarital partners. Note that we have not counted sexual relations with the spouse before marriage as premarital sex. Young men declared premarital sex more often than did older men. Women older than 50 years rarely declared premarital sex.


Among sexually active individuals, 28.4% of men (46 of 162) and 5.2% of women (7 of 134) declared more than one sexual partner in the previous 12 months. They were 28.8% (19 of 66) among never married and 13.8% (31 of 224) among currently married.

Variables found to be associated with multipartnership are listed in Table 3. Having more than one partner in the previous 12 months was associated with most of the variables examined, yet most of them are intercorrelated, so we then performed a multivariate analysis (Table 4). Men were more likely to engage in multipartnership than women [odds rate (OR) = 4.4], as were persons who had their first intercourse earlier than those who had it late (OR = 3.3), rural migrants rather than sedentary people and urban migrants (OR = 2.7), and those with a greater knowledge of AIDS. This means that individuals with more risk of infection are those more informed about it. Age at the time of interview has been included into the logistic model to exclude its possible confounding effect on the correlation between age at first intercourse and multipartnership.

Frequency of Sexual Acts

Frequency of sexual acts was 2.3 acts of sexual intercourse in the previous 4 weeks on average for the whole sample. It varied heavily according to (a) the marital status, ranging from 0.1 act for never-married people without a regular partner to 3.2 for currently married people with a spouse present at the time of the survey; (b) sex—among sexually actives, women tend to declare fewer acts than men (2.7 vs 2.4), although the difference is not significant; and (c) age, ranging from <2 acts in the previous 4 weeks for those aged <30 or >60 to >3 acts for those aged 30-60. In addition, 51% of the whole sample did not declare any acts in the previous 4 weeks. Frequencies of those having declared 1, 2, 3, 4, 5, 6, 7, and ≥8 acts in the previous 4 weeks were 4%, 8%, 10%, 9%, 7%, 4%, 2%, and 5% respectively.


The study shows a dramatic change in premarital sexual habits among the younger generation of the rural community studied when compared with their parent's generation: notably, earlier sexual debut and an expansion of sexual premarital relations. In addition, multipartnership was found to be associated with earlier sexual debut and temporary migration. Since migrations are widespread, they may be a serious threat to disease spread.

All questions were retrospective covering from 4 weeks to the entire lifetime of those interviewed. It is well known that answers become all the more inaccurate depending upon the time that has elapsed. In addition, the reliability of investigations of sexual behavior is questionable. A validity survey conducted among married couples of the same population (15) has revealed a fairly high reliability of answers regarding recent marital sexual acts. It appears that distortions stemmed mainly from memory bias and not from deliberate refusal to tell the truth. Another problem could be the possibly standard answers of older respondents. They may have consciously or unconsciously omitted remote sexual relationships that did not correspond with social standards. This could be the case for both premarital and extramarital sexual relations. For instance, the authors' qualitative observations suggest that women underreport extramarital sexual activities.

The marriage survey has shown that the age at first marriage has increased by 7 years for men and 3 years for women over the last 30 years. In Senegal as a whole, an increase in age at marriage has been reported (16), but it is much more marked for women than for men. This trend is associated with education and therefore the increase is greater in urban areas. The conditions are clearly different in the study area, and the older age at first marriage for men is due partly to economic considerations. Indeed, the future husband must pay a marriage compensation or bufos in the form of palm wine or more often in the form of cash. The increase in its cost and the increasing importance of the monetary system explain the growing difficulty to achieve marriage.

Though the sampling framework did not produce a sample totally representative of the whole population, apart from the divorced and widowed people who were each underrepresented, all categories of the population regarding age, sex, marital, and migration status were represented.

On the one hand, data on ages at first marriage were rather precise and unbiased, thanks to the method of estimation used. On the other, declarations of age at first intercourse were probably imprecise. Nevertheless, the pattern of change in age at first intercourse is remarkable: it has decreased a great deal for men but not at all for women. The decrease for men is exaggerated: there is an inconsistency in that men born in the 1925-40 cohort, who would for the most part marry in the period 1945-64, report first sexual contact some years after the mean age of marriage. However, the trend is probably true. There is no change for women, so a systematic bias for men is unlikely.

The opposite trends in age at marriage and in age at first intercourse are consistent with the observed increase in premarital sex. Multipartnership is more frequent among younger people. This result, which has been found by other studies in rural Africa (8,17), could stem from either age or cohort effect. However, the finding that multipartnership was related to the age at first sexual intercourse supports a cohort-effect contribution. In addition, a cross-national study (18) has shown a similar relation, at the societal level, between the frequency of premarital sex and the frequency of extramarital sex.

The correlation between at-risk behavior and temporary migration is consistent with results of the serological survey conducted in 1990 (12) showing that migration was one of the main factors associated with the spread of HIV infection. Hence, multipartnership, migration, and HIV infection are linked and constitute a possibility for HIV entrance into the community.

Finally, we noted the low coital frequency in the sample. This is partly related to the high proportion of people who declared no sexual acts during the recent period (the last 4 weeks). Those in this situation who were married were often temporarily separated from their spouses because of seasonal migration.

This study has pointed out a dramatic evolution in sexual habits. Such a change is associated with an enhanced sexual network, thereby opening a path for the dissemination of HIV. This situation has worsened because, as described elsewhere (19), condom use is still infrequent in the area (only 9% of sexually actives in 1989 declared having ever used a condom). Subsequent research is consequently in progress to identify the best way to induce behavioral change and to help design effective HIV prevention programs.

Acknowledgment: The authors are indebted to B. Cohen and M. Caräel for critical review of the manuscript and to M. Lefebvre, sister Jeanne-Marie, and sister Marie-Joëlle for their assistance. Thanks are due to the Ministère de la Santé, and Comité National Pluridisciplinaire de prévention du Sida, Sénégal, in particular I. Ndoye, for their interest and help. This research was supported by the Muséum National d'Histoire Naturelle, the Centre National de la Recherche Scientifique (URA49), the Institut National d'Etudes Démographiques, and the Agence Nationale de Recherches sur le Sida, Paris, France.

FIG. 1
FIG. 1:
. Mean age at first marriage by period of marriage. Numbers in each marriage cohort (45-54, 55-64, 65-74, 75-84, and 85-92) are 198, 263, 154, 142, and 133, respectively, for men, and 170, 232, 145, 159, and 215, respectively, for women. Black squares, men; white squares, women. Source: marriage survey.
FIG. 2
FIG. 2:
. Mean age at first sexual intercourse related to period of birth: confidence interval (95%) error bars are shown. Source: survey on sexual practices. Black squares, men; white squares, women.


1. Adamchak DJ, Mbizvo MT, Tawanda M. Male knowledge of and attitudes and practices towards AIDS in Zimbabwe. AIDS 1990;4:245-50.
2. Hoegsborg M, Aaby P. Sexual behaviour, use of condoms and perceptions of AIDS in an urban area of Guinea-Bissau with a high prevalence of HIV-2. In: Dyson T, ed. Sexual behaviour and networking: anthropological and sociocultural studies on the transmission of HIV. Liège, Belgium: Derouaux Ordina, [International Union for the Scientific Study of Population], 1992:203-31.
3. Govender V, Bhana R, Pillay A, Panchia R, Padayachee GN, De Beer M. Perception and knowledge about AIDS among family planning clinic attenders in Johannesburg. S Afr Med J 1992;81:71-4.
4. Müller O, Barugahare L, Schwartländer B, et al. HIV prevalence, attitudes and behaviour in clients of a confidential HIV testing and counselling center in Uganda. AIDS 1992;6:869-74.
5. Lindan C, Allen S, Caraël M, et al. Knowledge, attitudes, and perceived risk of AIDS among urban Rwandan women: relationship to HIV infection and behavior change. AIDS 1991;5:993-1002.
6. Wilkins A, Ricard D, Todd J, et al. The epidemiology of HIV infection in a rural area of Guinea-Bissau. AIDS 1993;7:1119-22.
7. Barongo LR, Borgdorff MW, Mosha FF, et al. The epidemiology of HIV-1 infection in urban areas, roadside settlements and rural villages in Mwanza Region, Tanzania. AIDS 1992;6:1521-8.
8. Schopper D, Doussantousse S, Orav J. Sexual behaviours relevant to HIV transmission in a rural African population: how much can a KAP survey tell us? Soc Sci Med 1993;3:401-12.
9. Kane F, Alary M, Ndoye I, et al. Temporary expatriation is related to HIV-1 infection in rural Senegal. AIDS 1993;7:1261-5.
10. Serwadda D, Wawer MJ, Musgrave SD, Sewankambo NK, Kaplan JE, Gray RH. HIV risk factors in three geographic strata of rural Rakai District, Uganda. AIDS 1992;6:983-9.
11. Mulder DW, Nunn AJ, Kamali A, Nakiyingi J, Wagner HU, Kengeya-Kayondo JF. Two-year HIV-1 associated mortality in a Ugandan rural population. Lancet 1994;343:1021-3.
12. Pison G, Le Guenno B, Lagarde E, Enel C, Seck C. Seasonal migration: a risk factor for HIV infection in rural Senegal. J Acquir Immun Defic Syndr 1993;6:196-200.
13. Enel C, Pison G, Lefebvre M. Migrations and nuptiality changes: a case study in rural Senegal. In: Bledsoe C, Pison G, eds. Nuptiality in Sub-Saharan Africa: contemporary anthropological and demographic perspectives. Oxford: Clarendon, 1994:92-113.
14. Le Guenno B, Pison G, Enel C, Lagarde E, Seck C. HIV-2 prevalence in three rural regions of Senegal: low levels and heterogeneous distribution. Trans R Soc Trop Med Hyg 1992;86:301-2.
15. Lagarde E, Pison G, Enel C. Reliability of reports of sexual behavior: a study of married couples in rural West Africa. Am J Epidemiol 1995;141:1194-200.
16. Pison G, Hill K, Cohen B, Foote K, eds. Population dynamics of Sub-Saharan Africa: the case of Senegal. Washington, DC: National Academy Press, 1995.
17. Konings E, Blattner WA, Levin A, et al. Sexual behavior survey in a rural area of northwest Tanzania. AIDS 1994;8:987-93.
18. Caraël M. Sexual behavior. In: Cleland F, Ferry B, eds. AIDS in the developing world. London: Taylor and Francis, 1995:75-123.
19. Enel C, Pison G. Sexual relation in the rural area of Mlomp (Casamance, Senegal). In: Dyson T, ed. Sexual behaviour and networking: anthropological and socio-cultural studies on the transmission of HIV. Liège, Belgium: Derouaux Ordina, [International Union for the Scientific Study of Population]. 1992:249-67.

Africa; HIV; Migration; Senegal; Sexual behavior; STD

© Lippincott-Raven Publishers.