The initiation of sexual intercourse is a milestone in the physical and psychological development of men and women in all societies, and both the timing of this event and the context within which it occurs can have immediate and longer term consequences for the individual. There are possibly serious health and social consequences for women who begin to have intercourse while very young or not yet married, especially if they become pregnant and have either an unplanned birth or, in some settings, an unsafe abortion. Some instances of very early sexual intercourse are involuntary--for example, when a young person is raped, is the victim of incest or turns to prostitution because of financial need. Moreover, first intercourse marks the beginning of young people's possible exposure to the risk of contracting sexually transmitted diseases (STDs).

Analysts have commented on differences in expectations and values regarding male and female sexual behavior. For example, in many societies, it is excusable for men to become sexually active at a significantly earlier age than women, and even expected that they will do so.1 Similarly, women but not men are commonly expected to be sexually inexperienced when they marry. Until recently, however, most surveys on sexual behavior and fertility--including almost all fertility surveys conducted in developing countries from the 1960s through the early 1980s--focused only on women.2 Consequently, knowledge about sex-based differences in age and marital status at first intercourse has been derived largely from qualitative research and anecdotal evidence, rather than from empirical, quantitative data.

In some industrialized countries, starting in the mid-1970s, as researchers became concerned about patterns of adolescent childbearing and abortion--and, later, about the spread of STDs--they began to examine the sexual attitudes, knowledge and practices of both male and female adolescents.3 However, until the 1980s, the study of male sexual behavior was still rare, even in the industrialized world.4

Whatever the reasons for this limited focus, the substantial neglect of men's sexual behavior did not begin to change until realization of the scale and seriousness of the global HIV and AIDS epidemic became widespread. By the mid-1980s, changing epidemiologic concerns and an increased awareness of the importance of men's reproductive role and health needs led to heightened interest in including men in family planning and fertility surveys, and various surveys began to interview men as well as women.

In addition to large-scale surveys conducted at the national or city level, a number of smaller scale studies and country or regional reviews of male and female adolescent sexual behavior have contributed to a growing understanding of the complex social issues that influence early sexual behavior.5 Some of these studies and reviews focus on issues of particular relevance in certain parts of the world: young women's participation in commercial or informal prostitution in Asia and elsewhere; very early and sometimes forced marriages of women in South Asia; the phenomenon known in parts of Sub-Saharan Africa and the Caribbean as "sugar daddies"; and patterns of casual dating in Western countries that might involve an individual in sporadic sexual encounters or many consecutive short-term relationships. These inquiries are valuable because they increase our understanding of the possible circumstances surrounding the sexual behavior of adolescents; however, they do not yield quantitative information on behavior among the youth population as a whole.

While acknowledging the valuable contribution of qualitative research on young people's sexual and reproductive behavior, this article takes a different approach. It presents quantitative, population-level information from 14 countries on male-female differences in age and marital status at the time of first sexual experience. The countries included here represent all of the world's major geographic regions (Sub-Saharan Africa, Asia, Latin America and the Caribbean, and the developed world); all had national surveys containing comparable information on the sexual activity of young men and young women.

Measurement of Sexual Activity

Despite the progress made in increasing the representation of both sexes in national and community-level surveys on sexual and reproductive behavior, accurate measurement in this area remains beset with difficulties. Many survey respondents do not or cannot always openly or truthfully answer questions dealing with the intimate topic of their sexual behavior and practices. Understandably, adolescents, especially if they are unmarried or live in settings where sexual relationships outside marriage are censured, are probably even more likely than adults to be reticent about this area of their behavior. And very young teenagers, who are only just beginning to develop a sense of their sexuality, may be especially unlikely to want to discuss this part of their lives. However, the opposite problem also may be encountered; some young men overreport their sexual activity to give the impression that they are conforming to what they think society expects of them.6 These differential biases increase the difficulty of accurately comparing the experiences of males and females.

The fact of involuntary intercourse may also affect accuracy; misreporting is more likely when an individual's experience includes nonconsensual sex. Finally, methodological issues, including some that are linguistic or semantic, may play a role.7 For example, it is not always clear how respondents define for themselves terms used in studies of sexual behavior, such as "sexual intercourse."

It is difficult for researchers to evaluate the extent to which biases occur, but their concern is evidenced by the substantial literature dealing with these issues.8 Clearly, any interpretation of the data presented here must take into account that inaccuracies exist; whether the measures used are valid overall, and equally reliable for both sexes, can be tested only by more in-depth studies.

Nevertheless, the information in this article provides the best recent, nationally representative measures of sexual intercourse among young people in a range of countries. And since this analysis includes young adults aged 15-24, who are reporting on recent experience, the results should be little affected by respondents' ability to recall the dates of events or their age when particular events occurred.

Data Sources

The data were obtained from a variety of nationally representative sample surveys of men and women (Table 1). Eight of the countries participated in the Demographic and Health Survey (DHS) program, which collected information from men and women of reproductive age (15-44 or 15-49). Two had data on 15-24-year-olds from a Young Adult Reproductive Health Survey (YARHS), for which it was possible to obtain special tabulations; another two had information from independent national youth surveys with designs similar to that of the YARHS. Data for Great Britain came from the 1990-1991 National Survey of Sexual Attitudes and Lifestyles (NSSAL), which included men and women aged 16 and older. For the United States, data were from the 1991 National Survey of Men, which included 20-39-year-olds; the 1995 National Survey of Adolescent Men (NSAM), which covered 15-19-year-olds; and the 1995 National Survey of Family Growth (NSFG), which interviewed women aged 15-44.

Even though these surveys used somewhat different questions, their research protocols were generally similar enough that we could create comparable measures of age at first intercourse and of current sexual activity. The YARHS asked respondents in what month and year they first had intercourse; those who could not provide the date were asked their age at first intercourse. The DHS design generally requested only the age at first intercourse. The DHS and YARHS included very similar questions about recent sexual activity. Currently sexually active is defined as having had intercourse in the month before the interview, except for women in the United States, for whom the measure is based on intercourse in the three months before the interview.

Two of the U.S. surveys focused more specifically on heterosexual intercourse. The NSFG asked: "Think back to the very first time in your life that you ever had sexual intercourse with a man. In what month and year was that?" Likewise, the NSAM question on first intercourse was: "When did you have sexual intercourse with a female for the first time, in what month and year?" In Great Britain, the NSSAL took a similar approach: "How old were you when you first had sexual intercourse with someone of the opposite sex, or hasn't this happened?" In all other countries, an unknown, though probably small, proportion of reported sexual activity may be with same-sex partners.

Some differences in the design and execution of surveys may have had small effects on the quality of data. The British survey is likely to provide the best, most accurate source of information on sexual behavior, because this was its only focus; the researchers made intensive efforts to design a questionnaire that would maximize the quality of the data. By contrast, the other surveys covered a much broader range of topics, and questions on sexual behavior made up a relatively small part of the interview. The YARHS took steps to improve the quality of response--for example, by matching the sex of the interviewers with that of the respondents. In Ghana, however, interviewers were primarily male, and in other DHS surveys, they were mixed or primarily female.

All the surveys sampled unmarried and married persons, and asked all respondents, regardless of their marital status, questions about sexual behavior. (Marriage is defined to include consensual, or cohabiting, unions and, in Jamaica, visiting unions.) The surveys also asked the date of first marriage; premarital sexual intercourse could thus be measured by comparing respondents' dates or ages at first intercourse and at first marriage. A sexual relationship is classified as premarital only if it began at least one year prior to marriage. This strategy allows for inaccuracies that might result from respondents' reporting rounded ages. At the same time, this approach is conservative because it excludes sexual relationships that begin a few months prior to marriage.

Levels of nonresponse to questions on age at first intercourse and on patterns of recent sexual intercourse were quite low--typically less than 5%.9 This suggests that in general, respondents are not as uncomfortable discussing these issues as had been believed.


Marital and Nonmarital Sexual Initiation

In all of the study countries except the Philippines, Thailand and Peru, roughly one-third or more of young women 15-19 have had sexual intercourse; in four countries (Ghana, Mali, Jamaica and Great Britain), about three in five are sexually experienced (Figure 1). Between about one-half and three-quarters of young men aged 15-19 in seven countries (Brazil, the Dominican Republic, Haiti, Jamaica, Peru, Great Britain and the United States) have ever had intercourse, but one-third or fewer in Ghana, Zimbabwe, the Philippines and Thailand have done so.

The surveys reveal that the context of early sexual experience is often very different for young men and for young women, especially in developing regions. In all represented countries, sexually experienced adolescent women are much more likely ever to have been married than are sexually experienced adolescent men. For example, in Zimbabwe, where about one-third of both female and male 15-19-year-olds have had intercourse, two-thirds of these young women (21% overall) have been married, compared with almost no adolescent men.

The same pattern holds in most of the developing countries studied, regardless of gender differences in the proportions sexually experienced. Only Jamaica differs, in that relatively high proportions of both male and female sexually experienced teenagers have ever been in a union. The explanation for this surprising finding lies in a category of union in Jamaica that is not found elsewhere, the "visiting" relationship, in which partners have regular sexual relations but do not cohabit. This type of union is common and broadly approved by social custom in Jamaica, but is not equivalent to being in union, as defined in other countries, where only legal marriages and cohabiting unions are socially recognized.

In Great Britain and the United States, overall levels of sexual experience among young men are about as high as those in Brazil and Jamaica; levels for young women are comparable to those in Ghana, Mali and Jamaica. However, the vast majority of British and U.S. sexually experienced adolescents of both sexes have never been married (and most ever-married adolescent women became sexually active before marriage10).

When young adults reach their early 20s, their marital and sexual situations change dramatically (Figure 2). Overall levels of sexual experience rise to about 80% or more for women in seven countries (the four Sub-Saharan countries, Jamaica, Great Britain and the United States) and for men in all but two countries (the Philippines and Thailand). The Philippines is distinctive because of the low overall reported levels of sexual experience among both young men and young women--no more than around 50% of 20-24-year-olds of either sex are sexually experienced.

As is the case among teenagers, the context of sexual intercourse among those in their early 20s tends to be very different for men and women. Most sexual relationships among men remain nonmarital (again except in Jamaica), and most sexual intercourse among young women continues to occur within marriage. Young adult women in the Philippines and Thailand report extremely low levels of nonmarital sexual intercourse.

The Pace of Sexual Initiation

The survey data permit us to examine the pace at which men and women in their early 20s initiated their sexual lives. This analysis provides a more complete picture of adolescent sexual behavior than can be drawn using data collected from teenagers (who are, on average, about 17.5 years old at interview).

Among women, sexual experience before age 15 is rare.* In Ghana, Mali, Tanzania, Jamaica and the United States, at least one in seven women aged 20-24 had had sex before their 15th birthday; about three in 10 had done so before age 16, as had more than half in Mali (Table 2). Levels are lower--for the most part considerably so--in the remaining countries.

Very early sexual experience appears to be somewhat more common among men than among women in all countries except Ghana, Mali, the Philippines and Tanzania. More than four in 10 Jamaican men aged 20-24 had had intercourse before their 15th birthday, as had one-third of their counterparts in Brazil and the United States, and about one-quarter in Costa Rica and the Dominican Republic. Men in several other countries reached these levels of sexual experience by age 16, but in Zimbabwe, the Philippines and Thailand, even by their 16th birthday, no more than 15% of men had initiated intercourse.

In Ghana and Mali, substantially higher proportions of women (88% and 94%, respectively) than of men (70% and 64%, respectively) are sexually experienced by age 20. By contrast, in Thailand, Brazil, the Dominican Republic and Peru, the proportion of men with sexual experience by their 20th birthday is 18-28 percentage points higher than the proportion among women. In most of the remaining countries, gender differentials are very small, despite gender differences in marriage levels.

Country-level differences in the pace of sexual initiation can also be captured by comparing the median age at first intercourse (the age by which 50% of all individuals in the group had begun sexual relations). The median age for women ranges from 15.8 in Mali to 19.6 in Peru and exceeds 20 in the Philippines and Thailand. Contradicting the widely accepted view that males are generally more sexually precocious than females, a similar range is found among men--from 15.4 in Jamaica to 19.0 in Thailand and more than 20 in the Philippines. In Latin America and the Caribbean, men typically initiate sex at a younger age than women; the largest differentials (more than two years) are found in Brazil and Peru. Only in Ghana and Mali do women first have intercourse at a substantially younger age than men; in the remaining countries, men and women first have sex at roughly the same age.

Patterns of early sexual initiation (i.e., before age 17) vary widely by country and gender (Figure 3). Among women, the proportion having sex by age 17 in Mali, Jamaica, Ghana, the United States and Tanzania is 7-10 times that in Thailand and the Philippines. The proportion of men with sexual experience before their 17th birthday in Jamaica, the United States and Brazil is 9-10 times the level in the Philippines. The gender gap is very large in Mali and Ghana (where higher proportions of women than of men become sexually active early) and in the Dominican Republic, Brazil, Costa Rica, Peru and Thailand (where the reverse is true).

Premarital Intercourse Before Age 20

Table 3 examines levels of premarital sexual intercourse by the end of the adolescent years. The rates are also calculated by urban-rural residence, in an attempt to explore possible differences in sexual behavior related to modernization and to rapid social change, or its absence.

The proportion of women who had their first sexual experience before age 20 and while still single is 75-86% in the two developed countries and Jamaica; 38-58% in Ghana, Tanzania, Brazil and Haiti; and 25-32% in Mali, Zimbabwe, Costa Rica and Peru. It is much lower in the Dominican Republic than in the other Latin American countries (15%). However, since overall levels of sexual intercourse during adolescence are about the same in the Dominican Republic as elsewhere in the region, the lower level of premarital sexual intercourse probably reflects that higher proportions of women enter a union less than one year after sexual initiation or become sexually active within their first union.

In the Philippines and Thailand, premarital sexual relationships among young women are rare; only 4-6% of 20-24-year-olds initiated sexual activity as unmarried teenagers. While some women may have failed to report a premarital sexual relationship, these low levels, if typical of Southeast Asia, suggest that this region has a distinct pattern.

Except in the Philippines, the majority of young men started to have sex while they were single and before their 20th birthday. While the majority is a bare one in Mali and Thailand (56% and 54%, respectively), the level of premarital sexual intercourse among young men is substantially higher in all other countries; in Brazil and Jamaica, the proportion reaches roughly 90%.

Urban residence has a differential influence on young women's and young men's sexual and marriage behavior. For example, in Zimbabwe, Thailand, Costa Rica, the Dominican Republic and Peru, the total proportion who initiated sexual activity by their 20th birthday (i.e., combining those who did so within and outside marriage) is considerably lower among urban women than among rural women. In contrast, urban men in all countries are about as likely as or more likely than their rural counterparts to have had sex by age 20; the differentials are not very large, except in Mali.

In only two countries (Mali and Brazil) did a substantially higher proportion of urban than of rural women first have intercourse as unmarried teenagers. Furthermore, among women in Zimbabwe, Haiti, Jamaica and Peru, this proportion is lower among urban than among rural women. Among men, Mali, the Philippines, Haiti and Peru are the only countries in which urban dwellers appear somewhat more likely than rural residents to have had premarital intercourse during adolescence (a difference of 10 percentage points or more). These data do not support the view that urban dwellers are more susceptible than residents of rural areas to changing values that increasingly tolerate premarital sexual relationships.

One explanation for these findings may be that the expectation of early marriage is lower for urban than for rural women, and that in many settings, urban women may be expected to complete secondary school before marrying. On the other hand, men may have more opportunities for entering nonmarital relationships, and in both urban and rural settings, they are almost equally unlikely to marry before age 20. The proportion of women who become sexually active within marriage is substantially higher among rural than among urban women in all but three countries for which these data are available.

Some young people do not become premaritally sexually active before age 20, but follow other paths. Some 20-24-year-old survey respondents became sexually active within marriage and before age 20; some, within marriage but after age 20. Others entered into premarital sexual relationships after age 20, while some were still unmarried and sexually inexperienced.

Each of these groups has different reproductive health and education needs. For example, young men and women who begin sexual intercourse as adolescents and before marriage--a substantial group in almost all countries, and a majority in several--have an urgent need for contraceptive services. Married adolescent women who begin childbearing soon after marriage need prenatal and maternity services.

Nonmarital Sexual Intercourse

Much sexual intercourse among never-married adolescents is sporadic. In settings where young men frequent commercial sex workers, where teenagers change partners with some regularity or where young migrant workers have sexual relationships wherever they happen to be living, young people's sexual activity likely occurs at widely spaced intervals or with a number of short-term partners.

Although relatively high proportions of never-married teenagers and young adults have had sex, much lower proportions are currently in a sexual relationship. For example, in Ghana, where 49% of never-married adolescent women have ever had sexual relations, only 23% have done so in the past month (Table 4). In fact, the proportion of adolescent females who are currently sexually active generally represents only about one-third to one-half of all those who have initiated sex.

Sporadic sexual intercourse also appears to be common among never-married men aged 15-19. In many developing countries, the proportion in this age-group who are currently sexually active is half--or even less--the proportion who have had some sexual experience.

A pattern of sporadic sexual activity characterizes young adults in their early 20s as well, especially in developing wregions. No more than about one-third of never-married sexually experienced women aged 20-24 are currently sexually active in six of 12 countries with this information. By comparison, about half or more of sexually experienced young men are currently sexually active in nine of the 11 countries with information.

The Partner Relations (PR) surveys, carried out by the World Health Organization, also measured current sexual intercourse among never-married adolescents; while the measure used in these surveys is not strictly comparable to the measure we used, the results in countries included in our analysis show very similar patterns. For example, according to the PR survey in Thailand, 29% of never-married men and 1% of never-married women aged 15-19 had had sexual intercourse in the 12 months preceding the survey. In Manila (Philippines), the proportions were 15% among adolescent men and 0% among adolescent women. And in Rio de Janeiro (Brazil), these proportions were 61% among never-married males 15-19 and 9% among females.11


Sexual activity takes a very different course for young men and young women in most of the 14 countries included in this study. Among men, most sexual relationships during the teenage years are nonmarital. Among women, a good proportion--and the largest proportion in some countries--occur within marriage. Moreover, because of these tendencies and because of age differences between partners, adolescent sexual activity is not always more common among men than among women, even though in some countries, males are more likely than females to become sexually active at a very early age. For never-married young people, particularly men, sexual intercourse appears to be very sporadic; therefore, it probably involves a number of partners over time.

Are the findings of this study--especially those pertaining to very early age at sexual initiation--to be believed? It may be difficult to accept that at least one in four young men in Brazil, Costa Rica, the Dominican Republic, Jamaica and the United States have had intercourse before their 15th birthday. Nevertheless, even after allowing for some exaggeration, an early start to male sexual activity probably is not uncommon in these countries.

Is it also to be believed that similar levels of sexual experience are not attained by Thai males until their 17th year and by Filipino males until their 19th year? It is somewhat reassuring that the PR surveys support this finding. However, in both types of surveys, unmarried young men in these two countries may underreport their sexual activity, because of cultural reasons.

And what of the findings showing that 30% or more of females in Ghana, Mali, Tanzania, Jamaica and the United States have initiated sexual relationships by the end of their 15th year? With whom are these girls having sex, and how old are their partners? And are these sexual relationships consensual? These questions cannot be answered until more probing qualitative research is conducted among young adolescents. Data from the YARHS surveys in Latin America show that the partners of adolescent women who have a premarital sexual relationship are typically 3-4 years older; that is, young women's partners are in their early 20s.12 In the United States, 60% of the partners of adolescent mothers aged 15-17 are three or more years older.13 In some countries, such as Mali, the large majority of adolescent women become sexually active within marriage. However, the age gaps between spouses are often large: Husbands are, on average, 4-12 years older than their wives in 13 countries of Sub-Saharan Africa, and 3-8 years older in five developing countries in other regions.14 What are the implications for the autonomy of young women in such marriages?

Other perturbing questions remain unanswered. In countries where very high proportions of young men initiate sexual activity before age 15 and young women postpone having sex, are young men initiating intercourse mostly with prostitutes? How common is it for adolescents to have multiple sexual partnerships over a short period of time, or to have overlapping partnerships? Evidence from Latin America suggests that this is a common pattern and that young men who begin sexual intercourse at very young ages are more likely than others to have their first sexual experience with a prostitute.15 This factor, along with the findings that men are likely to have sporadic relationships and to have several partners during their adolescent and young adult years, suggests an increased chance that they will become infected with an STD (possibly HIV) by the time they marry.

Many health experts acknowledge that sexual experience during adolescence in and of itself does not necessarily involve risk. If young couples use effective methods of protection against pregnancy and STDs, and if their decision to enter into a relationship is voluntary, physical and psychological risks are minimized. The concern is pressing, however-- especially for women--when sexual relationships are involuntary or not sanctioned by law or custom and when contraceptive use is absent or inadequate.

Nonmarital sexual relationships at any age can expose women, particularly those who are poor or lacking education, to an uncertain future if they become pregnant or give birth. But adolescent women, who are usually not yet able to support themselves, let alone any children they might have, are likely to be even more undermined by the lack of social, financial and legal support associated with nonmarital relationships.

The consequences of nonmarital sexual intercourse are somewhat different for young men than for young women. For young men who are not in a mutually monogamous relationship or who do not use condoms consistently, the severest problem is the high risk of contracting or transmitting STDs. Furthermore, if unprotected coitus occurs and the young woman becomes pregnant, and if the young man does not walk away from the situation, his life plans are likely to be thrown into disarray.

Both the conditions and the implications underlying early sexual relationships among adolescents are likely to be widely dissimilar from one cultural setting to another. In developed and developing regions, many nonmarital adolescent sexual relationships are mutually voluntary and monogamous. Some young people use effective protection and prevent pregnancy and infection; some relationships soon lead to marriage; and some adolescent women who marry postpone their first birth.

However, many early sexual relationships occur within a less-desirable context and involve almost certain risk. Some young men have their first sexual encounter with a prostitute at an early age and then have sporadic relationships with different partners in their later teens. Some young women become sexually active at an early age as a result of rape, incest or coercive pressure to engage in sexual intercourse or to enter prostitution. In some countries where schoolgirls are likely to exchange sexual favors for financial reasons (e.g., to help with tuition payments), an unmarried adolescent woman's partner could be a much older, married man.

The recent trend of an increasing age at marriage among women suggests that levels of sexual activity during adolescence may be declining. However, the picture is mixed: An analysis of data from 36 developing countries shows that although the proportion of women who marry by age 18 has declined in all but a few countries, the proportion who initiate sex by age 18 has dropped much less; as a result, a growing proportion of women are becoming sexually active before marriage and at a young age.16

The current situation regarding contraceptive use among adolescents is also mixed: In some countries, use is increasing substantially, but levels are still low. For example, in Ghana, the proportion of married adolescents using a modern contraceptive method increased from 2% to 7% over the past decade.17 The moderate levels of use found among married adolescents in some countries (30-46% in Brazil, Colombia, Indonesia, Thailand and Zimbabwe) are encouraging. In addition, in most countries of Sub-Saharan Africa, the majority of female adolescent contraceptive users are unmarried, and unmarried sexually active adolescents are more likely than married adolescents to know of, approve of and ever have used contraceptives. Even so, in a study of 26 developing countries, only in five were 25% or more of unmarried sexually active adolescents using a method.18

The data reported here, though limited, show that in most of the countries included, a high proportion of adolescents and young adults are potentially at risk for a range of poor reproductive health outcomes. Their actual level of risk depends on whether they are using effective means of protection against both pregnancy and STDs, and using them consistently.

If educators and planners are to be able to provide for the reproductive health needs of adolescents, it is clearly important to encourage quantitative and qualitative research that will provide more accurate measures of teenagers' sexual behavior and a better understanding of the circumstances surrounding their sexual relationships. The quantitative data presented here take a first, essential step in that direction.