This study is aimed at understanding premarital sexual activity in India using data from the third round of the National Family Health Survey. At the national level, reported premarital sex is still fairly low among women and somewhat higher among men. The data show that ever married women are more likely to report premarital sex compared to currently unmarried women. Among unmarried women and men who report premarital sexual activity, the distributions vary by place of residence, occupational status, level of education, and level of household wealth. Unmarried women who are aware of hiv/aids and sexually transmitted diseases are less likely to have had premarital sex. Notably, there is a significant gender dimension in reporting of premarital sexual activity, pointing the way for further research.
NATIONAL FAMILY HEALTH SURVEY-3november 29, 2008 EPW Economic & Political Weekly54Premarital Sex in India: Issues of Class and GenderLekha SubaiyaThis study is aimed at understanding premarital sexual activity in India using data from the third round of the National Family Health Survey. At the national level, reported premarital sex is still fairly low among women and somewhat higher among men. The data show that ever married women are more likely to report premarital sex compared to currently unmarried women. Among unmarried women and men who report premarital sexual activity, the distributions vary by place of residence, occupational status, level of education, and level of household wealth. Unmarried women who are aware ofHIV/AIDS and sexually transmitted diseases are less likely to have had premarital sex. Notably, there is a significant gender dimension in reporting of premarital sexual activity, pointing the way for further research.Lekha Subaiya (lekha@isec.ac.in) is at the Population Research Centre, Institute for Social and Economic Change, Bangalore.In the last two decades, the attention devoted to research on sexuality in India has been on the rise, almost solely due to the advent of AIDS. Until then, sexuality was considered a private matter in traditional, family-oriented India, and sex outside marriage was treated as non-existent in the public arena. IntroductionThe pattern of the spread of HIV/AIDS in India challenged the widely held belief that there were low rates of multi-partner sexual activities and other risky sexual behaviour in the country. Since then, researchers have paid more attention to non-marital sexu-ality in India, albeit from the perspective of sexual behaviourasit relates to sexual health (for example, Nag 1996; Verma et al 2004). However, the issue of premarital sex in India remains a poorly explored topic. Not enough is known about the levels, trends and regional patterns in sexual activity before marriage in India. A review of the literature on adolescent sexuality by Jejeebhoy (2000) showed that anywhere up to 10% of unmarried girls and women and 20-30% of unmarried boys and men have been sexu-ally active. Given the conservative attitude towards non-marital sexuality, even this vague figure for premarital sex is revealing. Further, there are indicators that the numbers are likely to be higher than those reported by women, especially in poor and rural areas. In a community-based study of rural tribal women in Maharashtra, Bang et al (1989) found that nearly half of all un-married girls were sexually active. The methodology used in thisstudy included women’s self-reported clinical histories and physical examinations. At the same time, in-depth interviews and reports from men suggest that women are likely to under-report premarital sexual activity. Bhende (1994) found that there were higher rates of sex-ual activity among poorly educated adolescents in a slum area of Mumbai than was directly reported by them. In-depth interviews of medical practitioners and community leaders, along with young boys and girls and their mothers, indicated more activity than was reported by the adolescents and their mothers. Inter-views with men also provided indirect evidence that they find sexual partners among the unmarried girls within their own community. Male respondents in a qualitative study in rural Gujarat indicated that unmarried girls in their own village were available for sex (Joshi et al 2004). That women are likely to have under-reported in these studies illustrates that there are strong cultural norms against young girls and women engaging in pre-marital sexual activity. The arranged marriage system still dominates Indian culture, and chastity is highly valued within it. According to societal and familial norms, premarital sex is not allowed, and families go to
NATIONAL FAMILY HEALTH SURVEY-3Economic & Political Weekly EPW november 29, 200855great lengths to protect the chastity of unmarried youth, espe-cially girls. In this scenario, it is necessary to understand the linkages to premarital sex. It is important to study sexuality in the Indian context, and changes thereof, because the start of sexual activity coincides with the time that one transitions to adulthoodand begins to take on adult roles. Sexual exploration should be entered into with full knowledge and awareness so that individuals can make healthy choices for themselves. In the absence of the ability to make healthy choices through lack of knowledge or lack of power, or both, the consequences can be dire. These include sexual coercion, unwanted pregnancy, abortion and its consequences, and sexually transmitted diseases (STDs) (Joshi et al 2004; Mehta et al 2004). Premarital Sex in IndiaIn general, Indian men and women are not expected to have sex before marriage. While there is some laxity with regard to men’s sexual behaviour, women’s chastity is still greatly valued. How-ever, studies show that there is premarital sexual activity in the country, with the rates being higher for men than women in every study. As mentioned earlier, Jejeebhoy’s review of studies on adolescent sexuality (2000) showed that the rates for women ranged from 0-10% of unmarried girls and women and from 20-30% of unmarried boys and men, depending on the study. These studies varied in type from self-administered question-naires in magazines and respondent interviews to reports from clinical examinations. They were not comprehensive and tended to cover sub-populations. The first type of study tended to cover urban upper and middle class populations due to self-selection in readership, while the community studies were mostly of students, and poor and tribal populations. In the popular media, there is now a sense that sex before mar-riage is on the rise with the social and economic changes brought about by globalisation. The increasing exposure of youth to western culture is thought to have effected a change in moral attitudes towards sex before marriage. That premarital sex is entering popular culture was reflected in a mainstream Hindi movie which had the protagonists, played by popular actors, living together and having a child before marrying each other. Further, certain trends such as the increase in urbanisation, financial independence among young women through employ-ment, and the age at marriage support this argument, as do studies gauging the attitudes of youth to premarital sexuality. A 1991 study, which surveyed school students who were in the ninth to 11th standards in urban and rural areas of Uttar Pradesh, Rajasthan, Haryana and Delhi through self-administered questionnaires, found that only a marginal majority (51%) agreed with the state-ment “unmarried people should not have sexual intercourse” (FPF andORG 1992). However, on breaking down the group by gender, the trend towards liberal moral attitudes was more pronounced among male students than female ones.While these studies focused on establishing the levels of non-marital sexual activity in a socially conservative society to link it with sexual health, and not to determine the correlates of pre-marital sexual activity, the findings do lend themselves to specu-lation on the probable motivations. There is some evidence from qualitative studies that women see a premarital relationship as a last chance at romance and emotional happiness before an arranged marriage removes the ability to focus on their own needs. Researchers who analysed data collected for the Sexual Health Project by the Deepak Foundation in a rural area of Gujarat suggest that the period between the onset of menarche and marriage presents a “window of opportunity” for fulfilling women’s needs for love and tenderness through sexual relation-ships (Mehta et al 2004). The study also found that gifts of trinkets and cash were made to women in non-marital sexual relation-ships by their partners and that these were desirable, which suggest that economics can be a factor in such relationships.An ongoing project by the Population Council and the Indian Institute of Population Sciences (IIPS), Mumbai, is in the process of collecting data on youths aged 10 to 29 in six states – Bihar, Maharashtra, Tamil Nadu, Andhra Pradesh, Jharkhand and Rajasthan. To date, information in the form of fact-sheets is publicly available for Maharashtra, Tamil Nadu and Jharkhand (Popula-tion Council andIIPS 2007). These population-level data reveal that there is premarital sex among female and male youth, but that at the same time, a very marginal percentage of the respond-ents in each state is engaged in discussions on sexual and repro-ductive systems, including pregnancy, by their parents. These two trends are suggestive of the situation in India – while sexuality is rarely talked about in the open and often considered to be non-existent, in reality, the situation is different.Knowledge and awareness of STDs is significantly low in the Indian population. Data from the Reproductive and Child Health Study indicate that only 44% of women and 53% of men are aware of STDs (IIPS 2006). Awareness ofHIV/AIDS is higher, with 60% of reproductive age women, and 84% of men in the third National Family Health Survey (NFHS-3) having heard of it (IIPS and Macro International 2007). It is notable that while awareness of HIV/AIDS is higher among younger women and men, the young-est age group is the least likely to know about STDs. HIV/AIDS, which has received much attention in the media, is better known asSTDs in general. Premarital sexual activity in a society which sanctions against it can have serious implications for those without the power of self-determination. Societal control of the sexuality of unmarried girls results in a blinding lack of information on issues relating to it. Girls are unable to develop the ability to determine a healthy sexuality for themselves. Sexual expression in the absence of knowledge and the ability to make healthy decisions can result in sexual abuse and exploitation. There is evidence from interviews with men of sexual coercion of unmarried adolescent females despite this group being strictly supervised, or even because of it(Sodhi et al 2004). Other obvious implications for girls who engage in premarital sex and at the same time operate in an environment where it is taboo are unwanted pregnancies and abortions, and the resulting physical and social complications (Mehta et al 2004; Joshi et al 2004). An intervention programme by the group Swaasthya in a settlement in Delhi in 1999-2001 that targeted adolescent girls and women to develop their know-ledge and coping skills with regard to reproductive health and sexual risk worked effectively to make them less vulnerable to
NATIONAL FAMILY HEALTH SURVEY-3 november 29, 2008 EPW Economic & Political Weekly56negative sexual and health outcomes (Sodhi et al 2004). Such programmes to increase knowledge and awareness of sexual and reproductive health, which enable females to determine their actions, are an obvious place for policy intervention.To the extent that choice is being exercised, higher levels of premarital sexual activity represent a breaking down of social norms. If moral attitudes towards sex have changed, and young women and men are having sex out of choice, this trend could have long-term implications for a society like India. It would sug-gest the formation of new gender relations, in which the ability to make decisions is possessed by both. On the other hand, given the class and gender hierarchies that are present in India, the ability to make this decision is likely to be tempered by the position anindividualoc-cupies in the social strata. In other words, one’s socio-economic class and one’s gen-der is likely to determine how much power one has in making decisions about sex. Until now, population-based, national-level data on unmarried men and women have not been available. TheNFHS-3 in-cluded unmarried women and men in its survey population. This study is aimed at understanding premarital sexual activity in India. A major limitation of the study is that sexual activity per se is limited to actual sexual intercourse, whereas in actu-ality, the term is wider in scope and in-cludes a range of activities, including petting, hugging, kissing, and oral sex. Data and MethodsThe data used for this study are from the NFHS-3 conducted in two phases in India under the aegis of the Demograhics and Health Surveys (DHS) programme from November 2005 to August 2006. The DHS are nationally representative, population-based surveys of women and men of reproductive age. Information is collected on various populations and health char-acteristics, including fertility,family planning, child mortality, maternal and child health, nutrition, and individual and household background characteristics. A final report based on the findings of the NFHS-3was released in 2007 (IIPS and Macro International 2005-06). For the NFHS-3, a total of 124,385 women aged 15 to 49 years and 74,369 men aged 15-54 years from 109,041 households were interviewed. This analysis includes all women and men, unmarried and married. For the first time, information was collected from both married and unmarried men and women in the NFHS-3. There were 25,462 women and 25,382 men among the unmarried re-spondents for whom there was information on age at first inter-course, and 96,000 women and 48,988 men among the married respondents for whom there was information on age at first in-tercourse and age at marriage. From this information, it is possi-ble to gauge whether the respondents had intercourse before marriage. The methods used in the analysis are solely descriptive in nature, and the aim of this section of the study is to describe premarital sex across all the states by various social, demographic and economic characteristics. To gauge the reliability of the data, the rates from the NFHS-3 are compared with that from the in-depth study of youth aged 15 to 24 years mentioned earlier. The Youth: Situations and Needs study is being carried out in six states and data for three are publicly available in the form of fact sheets(Population Council and IIPS 2007). Table 4 (p 59) shows the comparable figures. For each state, the rate for the study is reported, followed by the rates from the NFHS-3 for each demographic group. The percentage is of the rate for the NFHS-3data to that of the youth study. It can be seen that in all cases but one (Maharashtra) the rates from the NFHS-3 data are less than that of the youth study. It can also be seen that for each state, unmarried women were more likely to under-report than married women. This comparison shows that a national-level survey on demography and health as a whole is likely to underestimate levels of non-marital sexuality compared to a survey that targets the sexual and reproductive health of a particular population, in this case, youth.Premarital Sex by GenderThe percentage of women and men who report premarital sexual intercourse are presented in Table 1. The data shows that about 1.8% of all women and 12% of all men report that they had sexual inter-course before marriage. The wide disparity between men and women in the rates of premarital sex indicates that there are dif-ferent rules for men and women with regard to non-marital sexuality in India. When breaking the numbers down by marital status, it is apparent that signi-ficantly more ever married women (2%) report premarital sex compared to never married women (0.7%), whereas slightly more never married men report premarital Table 1: Percentage of Persons Who Report Having Sexual Intercourse Prior to Marriageby Selected Characteristics (NFHS-3, 2005-06) WomenMenMarital status Never married 0.7 13.5 Evermarried2.011.2Age 15-19 0.8 8.8 20-24 1.5 16.0 25-29 1.814.9 30-342.212.5 35-39 2.311.5 40-442.59.5 45-492.210.0 50-54 – 10.4Education None 2.012.4 Primary1.813.0 Secondary or higher 1.5 11.6Media exposure No or infrequent 2.0 11.2 At least once a week 1.6 12.2Occupation Do not work 1.4 6.7 Have an occupation 2.3 12.7Residence Rural 1.812.6 Urban1.710.9Wealth quintile Lowest 2.2 12.2 Second1.813.6 Middle1.712.7 Fourth1.611.9 Highest1.610.1State* HimachalPradesh2.2– Punjab2.717.2 Uttarakhand4.7– Haryana3.316.4 Uttar Pradesh – 14.5 Sikkim – 25.5 ArunachalPradesh4.023.9 Nagaland6.127.7 Mizoram10.245.3 Tripura4.418.1 Jharkhand9.8– Orissa –12.3 MadhyaPradesh–12.7 Gujarat2.615.2 Maharashtra4.312.8 AndhraPradesh–15.6 Goa 4.7– Kerala3.314.6 TamilNadu2.8–Total 1.812.0Sample size 124,385 74,369* Only states with higher than average percentage shown.
NATIONAL FAMILY HEALTH SURVEY-3Economic & Political Weekly EPW november 29, 200857sex (13.5%) than ever married men (11.2%). Among women, the percentage of those who reported premarital sex increases mar-ginally with age, from 0.8% of those aged 15 to 19 years, to 2.2% among those aged 45 to 49 years. For men, there is no clear rela-tionship with age, with those aged 20 to 24 years (16%) and those aged 25 to 29 (14.9%) reporting higher than average levels.While there appears to be a strong relationship between edu-cation and household wealth and the levels of premarital sex among women, these relationships are less clear for men. The percentage of women who report premarital sexual intercourse decreases with an increase in their level of education as well as with an increase in the wealth quintile their household is in, albeit marginally. But there is no discernible relationship between pre-marital sex and men’s level of education or wealth. Both men and women show substantial differences by occupa-tional status (whether the respondent has an occupation or not), with more of those who have an occupation reporting premarital sexual activity. Both men and women show little variation whether their residence is in a rural area or not.Table 1 also lists the states in which a higher than average per-centage report premarital sexual activity. More women in the north-eastern and south-western states report premarital sexual activity compared to women in other states, with the highest per-centages in Mizoram and Orissa. More men in the north-eastern states of Mizoram, Nagaland, Arunachal Pradesh and Sikkim as well as in Punjab and Haryana report premarital sexual activity.TheNFHS-3 collected data on eight cities, and percentages of premarital sexual acti-vity in each city are presented in Table 2. The results show that very few cities report higher than average premarital sexual activity. For women, these are Mumbai (4.2%) and Chennai (3.4%), while for men they are Meerut (11.9%) and Mumbai (11.7%). In only one of the cities is the level of premarital sexual activity higher than the state average, Chennai. Premarital Sex by Marital StatusTable 3 (p 58) shows the level of premarital sex according to social and economic char-acteristics and marital status. The distri-butions for unmarried women and men show that the level of premarital sexual activity for both groups varies by place of residence, occupational status, level of education, and level of household wealth. Women who live in rural areas are more likely to have had premarital sexual intercourse (0.9%)thanthose who live in urban areas (0.4%), those who have anoccupationhave higher levels of premarital sexual activity (1.2%) than those who do not (0.4%), and the percentage decreases with an increase in the education and wealth quintile. These relationships are also found among unmarried men. More men in rural areas report having had sexual intercourse (14.7%) than those in urban areas (11.9%), those who have an occupation have higher levels of premarital sexual activity (17.1%) than those who do not (7.2%), and the level of premarital intercourse decreases with an increase in the level of education.The distributions for married women show little variation by residence, occupational status and level of education, and while it does not vary for men, there is no clear pattern to any of these variables. Also, while the percentage who report premarital sexual activity generally decreases with an increase in the quintile of household wealth for married women and has an inverted U-shape for married men, again, the variation between levels is marginal. In sum, the results of bivariate analysis show that for unmarried women and men there is a variation in the incidence of premarital sexual activity by level of education, employment and level of wealth of the household, while the same is not true for married women and men. These results suggest that economic status is a significant factor for unmarried women and men but not so for married women and men. At the same time, for three of the five questions on attitudes towards sex roles, unmarried women who give gender equitable responses are less likely to have had premarital sexual intercourse than women who agree that a husband is justified in beating his wife. Again, there is no variation for married women. On the other hand, men who support more equitable gender roles are more likely to report premarital sexual activity. It is notable that for both groups of men, unmarried and married, the pattern of responses to this set of questions is similar.The distribution by marital status is im-portant for women, because while it varies by social and economic characteristics for unmarried women, it does not vary very much by the same characteristics for married women. So, a closer inspection of marital status and sexual activity is neces-sary. The comparison of age at marriage and age at intercourse proves to be reveal-ing. Of the 2% of married women whose age at first intercourse came before age at marriage, 1.6% had intercourse in the year before marriage and only 0.4% had inter-course earlier than that. The latter number is closer to the percentage of unmarried women who report that they have had sexual intercourse. This finding suggests that for a significant number of women who have premarital sexual intercourse, sexual activity leads to marriage. Either women marry the person they have sexual intercourse with (or have sexual inter-course with the person they are affianced to) or they get married subsequently. Thus sexual intercourse transitions into marriage for a portion of women. Class and Gender DimensionsEconomic class and gender are two dimensions of socio-economic inequality in India which interact with each other, and have to be examined separately as well as in relation to each other to Table 2: Percentage of Persons Who Report Having Sexual Intercourse Prior to Marriage by Selected City and State (NFHS-3, 2005-06)State/City WomenMenDelhi 0.5 10.3Delhi city 0.8 10.9Uttar Pradesh 0.8 14.5Meerut 0.311.9West Bengal 0.7 7.8Kolkata 0.63.2Madhya Pradesh 0.3 12.7Indore 0.28.9Maharashtra 4.3 12.8Mumbai 4.211.7Nagpur 2.210.7Andhra Pradesh 0.1 15.6Hyderabad 0.16.2Tamil Nadu 2.8 9.7Chennai 3.49.2Total % within cities sampled 2.0 9.6Total number in city sample 20,508 16,518Total % 1.8 12.0Total size 124,388 74,368
NATIONAL FAMILY HEALTH SURVEY-3 november 29, 2008 EPW Economic & Political Weekly58understand their influence on sexual health. There are large dif-ferences in access to economic resources as well as to information and care in the health system by economic class, including sexual and reproductive healthcare. Gender refers to the cultural values, roles, practices and characteristics that are ascribed to each sex. In each society, there are different expectations of boys and girls, which have an impact on sexual behaviour, sexual responsibilities, and the ability to access information about sex and healthcare (Rao Gupta 2000). Sexuality in India has to be understood in terms of economic class and gender differentials, since each di-mension influences an individual’s behaviour through the power he or she has to acquire information as well as make decisions.The percentages of premarital sexual activity among unmar-ried and married women found in the NFHS-3 raises an interest-ing question: how much of the pattern in the relationships between premarital sex and age, education, and wealth status were caused by these factors, and how much by biases in reporting, primarily because of gender and class issues.It is more than likely that the data that has been collected on sexual intercourse before marriage has been seriously under-reported, especially by women. Even though the anonymity of the interview was a required component of the survey procedure, anecdotal evidence from the field indicates that there was resistance to ques-tions about sexual intercourse, albeit in the domestic violence module. In each instance, community leaders and officials had to be placated with assurances that the questions on sexual inter-course were part of a health survey before the survey was allowed to proceed. It is also notable that unmarried women interviewers said they were greatly embarrassed at having to ask questions about sexual intercourse. If these are examples of overt resistance to the open mention of sexual intercourse, it is likely that the re-sponses from the interviewees was also inhibited. All this indicates that the understanding of the gender dimension operates even at the point of data collection, leave alone at the level of reporting.The higher rates of response from married women are sugges-tive of biases in reporting. Of course, married women include women of all ages, and there are more women in the older ages in the married group than in the unmarried group. However, a com-parison of women by marital status and age group suggest that in the younger ages (15 to 24), the rates are more similar between the two groups, though somewhat higher for married women. The higher rates in the older ages could be either due to the “abil-ity” to report on a sensitive topic like premarital sex or due to the higher likelihood of premarital sex among the older cohort and those with a later age at marriage. If there is no error in reporting the two measures, it might also be because it is easier for married women, who are expected to have had sex, to talk about it and admit to premarital sex. It is assumed that class has an influence on an individual’s abil-ity to access resources, including information and care in sexual and reproductive health. A person with more economic resources is more likely to have access to knowledge, and be more able to make healthy decisions about sex than a poor person. At the same time, when it comes to sex and women, gender operates within the context of a complete and enduring silence. As such, the inter-linkages between class and sexual choices are complex in a cul-ture with strong unwritten rules and regulations on womens’ bodies. So, the pathways through which class affects women and the negative consequences of ill-informed decision-making are an important aspect of sexuality in India, which need careful analy-sis and in-depth discussion. However, the overwhelming effect of gender on reporting makes it difficult to do a comprehensive class analysis. At the same time, the incidence of premarital sexamong Table 3: Percentage of Women and Men Who Report Sexual Intercourse Prior to Marriage by Selected Characteristics(NFHS-3, 2005-06) Unmarried Married UnmarriedMarried Women Women Men MenAge 15-24 0.7 1.3 11.9 14.0 25-34 0.9 2.1 20.3 12.0 35+ 1.7 2.4 23.4 10.2Age at marriage (women/men) Below 15-18 years – 0.7 – 6.1 14 to 18 years/19 to 21 years – 1.8 – 11.4 19+ years/22+ years – 3.5 – 13.1Education None 1.6 2.1 18.6 11.3 Primary 1.2 2.0 17.0 11.7 Secondary 0.5 2.0 12.7 11.5 Higher 0.3 2.0 11.5 8.3Occupation Do not work 0.4 1.7 6.2 12.1 Have an occupation 1.2 2.5 17.3 11.1Attitude to sex roles: Beating one’s wife Justified if wife goes out without permission 0.9 2.1 12.8 10.4 No 0.7 2.1 16.5 13.8 Justified if wife neglects the children 0.7 2.3 12.4 10.1 No 0.7 1.9 16.4 14.0 Justified if wife argues with her husband 1.1 2.2 12.3 10.4 No 0.6 2.0 17.3 13.4 Justified if wife refuses sex 1.5 2.0 13.2 10.8 No 0.7 2.1 20.6 15.6 Justified if wife burns the food 1.0 2.3 13.1 10.8 No 0.7 2.0 16.8 14.0Knowledge of sexually transmitted diseases Heard of STDs 0.6 2.0 14.1 11.8 No 1.2 2.1 8.1 8.4 Heard of HIV/AIDS 0.5 2.1 14.1 11.8 No 1.3 2.0 8.6 8.8Exposure to TV No or infrequent exposure 1.1 2.0 14.5 10.6 Almost every day 0.5 2.0 12.5 11.8Residemce Rural 0.9 2.0 14.7 11.6 Urban 0.4 2.2 11.9 10.2Wealth quintile Lowest 1.6 2.2 16.3 10.8 Second 1.2 1.9 16.6 12.2 Middle 0.8 2.0 13.8 12.0 Fourth 0.4 2.0 12.9 11.2 Highest 0.3 2.1 11.0 9.5Total % 0.7 2.0 13.5 11.1N 25,462 96,000 25,382 48,988
NATIONAL FAMILY HEALTH SURVEY-3november 29, 2008 EPW Economic & Political Weekly54Premarital Sex in India: Issues of Class and GenderLekha SubaiyaThis study is aimed at understanding premarital sexual activity in India using data from the third round of the National Family Health Survey. At the national level, reported premarital sex is still fairly low among women and somewhat higher among men. The data show that ever married women are more likely to report premarital sex compared to currently unmarried women. Among unmarried women and men who report premarital sexual activity, the distributions vary by place of residence, occupational status, level of education, and level of household wealth. Unmarried women who are aware ofHIV/AIDS and sexually transmitted diseases are less likely to have had premarital sex. Notably, there is a significant gender dimension in reporting of premarital sexual activity, pointing the way for further research.Lekha Subaiya (lekha@isec.ac.in) is at the Population Research Centre, Institute for Social and Economic Change, Bangalore.In the last two decades, the attention devoted to research on sexuality in India has been on the rise, almost solely due to the advent of AIDS. Until then, sexuality was considered a private matter in traditional, family-oriented India, and sex outside marriage was treated as non-existent in the public arena. IntroductionThe pattern of the spread of HIV/AIDS in India challenged the widely held belief that there were low rates of multi-partner sexual activities and other risky sexual behaviour in the country. Since then, researchers have paid more attention to non-marital sexu-ality in India, albeit from the perspective of sexual behaviourasit relates to sexual health (for example, Nag 1996; Verma et al 2004). However, the issue of premarital sex in India remains a poorly explored topic. Not enough is known about the levels, trends and regional patterns in sexual activity before marriage in India. A review of the literature on adolescent sexuality by Jejeebhoy (2000) showed that anywhere up to 10% of unmarried girls and women and 20-30% of unmarried boys and men have been sexu-ally active. Given the conservative attitude towards non-marital sexuality, even this vague figure for premarital sex is revealing. Further, there are indicators that the numbers are likely to be higher than those reported by women, especially in poor and rural areas. In a community-based study of rural tribal women in Maharashtra, Bang et al (1989) found that nearly half of all un-married girls were sexually active. The methodology used in thisstudy included women’s self-reported clinical histories and physical examinations. At the same time, in-depth interviews and reports from men suggest that women are likely to under-report premarital sexual activity. Bhende (1994) found that there were higher rates of sex-ual activity among poorly educated adolescents in a slum area of Mumbai than was directly reported by them. In-depth interviews of medical practitioners and community leaders, along with young boys and girls and their mothers, indicated more activity than was reported by the adolescents and their mothers. Inter-views with men also provided indirect evidence that they find sexual partners among the unmarried girls within their own community. Male respondents in a qualitative study in rural Gujarat indicated that unmarried girls in their own village were available for sex (Joshi et al 2004). That women are likely to have under-reported in these studies illustrates that there are strong cultural norms against young girls and women engaging in pre-marital sexual activity. The arranged marriage system still dominates Indian culture, and chastity is highly valued within it. According to societal and familial norms, premarital sex is not allowed, and families go to 
NATIONAL FAMILY HEALTH SURVEY-3Economic & Political Weekly EPW november 29, 200855great lengths to protect the chastity of unmarried youth, espe-cially girls. In this scenario, it is necessary to understand the linkages to premarital sex. It is important to study sexuality in the Indian context, and changes thereof, because the start of sexual activity coincides with the time that one transitions to adulthoodand begins to take on adult roles. Sexual exploration should be entered into with full knowledge and awareness so that individuals can make healthy choices for themselves. In the absence of the ability to make healthy choices through lack of knowledge or lack of power, or both, the consequences can be dire. These include sexual coercion, unwanted pregnancy, abortion and its consequences, and sexually transmitted diseases (STDs) (Joshi et al 2004; Mehta et al 2004). Premarital Sex in IndiaIn general, Indian men and women are not expected to have sex before marriage. While there is some laxity with regard to men’s sexual behaviour, women’s chastity is still greatly valued. How-ever, studies show that there is premarital sexual activity in the country, with the rates being higher for men than women in every study. As mentioned earlier, Jejeebhoy’s review of studies on adolescent sexuality (2000) showed that the rates for women ranged from 0-10% of unmarried girls and women and from 20-30% of unmarried boys and men, depending on the study. These studies varied in type from self-administered question-naires in magazines and respondent interviews to reports from clinical examinations. They were not comprehensive and tended to cover sub-populations. The first type of study tended to cover urban upper and middle class populations due to self-selection in readership, while the community studies were mostly of students, and poor and tribal populations. In the popular media, there is now a sense that sex before mar-riage is on the rise with the social and economic changes brought about by globalisation. The increasing exposure of youth to western culture is thought to have effected a change in moral attitudes towards sex before marriage. That premarital sex is entering popular culture was reflected in a mainstream Hindi movie which had the protagonists, played by popular actors, living together and having a child before marrying each other. Further, certain trends such as the increase in urbanisation, financial independence among young women through employ-ment, and the age at marriage support this argument, as do studies gauging the attitudes of youth to premarital sexuality. A 1991 study, which surveyed school students who were in the ninth to 11th standards in urban and rural areas of Uttar Pradesh, Rajasthan, Haryana and Delhi through self-administered questionnaires, found that only a marginal majority (51%) agreed with the state-ment “unmarried people should not have sexual intercourse” (FPF andORG 1992). However, on breaking down the group by gender, the trend towards liberal moral attitudes was more pronounced among male students than female ones.While these studies focused on establishing the levels of non-marital sexual activity in a socially conservative society to link it with sexual health, and not to determine the correlates of pre-marital sexual activity, the findings do lend themselves to specu-lation on the probable motivations. There is some evidence from qualitative studies that women see a premarital relationship as a last chance at romance and emotional happiness before an arranged marriage removes the ability to focus on their own needs. Researchers who analysed data collected for the Sexual Health Project by the Deepak Foundation in a rural area of Gujarat suggest that the period between the onset of menarche and marriage presents a “window of opportunity” for fulfilling women’s needs for love and tenderness through sexual relation-ships (Mehta et al 2004). The study also found that gifts of trinkets and cash were made to women in non-marital sexual relation-ships by their partners and that these were desirable, which suggest that economics can be a factor in such relationships.An ongoing project by the Population Council and the Indian Institute of Population Sciences (IIPS), Mumbai, is in the process of collecting data on youths aged 10 to 29 in six states – Bihar, Maharashtra, Tamil Nadu, Andhra Pradesh, Jharkhand and Rajasthan. To date, information in the form of fact-sheets is publicly available for Maharashtra, Tamil Nadu and Jharkhand (Popula-tion Council andIIPS 2007). These population-level data reveal that there is premarital sex among female and male youth, but that at the same time, a very marginal percentage of the respond-ents in each state is engaged in discussions on sexual and repro-ductive systems, including pregnancy, by their parents. These two trends are suggestive of the situation in India – while sexuality is rarely talked about in the open and often considered to be non-existent, in reality, the situation is different.Knowledge and awareness of STDs is significantly low in the Indian population. Data from the Reproductive and Child Health Study indicate that only 44% of women and 53% of men are aware of STDs (IIPS 2006). Awareness ofHIV/AIDS is higher, with 60% of reproductive age women, and 84% of men in the third National Family Health Survey (NFHS-3) having heard of it (IIPS and Macro International 2007). It is notable that while awareness of HIV/AIDS is higher among younger women and men, the young-est age group is the least likely to know about STDs. HIV/AIDS, which has received much attention in the media, is better known asSTDs in general. Premarital sexual activity in a society which sanctions against it can have serious implications for those without the power of self-determination. Societal control of the sexuality of unmarried girls results in a blinding lack of information on issues relating to it. Girls are unable to develop the ability to determine a healthy sexuality for themselves. Sexual expression in the absence of knowledge and the ability to make healthy decisions can result in sexual abuse and exploitation. There is evidence from interviews with men of sexual coercion of unmarried adolescent females despite this group being strictly supervised, or even because of it(Sodhi et al 2004). Other obvious implications for girls who engage in premarital sex and at the same time operate in an environment where it is taboo are unwanted pregnancies and abortions, and the resulting physical and social complications (Mehta et al 2004; Joshi et al 2004). An intervention programme by the group Swaasthya in a settlement in Delhi in 1999-2001 that targeted adolescent girls and women to develop their know-ledge and coping skills with regard to reproductive health and sexual risk worked effectively to make them less vulnerable to 
NATIONAL FAMILY HEALTH SURVEY-3 november 29, 2008 EPW Economic & Political Weekly56negative sexual and health outcomes (Sodhi et al 2004). Such programmes to increase knowledge and awareness of sexual and reproductive health, which enable females to determine their actions, are an obvious place for policy intervention.To the extent that choice is being exercised, higher levels of premarital sexual activity represent a breaking down of social norms. If moral attitudes towards sex have changed, and young women and men are having sex out of choice, this trend could have long-term implications for a society like India. It would sug-gest the formation of new gender relations, in which the ability to make decisions is possessed by both. On the other hand, given the class and gender hierarchies that are present in India, the ability to make this decision is likely to be tempered by the position anindividualoc-cupies in the social strata. In other words, one’s socio-economic class and one’s gen-der is likely to determine how much power one has in making decisions about sex. Until now, population-based, national-level data on unmarried men and women have not been available. TheNFHS-3 in-cluded unmarried women and men in its survey population. This study is aimed at understanding premarital sexual activity in India. A major limitation of the study is that sexual activity per se is limited to actual sexual intercourse, whereas in actu-ality, the term is wider in scope and in-cludes a range of activities, including petting, hugging, kissing, and oral sex. Data and MethodsThe data used for this study are from the NFHS-3 conducted in two phases in India under the aegis of the Demograhics and Health Surveys (DHS) programme from November 2005 to August 2006. The DHS are nationally representative, population-based surveys of women and men of reproductive age. Information is collected on various populations and health char-acteristics, including fertility,family planning, child mortality, maternal and child health, nutrition, and individual and household background characteristics. A final report based on the findings of the NFHS-3was released in 2007 (IIPS and Macro International 2005-06). For the NFHS-3, a total of 124,385 women aged 15 to 49 years and 74,369 men aged 15-54 years from 109,041 households were interviewed. This analysis includes all women and men, unmarried and married. For the first time, information was collected from both married and unmarried men and women in the NFHS-3. There were 25,462 women and 25,382 men among the unmarried re-spondents for whom there was information on age at first inter-course, and 96,000 women and 48,988 men among the married respondents for whom there was information on age at first in-tercourse and age at marriage. From this information, it is possi-ble to gauge whether the respondents had intercourse before marriage. The methods used in the analysis are solely descriptive in nature, and the aim of this section of the study is to describe premarital sex across all the states by various social, demographic and economic characteristics. To gauge the reliability of the data, the rates from the NFHS-3 are compared with that from the in-depth study of youth aged 15 to 24 years mentioned earlier. The Youth: Situations and Needs study is being carried out in six states and data for three are publicly available in the form of fact sheets(Population Council and IIPS 2007). Table 4 (p 59) shows the comparable figures. For each state, the rate for the study is reported, followed by the rates from the NFHS-3 for each demographic group. The percentage is of the rate for the NFHS-3data to that of the youth study. It can be seen that in all cases but one (Maharashtra) the rates from the NFHS-3 data are less than that of the youth study. It can also be seen that for each state, unmarried women were more likely to under-report than married women. This comparison shows that a national-level survey on demography and health as a whole is likely to underestimate levels of non-marital sexuality compared to a survey that targets the sexual and reproductive health of a particular population, in this case, youth.Premarital Sex by GenderThe percentage of women and men who report premarital sexual intercourse are presented in Table 1. The data shows that about 1.8% of all women and 12% of all men report that they had sexual inter-course before marriage. The wide disparity between men and women in the rates of premarital sex indicates that there are dif-ferent rules for men and women with regard to non-marital sexuality in India. When breaking the numbers down by marital status, it is apparent that signi-ficantly more ever married women (2%) report premarital sex compared to never married women (0.7%), whereas slightly more never married men report premarital Table 1: Percentage of Persons Who Report Having Sexual Intercourse Prior to Marriageby Selected Characteristics (NFHS-3, 2005-06) WomenMenMarital status Never married 0.7 13.5 Evermarried2.011.2Age 15-19 0.8 8.8 20-24 1.5 16.0 25-29 1.814.9 30-342.212.5 35-39 2.311.5 40-442.59.5 45-492.210.0 50-54 – 10.4Education None 2.012.4 Primary1.813.0 Secondary or higher 1.5 11.6Media exposure No or infrequent 2.0 11.2 At least once a week 1.6 12.2Occupation Do not work 1.4 6.7 Have an occupation 2.3 12.7Residence Rural 1.812.6 Urban1.710.9Wealth quintile Lowest 2.2 12.2 Second1.813.6 Middle1.712.7 Fourth1.611.9 Highest1.610.1State* HimachalPradesh2.2– Punjab2.717.2 Uttarakhand4.7– Haryana3.316.4 Uttar Pradesh – 14.5 Sikkim – 25.5 ArunachalPradesh4.023.9 Nagaland6.127.7 Mizoram10.245.3 Tripura4.418.1 Jharkhand9.8– Orissa –12.3 MadhyaPradesh–12.7 Gujarat2.615.2 Maharashtra4.312.8 AndhraPradesh–15.6 Goa 4.7– Kerala3.314.6 TamilNadu2.8–Total 1.812.0Sample size 124,385 74,369* Only states with higher than average percentage shown. 
NATIONAL FAMILY HEALTH SURVEY-3Economic & Political Weekly EPW november 29, 200857sex (13.5%) than ever married men (11.2%). Among women, the percentage of those who reported premarital sex increases mar-ginally with age, from 0.8% of those aged 15 to 19 years, to 2.2% among those aged 45 to 49 years. For men, there is no clear rela-tionship with age, with those aged 20 to 24 years (16%) and those aged 25 to 29 (14.9%) reporting higher than average levels.While there appears to be a strong relationship between edu-cation and household wealth and the levels of premarital sex among women, these relationships are less clear for men. The percentage of women who report premarital sexual intercourse decreases with an increase in their level of education as well as with an increase in the wealth quintile their household is in, albeit marginally. But there is no discernible relationship between pre-marital sex and men’s level of education or wealth. Both men and women show substantial differences by occupa-tional status (whether the respondent has an occupation or not), with more of those who have an occupation reporting premarital sexual activity. Both men and women show little variation whether their residence is in a rural area or not.Table 1 also lists the states in which a higher than average per-centage report premarital sexual activity. More women in the north-eastern and south-western states report premarital sexual activity compared to women in other states, with the highest per-centages in Mizoram and Orissa. More men in the north-eastern states of Mizoram, Nagaland, Arunachal Pradesh and Sikkim as well as in Punjab and Haryana report premarital sexual activity.TheNFHS-3 collected data on eight cities, and percentages of premarital sexual acti-vity in each city are presented in Table 2. The results show that very few cities report higher than average premarital sexual activity. For women, these are Mumbai (4.2%) and Chennai (3.4%), while for men they are Meerut (11.9%) and Mumbai (11.7%). In only one of the cities is the level of premarital sexual activity higher than the state average, Chennai. Premarital Sex by Marital StatusTable 3 (p 58) shows the level of premarital sex according to social and economic char-acteristics and marital status. The distri-butions for unmarried women and men show that the level of premarital sexual activity for both groups varies by place of residence, occupational status, level of education, and level of household wealth. Women who live in rural areas are more likely to have had premarital sexual intercourse (0.9%)thanthose who live in urban areas (0.4%), those who have anoccupationhave higher levels of premarital sexual activity (1.2%) than those who do not (0.4%), and the percentage decreases with an increase in the education and wealth quintile. These relationships are also found among unmarried men. More men in rural areas report having had sexual intercourse (14.7%) than those in urban areas (11.9%), those who have an occupation have higher levels of premarital sexual activity (17.1%) than those who do not (7.2%), and the level of premarital intercourse decreases with an increase in the level of education.The distributions for married women show little variation by residence, occupational status and level of education, and while it does not vary for men, there is no clear pattern to any of these variables. Also, while the percentage who report premarital sexual activity generally decreases with an increase in the quintile of household wealth for married women and has an inverted U-shape for married men, again, the variation between levels is marginal. In sum, the results of bivariate analysis show that for unmarried women and men there is a variation in the incidence of premarital sexual activity by level of education, employment and level of wealth of the household, while the same is not true for married women and men. These results suggest that economic status is a significant factor for unmarried women and men but not so for married women and men. At the same time, for three of the five questions on attitudes towards sex roles, unmarried women who give gender equitable responses are less likely to have had premarital sexual intercourse than women who agree that a husband is justified in beating his wife. Again, there is no variation for married women. On the other hand, men who support more equitable gender roles are more likely to report premarital sexual activity. It is notable that for both groups of men, unmarried and married, the pattern of responses to this set of questions is similar.The distribution by marital status is im-portant for women, because while it varies by social and economic characteristics for unmarried women, it does not vary very much by the same characteristics for married women. So, a closer inspection of marital status and sexual activity is neces-sary. The comparison of age at marriage and age at intercourse proves to be reveal-ing. Of the 2% of married women whose age at first intercourse came before age at marriage, 1.6% had intercourse in the year before marriage and only 0.4% had inter-course earlier than that. The latter number is closer to the percentage of unmarried women who report that they have had sexual intercourse. This finding suggests that for a significant number of women who have premarital sexual intercourse, sexual activity leads to marriage. Either women marry the person they have sexual intercourse with (or have sexual inter-course with the person they are affianced to) or they get married subsequently. Thus sexual intercourse transitions into marriage for a portion of women. Class and Gender DimensionsEconomic class and gender are two dimensions of socio-economic inequality in India which interact with each other, and have to be examined separately as well as in relation to each other to Table 2: Percentage of Persons Who Report Having Sexual Intercourse Prior to Marriage by Selected City and State (NFHS-3, 2005-06)State/City WomenMenDelhi 0.5 10.3Delhi city 0.8 10.9Uttar Pradesh 0.8 14.5Meerut 0.311.9West Bengal 0.7 7.8Kolkata 0.63.2Madhya Pradesh 0.3 12.7Indore 0.28.9Maharashtra 4.3 12.8Mumbai 4.211.7Nagpur 2.210.7Andhra Pradesh 0.1 15.6Hyderabad 0.16.2Tamil Nadu 2.8 9.7Chennai 3.49.2Total % within cities sampled 2.0 9.6Total number in city sample 20,508 16,518Total % 1.8 12.0Total size 124,388 74,368
NATIONAL FAMILY HEALTH SURVEY-3 november 29, 2008 EPW Economic & Political Weekly58understand their influence on sexual health. There are large dif-ferences in access to economic resources as well as to information and care in the health system by economic class, including sexual and reproductive healthcare. Gender refers to the cultural values, roles, practices and characteristics that are ascribed to each sex. In each society, there are different expectations of boys and girls, which have an impact on sexual behaviour, sexual responsibilities, and the ability to access information about sex and healthcare (Rao Gupta 2000). Sexuality in India has to be understood in terms of economic class and gender differentials, since each di-mension influences an individual’s behaviour through the power he or she has to acquire information as well as make decisions.The percentages of premarital sexual activity among unmar-ried and married women found in the NFHS-3 raises an interest-ing question: how much of the pattern in the relationships between premarital sex and age, education, and wealth status were caused by these factors, and how much by biases in reporting, primarily because of gender and class issues.It is more than likely that the data that has been collected on sexual intercourse before marriage has been seriously under-reported, especially by women. Even though the anonymity of the interview was a required component of the survey procedure, anecdotal evidence from the field indicates that there was resistance to ques-tions about sexual intercourse, albeit in the domestic violence module. In each instance, community leaders and officials had to be placated with assurances that the questions on sexual inter-course were part of a health survey before the survey was allowed to proceed. It is also notable that unmarried women interviewers said they were greatly embarrassed at having to ask questions about sexual intercourse. If these are examples of overt resistance to the open mention of sexual intercourse, it is likely that the re-sponses from the interviewees was also inhibited. All this indicates that the understanding of the gender dimension operates even at the point of data collection, leave alone at the level of reporting.The higher rates of response from married women are sugges-tive of biases in reporting. Of course, married women include women of all ages, and there are more women in the older ages in the married group than in the unmarried group. However, a com-parison of women by marital status and age group suggest that in the younger ages (15 to 24), the rates are more similar between the two groups, though somewhat higher for married women. The higher rates in the older ages could be either due to the “abil-ity” to report on a sensitive topic like premarital sex or due to the higher likelihood of premarital sex among the older cohort and those with a later age at marriage. If there is no error in reporting the two measures, it might also be because it is easier for married women, who are expected to have had sex, to talk about it and admit to premarital sex. It is assumed that class has an influence on an individual’s abil-ity to access resources, including information and care in sexual and reproductive health. A person with more economic resources is more likely to have access to knowledge, and be more able to make healthy decisions about sex than a poor person. At the same time, when it comes to sex and women, gender operates within the context of a complete and enduring silence. As such, the inter-linkages between class and sexual choices are complex in a cul-ture with strong unwritten rules and regulations on womens’ bodies. So, the pathways through which class affects women and the negative consequences of ill-informed decision-making are an important aspect of sexuality in India, which need careful analy-sis and in-depth discussion. However, the overwhelming effect of gender on reporting makes it difficult to do a comprehensive class analysis. At the same time, the incidence of premarital sexamong Table 3: Percentage of Women and Men Who Report Sexual Intercourse Prior to Marriage by Selected Characteristics(NFHS-3, 2005-06) Unmarried Married UnmarriedMarried Women Women Men MenAge 15-24 0.7 1.3 11.9 14.0 25-34 0.9 2.1 20.3 12.0 35+ 1.7 2.4 23.4 10.2Age at marriage (women/men) Below 15-18 years – 0.7 – 6.1 14 to 18 years/19 to 21 years – 1.8 – 11.4 19+ years/22+ years – 3.5 – 13.1Education None 1.6 2.1 18.6 11.3 Primary 1.2 2.0 17.0 11.7 Secondary 0.5 2.0 12.7 11.5 Higher 0.3 2.0 11.5 8.3Occupation Do not work 0.4 1.7 6.2 12.1 Have an occupation 1.2 2.5 17.3 11.1Attitude to sex roles: Beating one’s wife Justified if wife goes out without permission 0.9 2.1 12.8 10.4 No 0.7 2.1 16.5 13.8 Justified if wife neglects the children 0.7 2.3 12.4 10.1 No 0.7 1.9 16.4 14.0 Justified if wife argues with her husband 1.1 2.2 12.3 10.4 No 0.6 2.0 17.3 13.4 Justified if wife refuses sex 1.5 2.0 13.2 10.8 No 0.7 2.1 20.6 15.6 Justified if wife burns the food 1.0 2.3 13.1 10.8 No 0.7 2.0 16.8 14.0Knowledge of sexually transmitted diseases Heard of STDs 0.6 2.0 14.1 11.8 No 1.2 2.1 8.1 8.4 Heard of HIV/AIDS 0.5 2.1 14.1 11.8 No 1.3 2.0 8.6 8.8Exposure to TV No or infrequent exposure 1.1 2.0 14.5 10.6 Almost every day 0.5 2.0 12.5 11.8Residemce Rural 0.9 2.0 14.7 11.6 Urban 0.4 2.2 11.9 10.2Wealth quintile Lowest 1.6 2.2 16.3 10.8 Second 1.2 1.9 16.6 12.2 Middle 0.8 2.0 13.8 12.0 Fourth 0.4 2.0 12.9 11.2 Highest 0.3 2.1 11.0 9.5Total % 0.7 2.0 13.5 11.1N 25,462 96,000 25,382 48,988
