Good girls and boys: findings from a cross-sectional survey on adolescent rights, relationships, and sexuality in an urban informal settlement in India

Abstract Around 20% of India’s population are adolescents aged 10–19 years. Our objective was to strengthen program interventions on gender equity, health, and participation by gauging adolescents’ levels of understanding and opinions. In a cross-sectional survey, we interviewed 2005 adolescents on their opinions on rights, friendship and sex, sexual refusal and coercion, and communication with family, using a two-stage probability proportional to size sample. Opinions on gender allocations were generally equitable, although females supported clothing proscriptions. Premarital sex, multiple partners, masturbation and non-heterosexual partnerships were frowned upon. Few respondents said that they felt pressure to be sexually active, 79% said that sexual coercion was a form of violence, but 14% of older adolescents said that it would be unreasonable to refuse sex. Our interviews described young people negotiating the terrain between perceived normative expectations and contemporary aspirations, showing limited manoeuvring within assumed gender roles in which family control was prominent.

. Evaluation suggests that its success has been limited and that it has not succeeded in either increasing services for adolescents or improving the quality of the services they receive (Population Council & UniCEF, 2013). The recent National Programme for Youth and Adolescent Development (NPYAD) merges four schemes under the Ministry of Youth Affairs and Sports: Promotion of Youth Activities and Training, Promotion of National Integration, Promotion of Adventure and Development, and Empowerment of Adolescents (Government of India, 2015).
None of these initiatives has taken a particularly nuanced approach to gender and sexuality. Program content and approach tend not to address the structural and power inequities that women and girls face, and omit the development of skills to challenge socially accepted roles and expectations around sexuality, fertility, and work (Nanda, Das, Singh, & Negi, 2013). Sexuality is not prominent on their agendas and they have tended to reinforce hetero-normative behaviour and livelihoods, with the assumption of marriage as a young woman's ultimate goal and little scope for informed choices about sexuality and personal space. Programs that have included young men have been fewer and have likewise posed few challenges to sexual and gender norms. Studies have shown that the parameters of hetero-normativity and social sanction for boys to engage in sexual behaviours are often related to conceptions of masculinity that manifest in authority and power (Abraham, 2001;Santhya, Haberland, Ram, Sinha, & Mohanty, 2007). The subtle aspects of their sexuality are often not explored. A study of constructs of masculinity and their influence on men's wellbeing suggested that dominant norms and traditional beliefs about manhood were associated with greater risk-taking and risky sexual activity (Pradhan & Ram, 2010).
Despite restrictive social norms, there is increasing evidence that young people in India engage in premarital romantic and sexual partnerships (Jaya & HIndin, 2009). There are opportunities for social mixing, and young men and women have devised ways of developing romantic relationships, notwithstanding a social environment that ostensibly disapproves of interaction (Alexander, Garda, Kanade, Jejeebhoy, & Ganatra, 2007). TARSHI (Talking about Reproductive and Sexual Health Issues), a non-government organization running a helpline for sexual information, received over 59,000 calls from men seeking information on sexual anatomy and physiology in 2008. 33% of callers were young people (Tripathi & Sekher, 2013).
Home is a key place for sex education and parents' attitudes are vital. Sex education in schools remains controversial. In 2007, the media reported that the Ministry of Health was considering a ban on sex education in schools since it conflicted with Indian cultural values and might lead adolescents to experimentation and irresponsible behaviour (McManus & Dhar, 2008;Tripathi & Sekher, 2013). As recently as 2014, newspapers carried the story that the government of Maharashtra had yet to implement proposals for sex education and that public schools had no structure within which to deliver it (Porecha, 2014). People tend to think of sex education as confined to information on anatomical and biological differences, rather than gender and sexuality. A WHO report (2003) suggested that promotion of family life and sex education had resulted in delay in age of entry into sexual relationships, reduced partner numbers, and increased safe sex and contraception. Sex education in schools did not seem to encourage younger people to have sex. Nevertheless, both proponents and critics of sex education in Indian schools use the language of 'sexual restraint' , delaying the initiation of sexual activity until marriage (Tripathi & Sekher, 2013).
Information on how young people might begin and nurture relationships is scarce, though vital to address their needs. In the absence of other sources, young people rely on films and same-sex peers for information (Jaya & HIndin, 2009;Nair, Leena, George, Thankachi, & Russell, 2013b). As a result, they are likely to be poorly informed or misinformed. Despite evidence that appropriate information delays sexual initiation, prevents unwanted pregnancies, and lowers rates of STI transmission, social norms discourage discussion of issues related to sexuality in family settings, as parents often believe that this would imply approval. The cycle of poor communication keeps young people ill-informed and unlikely to receive parental support in sexual matters (Jaya & HIndin, 2009).
Research with adolescents has focused on access to health services, their knowledge of reproductive and sexual health, premarital partnerships, and risky sexual behaviours. Sexual health and wellbeing can be enhanced when adolescents are encouraged to express their sexuality candidly, and if they have been provided guidance on dealing with the issues positively. There is a need to create a social environment that allows expression of sexuality as a natural and important part of growing up. Such an environment will put adolescents in a better position to negotiate control and power over their sexuality, and will foster intolerance of violence.
A non-government organisation based in Mumbai, Society for Nutrition, Education and Health Action (SNEHA) manages the Empowerment, Health, and Sexuality for Adolescents (EHSAS) initiative with adolescents aged 11-19 years. The initiative includes individual and group activities and classroom sessions aimed at education on self-awareness, sexuality and its expression, soft and social skills, employability, and creative expression of sexuality through art and recreational activities. Group work with parents is conducted to create an enabling environment in which their maturing children can express their sexuality and build acceptance of it. Prior to the project, we did a needs-assessment to understand adolescents' knowledge and practices around sexuality, gender and violence. Our objective was to provide background information for our work in project EHSAS. We were beginning the intervention program on adolescent health and sexuality and wanted to gauge levels of understanding and opinions in order to develop learning materials and intervention strategies.

Setting
The study was done in Dharavi, an urban informal settlement in Mumbai with an estimated population of 750,000. Informal settlements (slums) are features of urbanization in India and have been described in two-thirds of cities and towns. The most recent estimate is that 41% of Mumbai's households are in such settlements (Chandramouli, 2011). UN-HABITAT characterizes them in terms of overcrowding, insubstantial housing, insufficient water and sanitation, lack of tenure, and hazardous location (Ministry of Housing and Urban Poverty Alleviation, 2010; United Nations Human Settlements Programme (UN-Habitat), 2003).
Dharavi covers 557 acres and is divided into 96 geographical clusters (Lewis, 2011). Our work with adolescents began on the backdrop of an existing program on prevention of violence against women and children, which itself began in 2000. Secondary interventions are provided through five counselling centres across Mumbai, linked with community mobilisation, health service, police, and legal support. Primary prevention is carried out through community mobilisation, mainly through group work and voluntarism. We currently run 180 groups of women, adolescents, and men. Key actions of these groups include collective support by community members and referral of survivors to support services.
Around 50 adolescent groups have been campaigning in communities since 2002. They carry out safety audits, mount performances and campaigns, undertake peer counselling, and participate in gender transformation programmes. More than a decade of work in the community means that our organisation is relatively visible in terms of our intent and credibility.

Design
We did a cross-sectional quantitative survey of adolescents' awareness of and opinions on rights, friendship and sex, sexual refusal and coercion, communication with parents, and general communication within families.

Participants
The inclusion criteria were that respondents be aged 11-19 years, irrespective of socioeconomic status, caste, religion, gender, disability, or education. Unfortunately, we were unable to interview adolescents with cognitive, hearing, or speech impairments because the interview team did not have the necessary skills to do so.

Tools
Questionnaire content was developed after focus group discussions with parents and adolescents divided into groups aged 11-14 and 15-19 years. The discussions addressed perceptions and practices around sexuality and sexual behavior, and the influence of gender, culture, and economic status on sexuality. Four areas of enquiry emerged: current awareness of rights, negotiation of control in relationships, levels of understanding of sexuality and opinions on it, and communication between adolescents and their families. The resulting questionnaire included six groups of predominantly closed categorical responses: family background, knowledge of rights, knowledge of and attitude to gender and violence, response to the participant's own and others' feelings of sexuality, and family influence and control over decision-making.

Procedures
We recruited 38 field interviewers after written tests and two rounds of interviews. They were evaluated on their writing and communication skills, non-verbal communication, and understanding of relevant issues, and given three days of training on issues around sexuality, sex and gender, counseling and initial response, and technical aspects of administering the questionnaire. We piloted 100 interviews in the presence of a supervisor who intervened if the interviewer appeared to be struggling. At the end of each day, the supervisor went through the questionnaires, identified problems in completion, and checked if the investigator had any doubts or concerns.
The subsequent survey was done by two teams, one for males and one for females, each allotted a supervisor who observed interviews throughout data collection and checked questionnaires for data quality. The team of interviewers began at a central point in each cluster and worked in pairs, visiting every 50 home on the left-hand side of a lane and visiting adjacent homes if no adolescent lived there. If this process did not yield sufficient potential interviewees, they made enquiries about adolescents in the lane. Interviews took between 45 min and one hour, and interviewers arranged to come back if they were incomplete. Given the sensitive nature of the interview, we assumed the possibility of social desirability bias. We arranged that interviewers be unknown to respondents. The questionnaire was administered in respondents' homes after signed consent from both them and their parents. We tried to conduct interviews without other people present. If a location such as a loft or verandah was not available, we booked a small hall in the vicinity of the cluster to afford respondents privacy. Interviewers were trained to avoid cues in body language or tone of voice that might influence respondents' answers. Interviews were reviewed and compared by supervisors and field research coordinator at the end of each day, and flagged for interviewer review if there were discrepancies. The two groups met every two weeks to compare data collection and recording procedures.

Sample size and selection
A sample size of 1000 in each of two groups (older compared with younger adolescents, or females compared with males) would have at least 90% power to detect a difference between groups of 8% in binary outcome proportions across the possible range of comparisons. We aimed, therefore, to achieve a two-stage sample of 2000 after interviewing 20-25 adolescents in each of 80 primary sampling units selected from 96 geographical clusters. The self-weighting sample of clusters was developed through Probability Proportional to Estimated Size (PPES), based on a population estimate of 750,000, an estimate that 20% of the population would be adolescent (the estimate from the third National Family Health Survey was 19% for households in slum areas of Maharashtra state; IIPS and Macro International, 2008), a sampling interval of 2981 and the first primary sampling unit defined by a random number between one and 2981.

Data management and analysis
Survey questionnaires and information from interviews were kept in locked cupboards. Anonymized electronic data were entered into a database in Microsoft Access (Microsoft Corporation) and stored on a server in intelligibly named files. Information was backed up weekly on an external hard drive. Only the program director and senior members of the research team were allowed access, passwords were changed fortnightly, and data were read-only where possible. No datasets included the names of respondents. We tabulated frequencies and proportions of outcome variables, by sex and age (11-14 years and 15-19 years) and for the whole sample. We compared response proportions and 95% confidence intervals for differences between older and younger age groups and between females and males using survey commands in Stata 13 (College Station, TX), followed by lincom commands (linear combinations of estimators). We summarised socioeconomic position by quintiles of asset indices derived from standardized weights for the first component of a principal components analysis of household durables (Filmer & Pritchett, 2001;Vyas & Kumaranayake, 2006).

Ethical concerns
Approval for the study was granted by the Multi-institutional Ethics Committee, Mumbai. The issue of sexuality required careful handling. The process of recalling distressing events might have disturbed interviewees, and there were family, social, and legal issues around the discussion. We were reassured when parents' involvement in formative focus group discussions was enthusiastic, but we piloted the questionnaire extensively to gauge potential response. We had already worked with adolescents in the area and a support network of counselors, community peer workers, and healthcare professionals was in place for respondents to call upon. We trained the interviewers on the issues around confidentiality and ensured confidential data management. We discussed the study with community leaders before the focus group discussions, and subsequently organized meetings with community representatives in each cluster involved in the survey.
Informed consent was obtained from adolescents and their parents. Potential respondents were allowed a day to decide, along with their parents, whether they would like to be involved. Interviewers made sure that they had an opportunity to ask questions and understood the study. Respondents received no financial compensation. They were given information and booklets about HIV and child sexual abuse, and were provided with our counseling centre helpline numbers. If the respondent reported abuse, subsequent visits were made to the home to arrange a meeting with a counselor and appropriate follow-up.

What do adolescents think of their rights?
We asked adolescents about their awareness of their rights to education, protection, participation, free speech, and play. If they had heard of a right, we asked them whether it applied more to boys or girls (Table 2). Although awareness -or understanding of the idea of rights -was higher in the older age group, greater proportions of girls than of boys said that they were aware of their rights at all ages. Levels of awareness were, nevertheless, not high. Only 63% of older girls were familiar with the right to education and 59% with the right to free speech. Of those who said that they had heard of individual rights, very few said that they applied differentially across the sexes. In situations of household food insecurity, most girls (90%) said that sons and daughters should have equal access to what there was to eat. Where there were differences in opinion between the sexes, boys were more likely than girls to suggest that daughters should have priority access to food (95% confidence interval for difference 15-23%). This pattern was not repeated when adolescents were asked about who should share in limited household cash. Again, girls were more likely to say that money should be shared equally between brothers and sisters, but boys who opted for one or the other were roughly equally divided between brothers and sisters as priority recipients. A greater proportion of girls said that they should be restricted in their choice of clothing (36%; 95% CI 24-47%).

Sexual refusal and coercion
Pestering someone for a date was generally not considered acceptable, although 13% of older males thought it was (Table 4) and a greater proportion of males thought so at all ages (difference 4-10%). Few adolescents said that they felt pressure to be sexually active, but 6% of older boys and 2% of older girls said that they or someone they knew had been coerced at some point. Nonetheless, only 78% of older adolescents felt that it was reasonable to refuse to have sex 79% saying that coercion would be a form of violence. Teasing a person of a different sex was unacceptable (97%). A greater proportion of girls (85%) than boys (61%) thought that when a girl said no to sex she meant it (difference 19-30%). Around two-fifths of boys felt that this was not the case or that it depended on the circumstances.

Communication with parents
Most adolescents said that their parents allowed them to choose their own friends, although males were more likely to say this (89%) than females (75%; difference 10-19%. Table 5). Around two-thirds were allowed to go out with their friends, but friends of another sex were less permissible (42% of males and 17% of females; difference 20-29%)) and meeting them without others present less so (11%). More than 85% of respondents said that they had discussed their studies with their parents, but more females (38%) than males (15%) said that they found this difficult (difference 15-31%). Most (79%) said that they had discussed their career, irrespective of gender, although, again, females (27%) tended to find this difficult (difference 2-11%). Females were more likely (12%) than males (7%) to say that they had discussed marriage (difference 1-8%). Males particularly said that they found it difficult to discuss their sexuality (88%).
Adolescents said that the best form of marriage was arranged (67%) and that their parents thought so too (91%). Nineteen percent of males said that their parents would allow them to marry a partner of their choice, but this opinion was shared by only 5% of females (difference 11-18%). Adolescents' expectations of their future life partners differed in some respects (Figure 1). Education, pleasant nature, and an understanding character were important to both sexes. Females were more likely to prioritise family background, men who made them happy, non-drinkers and non-users of drugs, work in business, and wealth. Males were more likely to prioritise looks, smart appearance, fair complexion, and religion. That men should not be drinkers or drug-users was much more important to women. Most of these requirements were echoed in respondents' ideas about what their parents would prefer. Family background was important, and females gave more priority to residence in the same city.

Norms
If we take their answers at face value, young women articulated a correspondence between injunctive norms, normative expectations, and empirical norms. They stayed away from young men, knew little about sex and didn't masturbate, disapproved of premarital relationships and multiple partners, accepted gendered dress codes (also found in the ARSH study) (Nair, Leena, George, Thankachi, & Russell, 2013a), did not discuss their future career much with their parents, and expected an arranged marriage with a man approved of by their parents, perhaps from the same city so that they could see their parents easily. This man would be from a compatible family background, educated and employed, and (hopefully) not a drinker. Their views represented, perhaps, limited manoeuvring on a background of an assumed gender role in which family control was prominent (Fatusi & Hindin, 2010;Jejeebhoy, 1998). Young men were more likely to approve of friendships with women, say that they knew something about sex (a trait described elsewhere) (Fatusi & Hindin, 2010;WHO, 2011), be amenable to asking women out, be likely to be in a relationship, and be less convinced that no meant no. They were, however, uncomfortable with the idea of masturbation and also disapproved of premarital relationships and multiple partners, as has been found in other studies (Guilamo-Ramos et al., 2012). The Hindi word for masturbation, swapnadosh, means 'fault of dreams' , and a sense of shame around it may reflect a culture in which sexuality and pleasure remain largely unexplored (Nair et al., 2013a).
These were not, to put it mildly, rebellious teenagers. Braggadocio was not the hallmark of their answers to our questions about their relationships and sexuality. What interested us was their apparent naivety and wholesomeness, manifest in a romanticism that characterized descriptions of sex.
were hetero-normative and conformed with societal norms, supporting abstinence until marriage and disapproving of masturbation, if they knew about it. There was some cognitive dissonance: coercion was frowned upon and largely unreported by young women, while young men said that it was unreasonable to refuse sex.
The young people we interviewed all lived in an informal settlement. We do not need to expand here on the difference between their lived reality -its emphasis on education and bourgeois and family aspirations -and the salaciously orientalised views of informal settlements as nexuses of transgression (criminal and sexual) that have characterized societies at least since Victorian London and New York. It is, however, interesting to speculate that precisely because of their families' lack of wherewithal these young men and women might have had more traditional views and aspirations than their counterparts among the wealthy.

Concerns
We have some concerns about our data-set that raise a number of provocative questions. First, asking young people about sex when they have been raised in an environment in which empirical norms are overlooked and injunctive norms constantly reiterated may lead to best behaviour bias. Young people's responses may have been processed at three levels: what they really believed, what they thought their parents believed, and what they thought that wider society expected.
Although the interviews were conducted by people of the same sex (McCombie & Anarfi, 2002), face-to-face interviews present their own set of challenges, even if culturally appropriate (Jejeebhoy, 1998). Interviewers were from the same community as respondents, and were trained and supervised carefully, but it is possible that their own attitudes could have cued certain responses. We note, however, that the urge to give politically correct answers did not extend to young people saying that they were comfortable with the idea of non-heterosexual relationships. This raises the possibility that cultural norms might have overridden contemporary ones. We were at first critical of our findings, but over months of re-examination and the implementation of our intervention to engage with adolescents, we now wonder whether their answers were honest and their relative lack of sexual knowledge and activity genuine.
When we designed the study, we were aware that inclusion of adolescents with disabilities was important. We trained the interviewers to identify and ask questions to adolescents with locomotor, visual, hearing, cognitive, and learning difficulties. Unfortunately, time pressure and lack of confidence on the part of interviewers meant that these groups were not represented adequately. Given our interest in inclusion, this was a lesson. We have prioritised inclusion in our subsequent work. Interviewers will follow protocols to identify disability, we will check the numbers regularly, and we will employ experts in research with people with disabilities to help with the interviews.

Final thoughts
Probably our biggest lesson from the study and our subsequent work is that research involves allaying the fears of a number of stakeholders beyond the participants. In developing the questionnaire, we were careful to respond to the concerns of colleagues and institutional review boards that direct questions about sexuality might be inappropriate. Adolescent gender and sexuality are relatively new areas for work in India, added to which are (we think, largely unfounded) concerns about the sensitivities of slum-dwelling communities. Parents whom we consulted in developing the study and before and during our program were generally enthusiastic about them, and it may be that the obliquity of some of the questions -for example, asking adolescents if they knew others who were in sexual relationships -would have been best avoided. It is likely that over a decade of engagement with the community has contributed to uptake of program activities, but it has been our experience that respondents welcome the opportunity to discuss the issues and are keen to be involved in subsequent activities. We have found that questions about adolescents' own experiences, asked in a supportive atmosphere, often yield direct answers. We are reasonably convinced that they are served better by respectful directness, and are acting on this in our current work.