This is a study, conducted among 4283 teenagers aged 16-20 years, describing two methodological approaches in surveying adolescents' sexual life, which were combined within a national survey. A computerized questionnaire was used. The sequence, contents, and wordings of the questions were based on the analysis of role-playing by 10 adolescents in a 5-day workshop. The tree-like structure of the questionnaire led each respondent from one module to another as questions were adjusted according to respondent's stage of development and type of sexual experiences. 2075 of the respondents were girls; 2208 were boys. 7.6% did not finish the questionnaire due to its length. It was found that by age 18 more than 50% of the sample were sexually active. Half of this segment discussed contraception before engaging in their first act of sexual intercourse, and a majority of them used condoms. However, only 14-35% discussed issues concerning HIV transmission. The participation of young people in the design of the questionnaire produced an enormous amount of usable information and increased the acceptability of the survey to the respondents. Moreover, it reduced the influence of professionals' perceptions, which tends to bias the concept of the survey as well as the sequence and wording of the questions. Role-playing produced more creative and spontaneous materials for the survey. This kind questionnaire is advantageous as a general model. Firstly, it strengthens the feeling of confidentiality; secondly, the focus group and role-playing approach in association with computerization favors a qualitative, flexible, and contextual approach to adolescents' sexuality.