PURPOSE: To describe two methodological approaches to surveying adolescents' sexual life which were combined within a national survey. METHODS: The questionnaire was designed during a 5-day workshop. Ten adolescents played the roles of imaginary adolescents involved in different situations related to sexual life. The information obtained through the analysis of these role-plays was used to establish the sequence of the questions, their content, and wording. The questionnaire was computerized so that it could be completed by the adolescents using laptop computers. Its arborescent structure, leading each respondent from one module to another, made possible the adjustment of the questions to each respondent's stage of development and type of sexual experience. RESULTS: A total of 4283 teenagers (2075 girls and 2208 boys, and 4 refusals) 16-20 years were included. Only 7.6% did not fill in the whole questionnaire. By the age of 18 years, more than 50% of boys and girls were sexually active. Half of sexually active teenagers have discussed contraception issue before engaging in their first intercourse, but much fewer (14-35%) have discussed issues related to HIV transmission. During their first sexual intercourse, the majority of the respondents used a condom (girls, 63%; boys, 58%) or condom plus oral contraception (girls, 11%; boys, 17%). CONCLUSION: Youth participation and role-play decrease the conceptual biases often associated with professionals' perceptions and bring tools enabling a better exploration of circumstances and negotiations surrounding the sexual encounter. The computerized questionnaire increases confidentiality, attractiveness, easy data collection, and, above all, adjustment of the questions to the respondent's level of experience. These methods could be more systematically used in surveys targeting sensitive issues related to adolescents' health.