PROJECT
Assessing and Improving the Measurement
of Sexual Behaviors
(continued)

Kenya 2000, 2002. In a study conducted in two districts, more than 6,000 interviews of unmarried adolescents aged 15–21 were collected as part of household-based surveys. Respondents were randomly assigned to three modes of data collection: face-to-face interviews, paper-and-pencil self-administered interviews, and audio computer-assisted self-interviews (ACASI). The purpose was to evaluate whether the increased privacy afforded by the two self-interviewing methods produced differential reporting of sensitive behaviors. Although the two self-administered methods did not produce more accurate reporting for certain behaviors (e.g., premarital sex, sex with a boyfriend, and alcohol use), these modes—in particular ACASI—elicited a higher percentage of positive responses to questions about more stigmatizing behaviors. Three articles based on these data were published.

Malawi 2004. In collaboration with researchers at the University of Pennsylvania, a household-based survey of adolescent girls was conducted in Balaka district in southern Malawi. The sample was experimentally randomized to face-to-face and ACASI interviews, in order to assess whether computerized interviews elicit higher reporting of sensitive sexual behaviors. The fieldwork also included testing for STIs and HIV status. . Results of this study have been presented at the Population Association of America Meeting in 2006 and a paper has been submitted for publication.

Brazil 2004–2005. As part of a larger clinical study of the effectiveness, acceptability, and feasibility of rapid tests for sexually transmitted infections (STIs) among women aged 18–40 in São Paulo, Brazil an experimental comparison of face-to-face versus computerized survey administration was conducted to assess the reporting of sexual and other STI risk behaviors. As with previous research in Malawi, the study design allows for comparisons between STI status and reported sexual activity by interview mode, however, in this case among an older population of women. Results of this study have been presented at the Population Association of America Meeting in 2006 and a paper has been submitted to the journal Sexually Transmitted Diseases.

Further Research 2006–2009. A series of methodological experiments and innovative research activities will be conducted on this subject in the coming years. Household and clinic-based research will be implemented in diverse settings and populations, including Malawi, South Africa, and Uganda. For instance:

• Uganda. In collaboration with the Ugandan Bureau of Statistics (UBOS) and ORC Macro International a sample of young women aged 18–24 will be randomly assigned to three different interview modalities (traditional face-to-face, computerized interview, and a conversational-style interview). Using the standard Demographic and Health Surveys (DHS) questionnaire for reproductive-aged women, self-reporting of sexual behavior and other sensitive topics will be evaluated by interview mode. Biomarkers for STIs will also be collected and used to validate self-reporting of risk behavior.

• South Africa. As a part of a series of studies in preparation for future microbicide clinical trials, an experiment evaluating face-to-face versus computerized interviewing is being conducted in the Population Council’s three Carraguard Phase 3 sites in South Africa to assess participant self-reported compliance to study protocols. The study will also investigate the reporting of sexual and other risk behaviors, collecting various STI biomarkers for validation of reporting.

• Malawi. In 2009, a large household survey of adolescents age 15–24 will be implemented in Malawi in collaboration with Ann Biddlecom at the Guttmacher Institute. A focus of this study is to evaluate the feasibility and acceptability of using handheld computers in a household-based survey. As with other studies, respondents will be randomly assigned to different interview methods, in this case, computer-assisted personal interview (CAPI) and ACASI. Prior to the survey, a series of cognitive interviews will be conducted to acquire an understanding of how respondents interpret, process, and form answers to questions regarding sexual behavior.

---