Using STI Biomarkers to Validate the Reporting of Sexual Behavior within an Experimental Evaluation of Interviewing Methods

Paul C Hewett, Population Council
Manoel Ribeiro, Faculdade de Ciências Médicas de Santa Casa de São Paulo
Sheri A Lippman, University of California, Berkeley
Heidi Jones, Population Council
Janneke van de Wijgert, International Antiviral Therapy Evaluation Center

The AIDS pandemic has increased the need for comprehensive data on risk behaviors in developing countries. However, if reporting is unreliable, the ability to evaluate interventions and monitor programs is compromised. To address the sensitive nature of questions in studies of sexual behavior and STI risk, researchers have turned to computerized interviewing techniques. The collection of STI biomarkers in conjunction with an experimental evaluation of face-to-face versus audio-computer assisted self-interviewing (ACASI) provides an opportunity to externally validate the reporting of risk behaviors. Data will be used from a study assessing STI screening and diagnostic techniques among women aged 18-40 in São Paulo, Brazil. Reduced form bivariate probit models that control for exogenous demographic indicators will capture the residual association between STIs and reported risk behavior by interview method. If higher associations are found with ACASI, strong evidence that computerized interviewing provides more accurate data on risk behaviors will be obtained.