The validity of injecting drug users' self-reports about sexually transmitted diseases: a comparison of survey and serological data.

Researchers studying the validity of injecting drug users' self-reports have often reported a high degree of correspondence between self-reports and several types of criterion variables. In this study, injecting drug users' responses to questions about prior infection with four sexually transmitted diseases were compared with serological test data. For three of the four diseases studied, discrepancies between survey and serological results were quite large, especially when respondents said they had no previous history of infection. Limitations of serological testing, the questions asked, respondent knowledge and the more traditional concern with respondent veracity are discussed as possible explanations for the observed differences. Study data suggest that use of drug injectors' self-reports to measure infection with sexually transmitted diseases or to validate other measures is a questionable practice. Problems encountered in interpreting study results provide support for more frequent use of experimental or quasi-experimental designs and of multivariate analytic techniques when conducting research on response validity.