Problems in using survey data include definitional problems, operationalization of concepts, recall bias, underreporting, question order, external validity, and the sex and ethnicity of interviewers. The article examines recent developments in survey administration, such as the use of audio computer-assisted self-interviews, discusses the need for careful assessment of interview techniques, and recommends improvements in methodology. It also emphasizes the importance of research ethics in the study of violence against women, not only for the potential emotional trauma to respondents, but also for the potential for actual revictimization. The article discusses the effect of varied definitions and the impact of context on prevalence estimates, and supports the idea that attention must be paid to protecting the confidentiality and safety of study participants. Notes, references