This article presents qualitative findings from a project designed to develop better methodological tools for clarifying women's and men's attitudes about intimate partner violence (IPV) in rural Bangladesh and their perceptions of norms about IPV in their communities. Cognitive interviews and focus-group discussions were used to explore respondents'subjective understanding of standard survey questions meant to elicit attitudes about IPV. We find that the proportion of participants who justified IPV under some circumstances increased when additional context was provided, suggesting that the Demographic and Health Survey (DHS) questions may underrepresent the proportions of people who condone IPV. Moreover, most people in this social context also believed that IPV often goes beyond socially sanctioned limits and that something should be done to stop it--an important caveat that the DHS questions are not designed to capture. These findings suggest that incorporation of additional context into questions on IPV attitudes would make responses to these questions more meaningful.