In this study, the author analyzes the websites of 32 universities' Institutional Review Boards (IRBs) to explore two specific issues: (a) how individual IRBs define "vulnerable populations," and (b) the guidelines those IRBs offer with respect to participants who may have limited or no English language proficiency. Analysis indicates wide variation in IRB guidelines. Despite a trend to develop separate IRBs for biomedical and social and behavioral research, social science IRBs still use language reflective of medical models for research. Second, there is great variation in the definition of "vulnerable populations," as well as in the requirements for including, or excluding, their participation in research. Finally, university IRBs offer little, if any, guidance about conducting research with participants whose first language is not English. The author offers specific recommendations for the ways in which qualitative researchers may advocate for changes in how IRBs conceptualize these types of research studies and participants. (Contains 3 tables.)