Assessing social risks has proven difficult for IRBs. We undertook a novel effort to empirically investigate social risks before an HIV prevention trial among drug users in Thailand and China. The assessment investigated whether law, policies and enforcement strategies would place research subjects at significantly elevated risk of arrest, incarceration, physical harm, breach of confidentiality, or loss of access to health care relative to drug users not participating in the research. The study validated the investigator's concern that drug users were subject to serious social risks in the site localities, but also suggested that participation in research posed little or no marginal increase in risk and might even have a protective effect. Our experience shows that it is feasible to inform IRB deliberations with actual data on social risks, but also raises the question of whether and when such research is an appropriate use of scare research resources.