In longitudinal research projects, respondents’ contact information such as address, telephone, and email can become out-of-date. Some researchers have searched Facebook for these unreachable respondents, with the hope of sending them a private message that asks them to provide the needed follow-up data. Published reports of research that used Facebook that way raise several questions: When should researchers conclude that a Facebook user is actually the respondent they are seeking? How should researchers word their message? To what extent is respondent tracing through Facebook likely to reduce nonresponse bias? Is communication using Facebook sufficiently confidential? This paper discusses these questions and presents Westat’s own use of Facebook to trace respondents in a longitudinal evaluation of an intervention for youth in foster care. The paper discusses the implications of ourwork for future research practice.