Online virtual worlds are 3D graphical environments in which users create avatars that live, work, and play-some are actual games; others are social networking sites. These worlds offer significant potential for expanding the study of social and economic behavior. One of these worlds, Second Life (SL), began in 2003 and now has more than millions of unique accounts held by virtual residents and approximately 800,000 users who log in each month. Although few results from SL surveys have been published, automated surveys are common in the virtual world. They are conducted through links to web surveys, by in-world 'survey bots' (programs that administer questionnaires), and through e-mail invitation to panel members recruited through SL. This article evaluates the effectiveness of recruitment techniques to complete a self-administered survey within SL. The authors tested several techniques: a billboard in an SL public space, classified ads in SL publications, paid search engine advertisements, and SL web forum posts. All survey recruits completed the same survey administered at the RTI SL facility kiosk. Our findings suggest that a diverse convenience sample can be obtained in SL at a minimal cost.
