We present the results of a randomized experiment to test the effectiveness and cost-effectiveness of response incentives for a stated-preference survey of a general community population. The survey was administered using a mixed-mode approach, in which community members were invited to participate using a traditional mailed letter using contact information for a representative sample of the community; but individuals completed the survey via the web, which exploited the advantages of electronic capture. Individuals were randomized to four incentive groups: (a) no incentive, (b) prepaid cash incentive ($2), (c) a low lottery (10 prizes of $25) and (d) a high lottery (2 prizes of $250). Letters of invitation were mailed to 3,000 individuals. In total, 405 individuals (14. 4%) contacted the website and 277 (9. 8%) provided complete responses. The prepaid cash incentive generated the highest contact and response rates (23. 3 and 17 VSports手机版. 3%, respectively), and no incentive generated the lowest (9. 1 and 5. 7%, respectively). The high lottery, however, was the most cost-effective incentive for obtaining completed surveys: compared with no incentive, the incremental cost-effectiveness ratio (ICER) per completed survey for high lottery was $13. 89; for prepaid cash, the ICER was $18. 29. This finding suggests that the preferred response incentive for community-based, stated-preference surveys is a lottery with a small number of large prizes. .