This article reviews literature concerning the accuracy of self-reports of homeless respondents. It reports the findings of a study that specifically addressed the issue of the accuracy of homeless adults' reporting of ambulatory medical visits as compared with studies of the general public. The study used a community-based sample population from a UCLA health study and compared answers given by 349 respondents in face-to-face interviews with information in visit logs of a health clinic serving this population. Respondents were asked to recall the number of visits made in the past 12 months and in the past three months during a first interview, and then those visits made during the four months following the first interview during a follow-up interview. Results indicated that while 10% of the respondents reported using this clinic, medical records at the clinic indicated that 21% had made actual visits. Homeless respondents were found to be accurate when reporting the presence or absence of ambulatory care visits, but not when reporting the number of visits made during a specific time frame. Findings of underreporting by respondents based on the time period they were asked to recall visits over were also discussed. The authors conclude that homeless persons are fairly accurate reporters though they may not report complex information or socially undesirable information. (Authors)