In 1992 a questionnaire was sent to 50-year-olds in two Swedish counties. These self-report data were compared with clinical observations with regard to number of teeth, removable dentures, caries, and periodontitis. Complete information from both data sources was obtained for 1041 persons. The relevant questionnaire item explained 71% of the missing tooth variance. An agreement of 0.91 (Cohen's kappa) was obtained for removable dentures. A question about problems in opening the mouth differentiated clearly with regard to measured mouth opening ability. Toothache and tooth sensitivity were reported with 95% probability when having 22 decayed teeth and with 46% when there were no decayed teeth (58% correctly predicted). Two teeth with pockets > or = 6 mm gave 5% probability and 22 such teeth gave 39% probability of reporting migration of front teeth. The main conclusion from this study is that there is good correspondence between subjective self-reports and clinical findings, especially for those conditions that are relatively easy for the patient to observe, such as the number of teeth and the presence of dentures. Thus questionnaire data can be used for information and screening about some well-defined oral conditions.