Objective: To measure the validity of self-reported dental visits from a diverse sample of adults. Methods: The Florida Dental Care Study was a longitudinal cohort study of a diverse sample of residents of north Florida, USA. In-person interviews and dental examinations were conducted at baseline, 24 and 48 months after baseline, with half-yearly telephone interviews in between. Dental record information was abstracted afterward. Results: Agreement between self-report and dental record at each half-yearly interview ranged from 84 to 91%. Validity did not differ between persons of key sociodemographic groups (sex, race, age group, rural/ urban residence, poverty status, level of formal education, or problem-oriented/ regular approach to dental care). In a single bivariate multiple logistic regression (two outcomes: (i) self-reported use; and (ii) use measured from the dental chart), odds ratio estimates over-lapped for each of the 20 predictors. Conclusions: Validity of self-reported dental care use was good. There would have been few differences in conclusions made about predictors of dental care use had chart data been available earlier.