Abstract Background: We examined the validity of self-reported number of teeth in middle-aged adults by using representative cohort data to compare corresponding self-reported and clinical values. Methods: This validity study is part of the representative 46-year-old follow-up of the Northern Finland Birth Cohort 1966 (NFBC1966) Study. Mailed questionnaires (n= 5950) requested information on self-reported number of teeth and background variables (education, tooth brushing and smoking), while clinical oral health examinations (n= 1891) assessed the number of teeth (the ‘gold standard’). The main analyses compared the self-reported and clinical values for the number of teeth in 1669 participants. Scatterplot and Bland-Altman plot served for visual analyses, and alternative correlation coefficients (Pearson, Spearman, intraclass) for numerical comparisons separately for men and women, with stratification according to background variables. Results: The clinical assessment revealed that the mean value for the number of teeth was 27.46 (SD = 2.38), while the corresponding value based on self-reported information was 27.48 (SD = 2.78). According to the Bland-Altman plot, the mean difference between the clinical and self-reported values was − 0.02 (95% limits of agreement, LoA: − 3.37 to 3.32). The observed ranges of intraclass correlation coefficients (ICC) among men and women were 0.72 to 0.95 and 0.72 to 0.85, respectively, depending on the background variables. Conclusions: Self-reported number of teeth in middle-aged Finnish adults agreed closely with the corresponding clinical measure.