Self-report measures of sexual maturation continue to be used to classify pubertal development even though their reliability remains in question. This study examined the accuracy of self-report measures by early adolescents in two settings. Standardized figured drawings depicting Tanner's sexual maturation scale (SMS) were shown to early adolescents at school (S1) and again in a clinical sitting (S2), and subjects were asked to rate their own pubertal development. Physical examination by a physician at S2 was used to corroborate sexual maturation. Participating in the study were 46 males, age 11-14 years (mean, 12.4, SD, 1.9), and 37 females, age 11-14 years (mean, 12.7, SD, 0.7). Concordance rate between physical examination and self-report of pubic hair development (males) at S1 was 58% (kappa = 0.35, p less than 0.0001) and 78% (kappa = .66, p less than 0.0001) at S2. Concordance rate of self-report of genital development at S1 and S2 was 27% (kappa = -0.06, p less than 0.49) and 44% (kappa = 0.18, p less than 0.04), respectively. Self-report of breast development demonstrated a concordance rate of 59% (kappa = 0.43, p less than 0.0001) at S1 and 72% (kappa = 0.59, p less than 0.0001) at S2. Concordance rate for self-report of pubic hair development in females was 58% at S1 (kappa = 0.42, p less than 0.0001) and 75% at S2 (kappa = 0.64, p less than 0.0001). There was a tendency for subjects to overestimate their development at early stages of maturation and underestimate development at later stages.(ABSTRACT TRUNCATED AT 250 WORDS)