Reliability and validity of retrospective self-reports of the age of pubertal onset using twin, sibling, and college student data

This study presents a methodology, derived from behavioral genetic research, designed to evaluate the validity of retrospective self-report inventories. The application of this technique to college student, twin, and sibling data demonstrated the reliability and validity of a retrospective self-report questionnaire on the timing of pubertal onset. Support was provided by four main findings: (1) substantial test-retest reliability coefficients for individual items (r ranged from .73 to .97) were obtained for a sample of college students; (2) nearly all of the monozygotic twin (MZ) intraclass correlations exceeded those for dizygotic twins (DZ); (3) the magnitude of the test-retest and intraclass correlations for recall of the timing of pubertal onset followed theoretical expectations for a trait demonstrating significant genetic influence (i.e., r for the same individual tested twice greater than r for MZs greater than r for DZs greater than r for siblings); (4) the expected secular trend of a cross-generational decrease in the age of sexual maturation was demonstrated for retrospective self-reports of family members representing three different generations. Moreover, the size of the mean differences across generations mirrored those reported for nonretrospective assessments of the actual age of pubertal onset. Finally, the value of retrospective self-report surveys in psychological research is discussed, and the relative merits of these instruments as indices of individual differences in pubertal development, as opposed to indices of the actual age of sexual maturation, are emphasized.