Background: Although adult reported childhood socioeconomic position has been related to health outcomes in many studies, little is known about the validity of such distantly recalled information. This study evaluated the validity of adults' reports of childhood paternal social class.
Methods: Data are drawn from the Aberdeen children of the 1950s study, a cohort of 12 150 people born in Aberdeen (Scotland) who took part in a school based survey in 1962. In this survey, two indices of early life socioeconomic position were collected: occupational social class at birth (abstracted from maternity records) and occupational social class in childhood (reported during the 1962 survey by the study participants). Between 2000 and 2003, a questionnaire was mailed to traced middle aged cohort members in which inquiries were made about their fathers' occupation when they were aged 12 years. The level of agreement between these reports and prospectively collected data on occupational social class was assessed.
Results: In total, 7183 (63.7%) persons responded to the mid-life questionnaire. Agreement was moderate between social class of father recalled in adulthood and that measured in early life (kappa statistics were 0.47 for social class measured at birth, and 0.56 for social class reported by the child). The relation of occupational social class to birth weight and childhood intelligence was in the expected directions, although weaker for adults' reports in comparison with prospectively gathered data.
Conclusions: In studies of adult disease aetiology, associations between childhood social class based on adult recall of parental occupation and health outcomes are likely to underestimate real effects.