The Retrospective Bias in Unemployment Reporting by Sex, Race and Age

Abstract This article examines differences in reported unemployment between a survey taken the week following the individuals' labor force experience and one conducted a year later on a retrospective basis. Using a limited number of annual observations, we find a relative understatement of unemployment in the retrospective survey for groups containing large proportions of women and youth. However, for certain other groups, namely central age males and older women, the retrospective method appears to overstate unemployment relative to the current reporting. To explain these findings we offer several hypotheses relating to the social roles of primary and secondary workers and the fact that the labor force member and the survey respondent are frequently not the same individual.