Biased estimators? Comparing status and statistical theories of gender discrimination

Abstract

Gender inequality in paid work persists, in the form of a gender wage gap, occupational sex segregation and a “glass ceiling” for women, despite substantial institutional change in recent decades. Two classes of explanations that have been offered as partial explanations of persistent gender inequality include economic theories of statistical discrimination and social psychological theories of status-based discrimination. Despite the fact that the two theories offer explanations for the same phenomena, little effort has been made to compare them, and practitioners of one theory are often unfamiliar with the other. In this article, we assess both theories. We argue that the principal difference between the two theories lies in the mechanism by which discrimination takes place: discrimination in statistical models derives from an informational bias, while discrimination in status models derives from a cognitive bias. We also consider empirical assessments of both explanations, and find that while research has generally been more supportive of status theories than statistical theories, statistical theories have been more readily evoked as explanations for gender inequalities in the paid labor market. We argue that status theories could be more readily applied to understanding gender inequality by adopting the broader conception of performance favored by statistical discrimination theories. The goal is to build on the strong empirical base of status characteristic theory, but draw on statistical discrimination theories to extend its ability to explain macro level gender inequalities.

Citation

Correll, S.J. and Benard, S. (2006), "Biased estimators? Comparing status and statistical theories of gender discrimination", Thye, S.R. and Lawler, E.J. (Ed.) Advances in Group Processes (Advances in Group Processes, Vol. 23), Emerald Group Publishing Limited, Bingley, pp. 89-116. https://doi.org/10.1016/S0882-6145(06)23004-2

Download as .RIS