Premium
Pay Discrimination Research and Litigation: The Use of Regression
Author(s) -
BLOOM DAVID E.,
KILLINGSWORTH MARK R.
Publication year - 1982
Publication title -
industrial relations: a journal of economy and society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.61
H-Index - 57
eISSN - 1468-232X
pISSN - 0019-8676
DOI - 10.1111/j.1468-232x.1982.tb00240.x
Subject(s) - library science , citation , state (computer science) , sociology , operations research , computer science , engineering , algorithm
THE USE OF REGRESSION ANALYSIS in empirical studies of pay discrimination within organizations and in court cases is a relatively recent phenomenon. Until the late sixties and early seventies, researchers interested in race and sex discrimination in pay had little in the way of established analytical constructs for investigating these issues. Even if such analytical constructs had been fully developed, researchers interested in empirical questions of this kind would still have had to contend with another problem: the limited availability of computer software suited to the kind of large-size data set (e.g., Census data, data on individual employees at particular firms) that would normally be used in such analyses. As regards litigation, for a considerable period after passage of the Civil Rights Act in 1964, employment discrimination cases were largely concerned with recruitment and hiring rather than with compensation, and frequently focused on persons in relatively unskilled jobs,. Detailed empirical analysis was not generally undertaken (see Ha~vard Law Review [ 19711 for a discussion of developments in discrimination litigation between 1964 and 197 1 ; and the review in Gwartney et al., 1979). Whether or not regression analysis might have been useful in such contexts, it does not appear to have been used. Times have changed. Many studies (especially Mincer's [ 19741 seminal work, which circulated widely in manuscript form prior to publication) have provided solid analytical foundations for empirical analyses of earnings; and development of powerful computer software such as the Statistical Package for the Social Sciences (Nie st al., 19'70) and the Statistical Analysis System (Service, 1972)-now permits analysis of larger data sets at much lower cost and greater speed than was previously possible. At the same time, the kinds of problems and issues that arise in litigation under employment discrimination laws have become more complex, making the use of