| Summary: | This text aims at looking at some advances and above all some problems coming from the study of « racial » discrimination (« whites » and « non whites ») in the labour market (hiring, promotion, wages) in the United States since the 1960’s by statistical models of mutiple regression. In a first part we describe the progressive generalization of this method in the 1960’s and 1970’s for this subjet, showing an evident lack of precision in the beginnings on methodological issues. In a second part we show that many important problems remain in spite of a tendency of methodological improvements. We suggest in conclusion a discussion about the usefulness, after half a century of practices, of trying to « explain » discrimination with this kind of tools.
English Title: How to Study “Racial” Discrimination on the Labor Market? The Use of Multiple Regression in the United States since the 1960s
Keywords: Racial discrimination, Multiple regression, History of statistical methods, History of social sciences in the USA
|
|---|
