摘要：This paper reviews five reports measuring discrimination in salaries between males and females at Canadian universities. All find some discrimination (3% to 8%), a result in accord with published research on the same topic. However, the approaches taken are quite different, often reflecting controversial decisions over which variables would be included to explain salary differentials. We examine the strengths and weaknesses of these reports. In particular, the focus on single equation models is a problem since some of the controversial vari- ables, which may be biased by discrimination, also contain some information which explains legitimate differences in salaries. Our review suggests that many of the models are probably misspecified. We conclude with a call for universi- ties to collect the information which is required to complete these studies expe- ditiously and accurately.
其他摘要：This paper reviews five reports measuring discrimination in salaries between males and females at Canadian universities. All find some discrimination (3% to 8%), a result in accord with published research on the same topic. However, the approaches taken are quite different, often reflecting controversial decisions over which variables would be included to explain salary differentials. We examine the strengths and weaknesses of these reports. In particular, the focus on single equation models is a problem since some of the controversial vari- ables, which may be biased by discrimination, also contain some information which explains legitimate differences in salaries. Our review suggests that many of the models are probably misspecified. We conclude with a call for universi- ties to collect the information which is required to complete these studies expe- ditiously and accurately.
