I want to suggest that caution is in order. The statistics used to report "wage gaps" such as these are medians and therefore they are aggregations. When you look at aggregated data you are unable to see the differences in the details that are aggregated. No one works in aggregate. People work individual jobs in specific settings. And people make different voluntary choices that result in the specific jobs in the specific settings. A "wage-gap" based upon aggregated information need not be the result of discrimination in the labor market.
The commentary by Carrie Lukas at National Review Online illustrates why caution is in order by telling us about Warren Farrell's Why Men Earn More. For example,
"Not only do women take more time out of the labor force and work fewer hours than do men, women also avoid jobs that require a great deal of travel or relocation. Men assume more high-risk jobs -- 92 percent of occupational deaths occur among men -- and endeavors that require braving the elements outdoors."It seems to me that "wage-gap" numbers cannot be relied upon as evidence of labor market discrimination.
". . .Dr. Farrell outlines how women can increase earnings, but notes that higher pay often comes at a price -- be it greater physical risk, more time on the road, or more hours in the office."