I'm genuinely interested to see the research that shows that women and minorities (other than those clearly disadvantaged by visa status) are being underpaid for the same qualifications and same work. I've worked at many companies, including as a hiring manager, and I have not observed anyone ever trying to intentionally underpay female or minority candidates. If anything, the female or minority candidates were sought after by multiple companies to improve their diversity statistics, so had a better bargaining position and eventual salary outcome than an equivalently qualified male non-minority candidate. On the other hand, I have heard with my own ears in company meetings (not my current employer) that we already employ "enough" non-minority males, and should prefer to hire minorities and women over others. I've personally experienced and also heard from others that they've been pressured to give female and minority candidates the benefit of the doubt when they would've failed interviews had they been someone else.

As an alternative explanation for any observed difference in pay... Anecdotally, there are lots of non-minority male nerds who spend ridiculous amounts of time reading tech books and working on work related side projects, to the point that they have no social lives and may never have a family. Even if we compare a sample of male and female candidates with the education and years of experience, a few outlier workaholic geeks (especially given power-law compensation) could throw off the averages and make it seem that women are discriminated against.

I'd also like to see research that controls for things like... what school did the individuals study at, not just their degree? A CSU grad will probably make less than a UC grad. What classes did they take? (To control for people who take minimums vs take intense courses.) How many hours per week do they devote to their career outside of work on average? How many hours do they work on average? And so on. If these are difficult variables to control for, that makes my point... we can't say it's a result of discrimination unless we've controlled for all the actual variables.

To zoom out for a minute... I'm very much interested in everyone getting equal opportunity and equal pay for equal work. But it does NOBODY any favors to try to make political points with sloppy or intentionally misleading research, because all it does is alienate potential allies and make them think the debate is not honest.