Equal prospects? How do we know?

Saw this story in the Naturejobs Newsletter: "Equal prospects for both sexes in science: Women as likely as men to be promoted, but fewer apply for jobs."

The way Natasha Gilbert seems to be spinning this story is as more evidence for a confidence gap: women don't face discrimination during promotion, and would succeed, if only they would make the effort.

But: is the best way to analyze these data to compare all women faculty to all male faculty (as this study does), or is it better to compare the women to the top 37%* of male faculty, and then see whether there are promotion discrepancies?  I.e., if it's really true that there are all these qualified women out there who aren't seeking faculty positions, who, if only they did, would bump out the bottom male faculty currently getting positions, then wouldn't that model predict that the women who do have the confidence to seek faculty positions are more qualified than male faculty as a whole, and wouldn't we predict women faculty to have a higher promotion rate than male faculty as a whole? Or does confidence not correlate with ability?

Not only would this experiment be hard to do, because I'm not sure what the best way is to identify the top 37% of male faculty, but also, the skills that get you hired might not necessarily get you promoted. (I.e., maybe you yourself were a very productive postdoc, but you turn out to be terrible at managing a lab.)


What do you think would be the best way to measure whether men and women of equal quality are being promoted at the same rate?


*Why 37%?  I got this ballpark figure from the fact that 27% of incoming faculty are women, so 73% must be men.  27/73 = 37.

This calculation is also problematic, though, because it assumes that postdocs are 50/50 women/men, and that those came from a pool of grad students that were 50/50, and those came from a pool of undergrads that were 50/50, etc.  If these ratios are less than 50/50 at earlier stages, then you would predict women faculty to be even more enriched for qualified candidates.  (Unless, once again, confidence doesn't correlate with ability.)  And that's not even considering the quality of the institutions that the postdocs are coming from.  In summary: the proper analysis here would be complicated indeed.