His characterization that the quality of papers from men must, by definition, by higher quality clearly establishes the fact that he is a textbook example of the problem.
He didn't say that they, must be of a higher quality. He said that it's a possibility that shouldn't be ignored. You can't just assume it's not true.
Personally, I think the problem is that we try to use science to evaluate things it's ill-suited to do. "How gender differences affect the experiences that PhD students have when moving into post-doctoral work" is not a subject that's best examined using the scientific method. If one wants to come to a real understanding of this issue I would suggest asking a bunch of PhDs, both male and female, to write essays about it from a purely subjective point of view, and put those essays into a collection. It would certainly be much more enlightening than survey data and whatever various statistics were compiled in an effort to make a scientific study out of the whole thing.
I love science. It's great method for discovering truth. But I hate it when people try to apply science to social issues.