Earlier quoted context omitted.
Let me first say that you are right that her comparison of 'cowboy' vs. 'good developer' is completely ridiculous. Talk about her prejudices showing in conjuring up that straw man. I've never worked with any developer that fitted in the cowboy category, while most did their best to fit in the 'good developer' category. However: Can't you say that about any form of discrimination? She could, but that's not what she is…
...female students in CS are being discriminated against. [...]but research in discrimination against women is crystal clear on this. [Citation needed.]
I can't do any better than Googling it; that's what I did years ago when I was looking to verify that women really were being discriminated against, because I figured we would surely be past such barbaric behavior. Since then, I wised up on the fact that we're just not really that much in control of our behavior. I'm not a psychologist, but I do know how to find and apprehend research in such a domain. Meta-analysis studies are quite convenient :).
[1] http://web.mit.edu/fnl/women/women.html
[2] http://www.fas.harvard.edu/~mrbworks/articles/1995_PsyRev.pd...
[3] http://econ2.econ.iastate.edu/classes/econ321/Orazem/darity_...
[4] http://economics.uchicago.edu/download/JLISTdisc.pdf
[5] http://www.tamu.edu/classes/psyc/payne/ID/ID%2520articles/He...