Crazy study. If women hide their gender, their rate of acceptance is higher than men. If they are openly female in their profiles, the acceptance rate is lower than men. I like to think that I am gender blind, that I only care about the quality of the submitted code, but maybe I’m not. This study certainly suggests that there is an institutional bias.
As always, sample population studies do not necessarily reflect the real situation. In this case, having the counterfactual to support (the data regarding women who did not openly identify their gender in their profiles) gives the report a lot of weight however.