> We'd have to look at a meta-study to weed out problems in individual studies

Or we could try not generalising outside the scope of an experiment. The experiment doesn't even cover all faculty hiring.