“Characterizing the test as insensitive to gender differences was enough to totally eliminate women’s underperformance in this experiment. Yet when the same test was characterized as sensitive to gender differences, women significantly underperformed in relation to equally qualified men.”

http://www.tandfonline.com/doi/full/10.1080/15298868.2012.68...

http://www.ncbi.nlm.nih.gov/pubmed/19487665