I did find it quite interesting that the men seemed to do worse with the values exercise than without it, and strange that the article didn't discuss this aspect. The idea I had was that if the men are subject to the same stereotype (that they are inherently better than women at physics), and then they find themselves in a class where the women are doing just as well as them, that they would lose some confidence. Since they only did slightly worse, I think this is perhaps a reasonable explanation.

If so, then I think it is ethical, because all you're doing is challenging an incorrect stereotype that was unfairly giving men a slight boost and women a significant handicap.