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.
Did either of you read the article? The control group picked "their least important values and wrote about why these might matter to other people." The article says nothing about how either men or women perform without a values exercise.
I don't think that exercise, in itself, would make anyone perform worse at physics. While writing about why your values are important to you could very plausibly increase your confidence, I don't see how writing about why other people might care about things you don't value would decrease your confidence or performance.
I think it mirrors the way women might experience a physics class. They might feel like the class is about things that other people value, and that might make them perform worse.
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.