This seems like an extremely flawed methodology - it would seem that if you are being shown photos of mates, you can see the attractiveness with your eyes, whereas simply being told the less attractive ones are “more intelligent as reported by their peers” (what does this even mean?) is not as easily visible. Since the mechanism of choice is visual, of course people are going to respond mainly to visual cues. I bet the results of this study drastically change if the women themselves are allowed to interact with mate choices and rate intelligence by their own metrics. What a silly way to go about this.