Going through the main study - I'm just picking out things that struck me as off, so this isn't intended to be especially even-handed. And I'm not saying these are necessarily things the authors didn't accept (they explicitly accept some of them), but they might change how far we take their conclusions:
- They're looking at a very particular group of men, specifically white British undergraduates between 18 and 42, which is interesting. They accept 'ethnicity' as a possible influence, so exclude other ethnicities.
- Their test group has a substantially heavier top limit for the men's own BMIs than the control - the test group ranged from 17.15 (just slightly under normal), to 31.64, which is into the obese range. The control group, OTOH, ranged from 16.53 (underweight) to only 27.76 (overweight). Since the standard distribution of BMI in both groups is the same, these are significant differences between the two groups.
This is interesting, because it's possible the men's own body sizes would affect their preferences. They reckon they discounted BMI by finding no significant differences across the groups, but I'm curious about this one - I wonder what another study focussing on the BMI of men would show? Here, I'm thinking it's interesting they want to discount this factor in the men, whereas they want to focus on it in the women - methodologically, why would they choose to do it that way around? (I mean, I know why, but it's interesting to look at which experiments get done, as well as how they're done).
Interesting point not flagged up in the BBC article - the stressed men not only preferred bigger women, they preferred a wider range of women to the control group.
They point out their study's limitations - they had the control group sit around doing nothing while the test group were being stressed, so maybe the control group's boredom accounts for something too.
My feeling is, while this study has some methodological drawbacks (largely much acknowledged by the authors), the main problem is they push their conclusions far beyond what the study can show, and their abstract makes a pretty unforgiveable generalization from men's responses to 'human' responses. It's not replicated in the article but it's dodgy. The BBC then generalizes it further.
It's fascinating that it's the size issue, men 'preferring bigger women', not the expansion of the range men prefer upwards in BMI, so men 'preferring a wider range of women', that's been picked up on.