I think the underlying assumption is that by selecting traits that one has relatively little control over (foot size, thigh length, etc.), then showing participants a physically unrelated trait (i.e. the face), that they're somehow insulating from contemporary culture.

One could still make the critique that contemporary culture values as attractive traits that are correlated with these other markers, whereas other cultures and times do not, but that argument starts getting a little convoluted itself.

Keep in mind that this work hasn't yet been published (just reported at a conference), so the actual paper will probably be a bit more careful about claims that they make.

Gibson Girl: http://en.wikipedia.org/wiki/Gibson_Girl