Uh oh, I think I've used this data in exams before. At first glance, I don't buy this study. So the deal is, there is a long-running german omnibus survey whic

Uh oh, I think I've used this data in exams before. At first glance, I don't buy this study. So the deal is, there is a long-running german omnibus survey which is (was?) administered face to face. It includes interviewer assessed and self-reported beauty. Those two correlate, though not perfectly. Now what the paper does (at least the preprint that is publicly accessible on ssrn) is to systematically test beauty against a bunch of dependent variables. Turns out, there is a somewhat stable but weak correlation with the belief that success is earned (vs. luck), and a weak correlation with voting for the german liberals. However, the entire paper smells of a fishing expedition. There's a ton of linear regressions with many variables, not all of which are very straight-forward predictors. The theoretical link is also somewhat contrived. So I would be very careful in interpreting this kind of result until someone does a pre-registered replication. Corrections and comments to this view are highly welcome.