3 ms·
Doesn't seem to be mentioned in any other comments or the paper itself, but this is Berkson's paradox. https://en.wikipedia.org/wiki/Berkson's_paradox https://
by ladberg 3y ago
Doesn't seem to be mentioned in any other comments or the paper itself, but this is Berkson's paradox.
https://en.wikipedia.org/wiki/Berkson's_paradox https://en.wikipedia.org/wiki/Berkson's_paradox
- projektfu 3y agoYou're saying that unattractive, unperformant fund managers have nothing to recommend them, and so that makes the correlation negative because they're censored from the sample? Otherwise it would be roughly flat, if they didn't let the unattractive bad managers go? Could be. But then, it's still a reasonable heuristic, because you might find that unattractive managers are uniformly pretty good or better, and attractive ones are 50-50. The best manager might be one of the really really good looking ones, but overall there is less selection pressure on attractive managers with respect to performance. (Halo effect?)
- ladberg 3y agoYeah exactly. And sorry I wasn't trying to disprove the study or anything, just providing the name of one of the effects they're describing.
- apienx 3y agoI don’t have access to the paper. I wouldn’t be surprised if they didn’t account for such sampling bias effects. Hope they did though (for science’s sake!).