The basic point, that *conditioned on test score*, a group with less test preparation will on average have a higher inherent ability than a group with more preparation is obvious enough. But the genius is in finding a situation where imposing a minimum quota on *males* improves results! That's bound to wind up the "meritocracy".
I wasn't as impressed by Sethi's claims for monotonicity, unconnected as they are to any empirical support; in that context, the "extremely likely" assertion has a contrary effect. I think it plausible that training is weakly observable (in the sense of a group average) but his model rests on the assumption that it is strongly observable (for each individual) and he ought to provide some empirical support for this assumption even if he can't find any for his conclusion. But whatever, that is a secondary question IMO.
Ha ha, that Schaede & Mankki paper is great!
The basic point, that *conditioned on test score*, a group with less test preparation will on average have a higher inherent ability than a group with more preparation is obvious enough. But the genius is in finding a situation where imposing a minimum quota on *males* improves results! That's bound to wind up the "meritocracy".
I wasn't as impressed by Sethi's claims for monotonicity, unconnected as they are to any empirical support; in that context, the "extremely likely" assertion has a contrary effect. I think it plausible that training is weakly observable (in the sense of a group average) but his model rests on the assumption that it is strongly observable (for each individual) and he ought to provide some empirical support for this assumption even if he can't find any for his conclusion. But whatever, that is a secondary question IMO.