First, I'm just a layman; maybe he really is full of it. But I thought his claim that IQ is validated using correlation against performance measures that (may?) have fat tails is reasonable: the correlation might not even exist. Is the normality assumption justified?
-
Maybe they did test for normality, I don't know. If so, then Taleb's just ignorant.
Of course these trivial things have been tested already. Any scatterplot would reveal some kind of fat tail or whatever his ad hoc hypothesis is. Here's a random one for income, log transformed or raw. http://humanvarieties.org/2016/01/31/iq-and-permanent-income-sizing-up-the-iq-paradox/ …pic.twitter.com/NO17jhgjUm
