Hacker News Reader

Comment by magicalist | Hacker News Reader

Ugh. That map is terrible. I'll just quote wikipedia on this one:

For 104 of the 185 nations, no studies were available. In those cases, the authors have used an estimated value by taking averages of the IQs of neighboring or comparable nations. For example, the authors arrived at a figure of 84 for El Salvador by averaging their calculations of 79 for Guatemala and 88 for Colombia. Including those estimated IQs, the correlation of IQ and GDP is 0.62.

It's not hard to find corroborating evidence, you can just make it up!

[1] http://en.wikipedia.org/wiki/IQ_and_the_Wealth_of_Nations#Na...

View on HN

More importantly it didn't even support the argument presented.

I guess what they intended was "African Americans have low IQ and so do Africans, ergo IQ is race-related". Yet African Americans have higher IQ than Africans, because they live in a more prosperous region, with less disease, more opportunities etc. so what does that tell us about IQ and race, and racial differences within countries? That maybe it's not all about race (and generally be on the lookout for stats that don't correct for factors like age/wealth/education/sex etc. I regularly see the "headline" figure being given uncorrected.)

A pretty big racial cognitive gap remains after adjusting for income, at least in America. Poor whites do better on cognitive tests than wealthy African Americans: http://glpiggy.net/2013/03/19/a-sociologist-looks-at-race-di...

African Americans are closer to European IQ levels than Africans, probably for several environmental and genetic reasons (better nutrition and schooling, ~25% European ancestry, etc)

That's true, that a gap remains when you adjust for a some obvious confounding factors, which makes it all the stranger that people are so determined to use the non-adjusted figure, it's as if the bigger number somehow resonates more with what they want to believe, than the more scientifically valid one. Maybe they're scared a that a smaller gap could be bridged by simply examining a few more variables.

Reply on news.ycombinator.com