[gwern]

> After further edit: There is recent research showing that learning can actually change brain neural connections

Yes, of course learning changes neural connections. How else would it work? Changing some chemicals won't get a brain very far.

But this is as far away from changing IQ as demonstrating the Casimir effect gets you to a tractor beam sucking in the _Millennium Falcon_. The famous study of the [London taxi](http://news.bbc.co.uk/2/hi/677048.stm) drivers shows they have changed connections, alright, even enlarged hippocampi - but nothing about increased brain mass or volume, because it's a zero-sum gain.

Now, as far as my specific assertions go. The 0.05 point is basic statistics which I feel no need to justify. The short-term point flows out of the former (there's simply more noise when you do all your sampling over a few days or minutes, say), and short-term studies are particularly vulnerable to issues like testing effects; the Balkans thing I alluded to in enough detail that you could easily have googled it if you were genuinely curious: http://www.google.com/search?num=100&q=balkans%20IQ%20io... And iodine deficiency isn't something that's purely long-term either - you can see effects after 24 weeks in this study: http://www.ajcn.org/cgi/content/full/83/1/108 (note the high significance).

The burden of proof is not on those who claim lack of malleability - not that there are any such people here - but on those who claim IQ boosts without good long-term IQ testing.