There are some who claim that it does not. In particular they claim that there is a longer tail on both ends of the curve than a normal distribution would pred

There are some who claim that it does not. In particular they claim that there is a longer tail on both ends of the curve than a normal distribution would predict. http://www.abelard.org/burt/burt-ie.asp http://www.abelard.org/burt/burt-ie.asp But there are others who refute that claim. Their analysis shows there are actually fewer supernormal outliers than you would expect for a normal distribution. http://www.ncurproceedings.org/ojs/index.php/NCUR2012/article/download/159/258 http://www.ncurproceedings.org/ojs/index.php/NCUR2012/articl... Either way it's a pretty close fit and the argument is just about whether there is a small variance several standard deviations from the mean. And, yes, there has always been disagreement about what IQ really is, what do IQ tests measure, what forms of intelligence aren't measured in IQ tests, etc. I've always settled for the tautological definition. IQ is what an IQ test measures.