Actually I agree with UT here, I suspect that IQ is a pretty ****ty measure of intelligence and that the well-known variance changes between men and women are probably in part social. Also I think it's a little strange to believe that very high intelligence is required for presence in top research academia; it's not really about being smart.
For my part I think IQ is a generally accurate measure of potential. A high IQ doesn't guarantee ability but someone with an average IQ is going to have a really hard time winning a Nobel Prize.
Trying to trace back through this conundrum of threadlocksplits has proven difficult, but this seems to be the most logical point of entry. Thus,
IQ is a meaningless measure of intelligence, though less so now than in the past. However, modern IQ tests are still cross-culturally useless and the true characteristics of intelligence, which haven't been adequately defined in the testing, may not even be measured. Considering a lot of psychology and biology has a great deal of difficulty determining even what intelligence IS, I have a lot of trouble with the notion that we can measure it objectively with a written (or oral) test.
As it is, IQ tests (generally speaking) can measure fairly accurately where (geographically) you were raised, what social "class" you belong to, some aspects of your morality, and the combined effects of how good your education was, your engagement was, and your memory is. It also depends on the quality of the test - I've scored anywhere from 116 to 140ish, with the "most reliable" score at 126, but in all cases it was fairly evident that the test incorporates the biases of the test designer. It's an utterly worthless number.