Summary: If you did well on the SAT or attended graduate school, you were more likely to become a professor or receive NSF grants.
This seems to measure people who have bought into "the system" in a particular way, rather than measuring exceptional people.
Paul Graham's essay today (http://paulgraham.com/lesson.html) seems especially apt. These people have learned to hack certain standardized tests.
Statistically most of them just aren’t that bright.
As measured by standardized tests?
But it's 30% research/academic work, 70% playing the system.
From my friends with PhDs, it seems like it's about 30% doing real stuff, 10% fucking around with grant funding paperwork and 60% trying to turn a single result into as many papers as possible in order to maximize impact scores and other metrics.