Why did this happen? A plausible answer follows.
First, neurotics are more likely to prefer math classes because they are terrified of being wrong. This is partly disposition and, I think, partly imprinting (maybe a response to lots of negative stimuli during development). Only math classes traditionally offer safety from getting wrong answers. Because of this, they are more likely to continue in math classes and work hard in them, and may study math to the detriment of other subjects, perhaps to the point of acquiring internal motivation to do so via fascination.
Second, neurotics who get advanced education in math are less likely to be capable of producing groundbreaking work of their own because neurosis means high latent inhibition, which means they are more likely to become educators. As educators, they are more likely to unduly punish small mistakes like missed minus signs with the same severity as big mistakes and incomprehension. This disincentivizes higher math education for equally high-IQ folks, ceteris pablum and such, who are not neurotic. They may argue that missing a minus sign is a really big deal because any amount of wrongness in an answer makes it wrong, which could have disastrous consequences, but the fact is that misplacing a minus sign on an exam that doesn’t matter, under a time limit, with no professional peer input, on material that a person learned last week, is not the same thing as bringing down the Challenger space shuttle.
I am not arguing in favor of incorrectness, I am saying come on people. Use your big dumb heads. Five point penalty max for that shit.
Third, as with many population dynamics a predominant personality type in a culture will eventually cause a tipping point, after which the personality feature becomes the default and deviance from the cultural norm is disincentivized at every step (even absent the gatekeepers in point #2). In cultures where people aren’t just plain mean, this manifests primarily as a communication problem. For instance, it is generally verboten (as I understand) to practice mathematical analysis outside of the French tradition because the French tradition gives the most exact definitions of the most abstract topics. You can’t define zero as “the integer between 1 and -1” or nobody will talk to you at math parties. You have to define it as the equivalence class [r, r], r ∈ ℕ. That’s how math people know you’re cool. You might get away with “the identity element of ℕ under addition” at work but you won’t get laid with any math groupies, no sir.
People who aren’t neurotic can still get by in math if they’re intelligent enough, because they will tend to make fewer errors for that reason, but sane folks have to be much smarter. The flipside of that is that the sane folks who make it through the grinder will tend to be much, much more creative and productive as working mathematicians. This is a simple result of intelligence and lower latent inhibition. They are also less likely to break down under the pressure of the singular professorial code of ethics: “Publish, yea, lest ye perish.” A neurotic person will be unable to publish anything except the most elementary extensions of existing theory, even if they can relax long enough to have an inspired idea. They’ll try to perfect that idea until they break down or die.
I mean, just imagine publishing an incorrect theory in a math journal. Your wife would leave you and your kids would disown you. Would your mother let you live secretly in her guest room? The oppressive shame would just be overwhelming, I can’t bear the thought even.