It's risky business to infer anything from job preferences though. Consider that (as far as I know) an extraordinarily large proportion of professional programmers are male -- obviously a cultural phenomenon. We certainly cannot infer that a talent for programming is located on the y-chromosome.

Edit after reading the paper associated with that link:

> Our aphantasic and hyperphantasic samples were opportunistic, in the sense that our participants had approached us spontaneously following publicity triggered by our original publication

This suggests that many of the participants actually read scientific publications (or reporting thereof). That's an obviously biased set of people who indeed are way more likely to work in STEM (and have an abundance of time to spend browsing the web), rather than as hairdressers.