MIT BLOG: Stereotypical Statistics

[Ryan Goble]

The summary below comes from my Hannah Arendt Center newsletter...

Stereotypical Statistics

An author identifying herself as Michelle G, a student at MIT, acknowledges, "There is empirical evidence to support the idea that males have a higher capacity for spatial reasoning than females." But Michelle G. dives deeply into these and other studies to argue that such "evidence" is a "factual misconception." "I'm guessing that you're familiar with common notions that men are spatial and logical thinkers, while females are more verbally proficient. A man being tested for spatial ability might assume that he's going to have an easier time than a woman of otherwise equal intelligence, his conclusion based not on sexism but on objective science. And statistically speaking, he's right. It is true that men score higher on spatial reasoning tests, though you might have caught on that there's a little bit more to this picture (why would a female MIT student publicize stereotypes that actively work against her?). If you're now wondering whether I'm about to throw some kind of feminist rant at you, I'll give you a 'well, sort of,' because calling out factual misconception is just as important as promoting feminist ideals here, and because I think those two go hand in hand anyway. I'll largely put the romance of egalitarianism aside, though, to talk about empiricism.... I think it's important to acknowledge the very rightful discomfort that arises when scientific studies attempt to trace such differences to biologically determined origins. Yet, across decades of research, no biological cause has actually been identified as a suitable explanation for the spatial reasoning discrepancy. Studies regarding testosterone and mental rotation, for example, found inconsistent or absent effects across cultures, prompting inquiries into 'differing cultural values' to account for the results. And gaps between men's and women's scores on some spatially-geared tests have significantly shrunk in the past few decades, which is interesting because noticeable evolutionary or nature-based development might take thousands of decades to take effect. ('Nurture'-based conditions are of course rapidly changing.) Still though, the gap has lingered, and a satisfying and empirically-supported explanation as to 'what gives' was not achieved until 2008, when researchers eliminated the performance gap under a single simple condition."