It's not measuring performance at calculus. It's measuring confidence in performance at calculus. This is where sexism and all those icky social things come into play. Most women have spent their lives up to that point hearing subtle and not-subtle variations of "girls can't math". So even if they're good at it, they tend to believe they aren't.

Meanwhile, college-age men tend to be brimming with confidence, walking examples of the Dunning-Kruger effect. Even if they're not good at calculus, they often think they are. This is partly a result of a society that says, directly and indirectly, that people like them are natural rulers.

That's how sexism works, in both directions.