The HN submission title (EDIT: The title is “ Attractive students no longer receive better results as classes moved online ” as I write this, in case it gets updated) isn’t from the paper and isn’t entirely accurate. The paper actually found that male students who scored higher in the beauty ratings continued to get higher grades as well:

> On the contrary, for male students, there was still a significant beauty premium even after the introduction of online teaching.

So the submission title’s claim that attractive students no longer receive better results when teachers can see their face isn’t true. The result was only detected for female students.

The fact that there is a discrepancy doesn’t give me a lot of confidence in the results. When you can only find a significant change after you start subdividing the group into different sub-groups it’s getting a little too close to p-hacking for my comfort. That’s not to say there isn’t a gender effect here, but the fact that males rating high on the beauty scores also got higher grades should suggest that this isn’t as simple as teachers biasing their grades based on what the students looked like.