Student evaluations of instructors, aka ‘how much do you like your teacher?’, are nearly ubiquitous. But should they be? An increasingly large body of evidence says heck no. For one, students are actually poor evaluators of whether their teacher is effective or not.

Secondly, student evals are straight up sexist. In two studies, a ‘natural’ one in France and a blinded study in the US, it found significant differences between the scores given to male instructors and female instructors. The US one is especially damning. It had the same two teaching assistants, one male and one female, leading 4 discussion groups for an online course. In 2 of them, the assistants used their real names; in the other two, they swapped names, so that the woman had the man’s name and the man had the woman’s. Everything that could be controlled was; assignments were returned at the exactly the same time, for example.

Since assignments were returned at exactly the same time in all four sections, the significantly lower rating for female instructors (what equates to about 16 percent of full scale) “seriously impugns the ability of SET to measure even putatively objective characteristics of teaching,” the paper reads.
Again, Stark and his colleagues found that, in contrast to the French data, perceived male instructors were rated significantly more highly not by male students but by female students. Male students rated the perceived male instructor somewhat significantly higher on only one criterion—fairness (p-value 0.09). But female students in the U.S. sample rated the perceived male instructor higher on overall satisfaction (p-value 0.11) and most aspects of teaching. Those include praise (p-value 0.01), enthusiasm (p-value 0.05), and fairness (p-value 0.04).

Let’s stop using these ineffective tools and find better ways to judge teachers’ effectiveness.