Now that student evaluations of faculty are taken more seriously than anyone ever dreamed—and posted online, thanks to the Texas Legislature—it probably makes sense to study them empirically. (One might suppose that study would precede policy, but no.)
A scholar from Sweden, Anamaria Dutceac, has posted an interesting article in Inside Higher Ed. She points to studies from around the world indicating that student evaluations reflect factors that don't have anything to do with teaching.
1. Humanities courses tend to get better evaluations than science courses, regardless of the variation within the respective curricula;
2. Courses with fewer students (the borderline is at around 20) get much more positive evaluations than large courses;
3. Courses at the advanced level get slightly better evals than those at the basic level;
4. Optional courses are better appreciated than obligatory ones.
Now this is interesting enough, but she also reports evidence indicating that age and gender are also related to evaluation scores. Combine the two, and older men get rated quite a bit higher by students than younger females.
We have all seen student comments on evaluations that are irrelevant to pedagogy. One favorite student observed that brown socks and blue pants don't go together, Dude. Thanks for the tip!
More seriously, there should be enough data now for academic research to examine these variables in the aggregate—gender and age especially. This subject deserves analysis. Dissertation anyone?