Teachers often find student evaluations rather frustrating. They are contradictory, short-sighted and sometimes insulting, especially when students did not put much effort in the class in the first place. Student evaluations are also biased towards teachers who are physically more appealing. And students, with their lack of experience and expertise, are not in a good position to evaluate an expert. What more could be said against student evaluations?
Michela Braga, Marco Paccagnella and Michele Pellizzari find that better teachers get worse evaluations. The way they measure teacher effectiveness is by looking at how students do in subsequent classes. They find that teachers matter, and substantially as the teacher can explain 43% of the standard deviation in subsequent grades. But the good teachers get a worse student evaluation, which is frightening, because administrators are getting the wrong message.
From the tables, I gather that higher ranked faculty teach better, but older and researchers with higher H-indexes do worse, which is rather contradictory. I wonder whether taking into account the obviously high correlation between some of the independent variables would take care of this, or other controls, like the attractiveness mentioned above.