Depends on the effect size. The effect size here, 0.68, is pretty insane for a trivial change of this kind. (if accurate) I think it is pretty clear, from ind

Depends on the effect size. The effect size here, 0.68, is pretty insane for a trivial change of this kind. (if accurate) I think it is pretty clear, from independent research, that student evals capture their eval-time happiness and little else. This is just another piece of evidence to that effect. An eval of this kind cannot really capture anything that useful. Esp. at a university, it would be better to ask people to eval their 101 prof.s at the end of their 201 courses -- ie., with sufficient distance to appreciate their impact on their learning.