I just got back my teaching evaluations from last semester's graduate algorithms class. Here is some of the more useful/interesting/insightful negative feedback.
- The evaulation seems a little flawed... Learning the material won't help you, which leads to hi-IQ people performing way better than hard workers. Is this fair?
- Be more willing to say "I don't know"—I caught a fair amount of fudging, which hurts your credibility!
- Not enough discussion of "big pictures", eg, general ideas of designing randomized algorithms instead of many examples
- Split into two semesters; do more stuff.
- HW solutions...When we make some mistakes, even typos, they subtract our grade, but they often have mistakes in the solutions they give us.
- Instructor occassionally (and in particular initially) gives the impression of being aloof, tyrannical, or both. (Instructor seems to be in fact neither.)
- Loose enforcement in the name of "courtesy" led to rules being viewed with disdain and made it hard for course staff to later enforce expectations.
- Sometimes too aggressive and hot tempered
Sad to say, these are all valid criticisms. (The first one is a bit exaggerated, but there is a core of truth to it. Talent helps.) And the aloof tyrant thing might explain why almost nobody came to my office hours, while my TA Erin's were constantly flooded.
This was the first time I taught a required algorithms course only for graduate students (as opposed to grads and undergrads together). Unlike some undergraduate classes I've taught, my graduate algorithms students repeatedly exceeded all my expectations. That's not to say that everyone did amazingly well, but the average levels of insight and algorithmic maturity—as well as the average homework and exam scores—were much higher than I expected.
Same with the evaluations. They didn't let me get away with anything!
So, if any of you students are reading this, thanks.