Rather than leaving it at that, I thought I'd do a little digging and see what educational esearchers have to say about this subject. After a bit of web-searching, here's an overview:

Your evals will improve if you are (in no particular order):

1. enthusiastic - (click here )

2. male (esp. in a male-dominated subject like mathematics - here for a bibliography)

3. teaching a large course and are perceived as "liberal, neurotic and extroverted"

4. teaching a small discussion-oriented class and are perceived as "gregarious, adaptable and supportive" (these last two are pulled from a nice overview of the research on teaching evaluations called "Instructor Evaluations and the Politics of the Classroom" )

I was somewhat surprised to not find anything on age or race - then again, I didn't spend more than half an hour looking. (It might also be an indication of the number of minority professors out there.)

My conclusions: Don't despair. If you can use evaluations (formal or otherwise) as a formative tool - to help you understand the effects of your teaching - that's great. Just pray that they aren't being used as a normative tool - to measure your "effectiveness" as a teacher, evaluations fail miserably.

The idea that we can easily measure something as complex and multi-faceted as teaching effectiveness is a very dangerous one. Unfortunately, it's an idea shared by many college administrators and by our current President.