I'm kind of conflicted about that. A strongly explanatory variable is valuable to have in order to control for its effects. Voluntarily removing it reduces the

I'm kind of conflicted about that. A strongly explanatory variable is valuable to have in order to control for its effects. Voluntarily removing it reduces the usefulness of the derived statistics and in this case I guess seems to make ratemyprofessor just a niche version of hotornot. Perhaps a middle ground would have been to allow students to rate "hottness" and explicitly controlled for it when ranking professors, but never making a given professor's "average hotness" viewable to users?