Sexy Professors are Better Professors (?)

I couldn’t read this report without hearing the 1970s Rod Stewart song in my head: “If you want my body and you think I’m sexy, come on sugar let me know.”

I’ve previously blogged on, the role of sexiness in student evaluations, and the limitations of anonymous student feedback. See here and here. This report took the entire dataset, regressed it for correlations between “quality” of instruction and professor “hotness,” and found a 0.64 correlation between the two. The authors try to discuss with a straight face the possibility that there may be a recursive effect where students find teaching brilliance as sexy…hah! (They write: “most student comments point toward Quality as a function of Hotness when they focus on physical characteristics of their professors that could be captured in photographs”). Ruling out this possibility, it is almost impossible to reach any other conclusion than that, in this dataset, sexiness contributes to assessments of professor quality.

So what take-away points can we get from this? The authors cite this as another reason to believe that student evaluations of teaching are generally unreliable. (“Taken as a whole, these self-selected evaluations from cast considerable doubt on the usefulness of in-class student opinion surveys for purposes of examining quality and effectiveness of teaching.”) If, in fact, student evaluations are influenced by such factors as professor attractiveness, then there is good reason to be suspicious of them. I am planning to attach this article as part of my tenure review package to explain some of my teaching evaluations (I’m making the highly defensible assumption that this factor is working against me, not for me).

It seems there might be another obvious conclusion to draw. If I want to improve my teaching evaluations, I should not invest more time in class preparation or subject material mastery. Instead, I should hit the gym.

Seriously, though, it would be easy to overinterpret this study as it relies on self-selected data (students opine at voluntarily). But this report has some very troubling implications for gender, age, race and physically challenged bias in student evaluations (the data also shows possible bias based on discipline–apparently, geeky scientists get hit hard on the sexiness-o-meter). At minimum, it is a good reminder that any evaluation metric (such as, in this case, the metric for evaluating professor teaching performance) must itself be evaluated for credibility. In the case of student evaluations, it is way too easy to overweight the precision of the “numbers,” when the entire numerical dataset might be skewed by bogus exogenous factors (like sexiness).