Wednesday, August 23, 2006

Woebetide the partisan hack professor 

Frequent reader and political scientist Steve Frank sent to the campus a synopsis of April Kelly-Woessner and Matthew Woessner's article "My Professor is a Political Hack: How Perceptions of a Professor's Political Views Affect Student Course Evaluations," published last month in PS, a journal of the American Political Science Association. The professors testified before the Pennsylvania legislative committee investigating academic freedom, and in some sense are following on the research agenda of Dan Klein (for instance, the likelihood of liberals getting higher SSCI counts than conservatives.)

The study uses survey results from thirty faculty teaching undergraduate political science at 29 colleges to 1,385 students. Interestingly, 27 of these faculty came from a postcard mailed to two hundred instructors (chosen at random) asking if they would agree to distribute the survey. 22 others wrote back to say they would not. You wonder -- why not just pitch the postcard requesting their participation?

But did it really matter to the sample? 23 of the thirty faculty are rated as left of center, with one person called "very liberal" and one "fairly conservative". When asked if they could be certain of this there was a good deal of variability. Also interesting to me -- the students were able to discern the ideology of the other students in the room, according to Kelly-Woessner and Woessner.

So, if you buy that students can tell the ideology of their professors, what does it matter? The authors explain:
[S]tudents rate faculty members who they perceived to be liberals more favorably on a number of faculty characteristics measures. As a whole, students are more likely to report that liberal professors �encourage students to express their own viewpoints,� and �work to provide a comfortable learning environment.� When professors are perceived as either liberal or Democrats, students are more likely to believe that their instructor �cares about students and their success.�
Those kinds of perceptions appear to be true both within a classroom and between classrooms. And it's likely that those things will be looked at by evaluative faculty committees and academic administrators in deciding who gains promotion and tenure.

Perhaps the most damning point is one for either conservatives or liberals to like. Kelly-Woessner and Woessner find that if you are perceived as being at an extreme in either direction, students systematically believe you are less able to give an "objective presentation". If I had a nickel for each time a faculty member at SCSU said to me that they were able to keep their politics out of the classroom so their extremism shouldn't be an issue, I'd probably quit my job. 35% of students were significantly different from their instructors in terms of party identification (versus 23% difference in ideology.)

How big a difference is it? On a five-point scale, the overall effect of being at least two points away (on a four-point party ID scale) was about 0.3 points. The highest and lowest overall rating averages in my department are less than one point, so 0.3 is meaningful. And being two points away could be simply your being a political moderate in a classroom full of liberal Democrats or conservative Republicans. It appears simply being away from the median view of students in either direction gets you in trouble.
Consequently, if the goal were simply to win the love and adoration of the students, clever instructors would merely pander to the median �voter.� By mimicking students� views and reinforcing long-held beliefs, professors might score well on student evaluations, while providing no useful information at all. Indeed, many students would be more comfortable with a course if they could skip the readings or forego exams. Yet, college is not Club Med. As instructors, we ought not to refine our pedagogy exclusively for the purpose of making students comfortable or improving course evaluations.
Jason Czarnezki notes that perhaps teaching to the median student ideology isn't a good thing:
I think it points to some flaws in student course evaluations--what do they measure?; can they be manipulated?; is there a gap between student perception of what is good teaching and what is actually good teaching?
The answer to the manipulation is usually yes. I can give you one easy example -- if you return an exam or term paper the same day you take evaluations, your evaluations will be influenced by whether the students did well or poorly. If I wanted to spike my evals, then, I give an easy exam a week before evaluation and return them before handing out the evals. Kelly-Woessner and Woessner find that Democratic faculty are more likely to be perceived as grading "fairly and consistently."

Now at SCSU this is less of a problem, since faculty can report whatever parts of their evaluations they wish; there are some areas where the differences are sufficiently small. And you don't necessarily have to use evaluations at all. It would be my preference not to use them, but they are still considered a normal part of the evidence one provides for effective teaching; it is quite normal for job ads for faculty to make evaluations part of the material submitted with applications, particularly at more teaching-oriented institutions.

Maybe as this fertility gap continues to grow, it will be more conservative or Republican faculty that are treated well in student evaluations. But that's not currently the case -- 43% of the students in the Kelly-Woessner and Woessner sampe self-identified as liberal versus 29% conservative, with extremely liberals at 13% versus extremely conservatives 6% -- and so at this time the bias does cut against the conservative or even moderate faculty member.