Tuesday, March 29, 2005

And, in an astonishing development... 

...college faculty are mostly liberal. (Gasp!)

Howard Kurtz reports on a study (you can get it here with free registration) by three political scientists of responses to six ideological questions, political and religious identification by 1643 academics at 183 universities across the Carnegie classification spectrum (from top-20 doctoral programs through comprehensive schools like us to small liberal arts colleges). From the abstract to the study:

A randomly based national survey of 1643 faculty members from 183 four-year colleges and universities finds that liberals and Democrats outnumber conservatives and Republicans by large margins, and the differences are not limited to elite universities or to the social sciences and humanities. A multivariate analysis finds that, even after taking into account the effects of professional accomplishment, along with many other individual characteristics, conservatives and Republicans teach at lower quality schools than do liberals and Democrats. This suggests that complaints of ideologically-based discrimination in academic advancement deserve serious consideration and further study. The analysis finds similar effects based on gender and religiosity, i.e., women and practicing Christians teach at lower quality schools than their professional accomplishments would predict.
Commentary offered by Todd Zywicki at the Volokh Conspiracy and by David French at FIRE's Torch focus mostly on the former finding; Zywicki points out that this confirms the findings of two earlier studies by Dan Klein.

Since Zywicki and French (and nobody else I found, though this is a bullet item on Memeorandum and I didn't check them all) didn't concentrate on the information that "ideologically-based discrimination in academic advancement deserve serious consideration and further study." It's the first time, according to the authors, that someone has tried to see if professional advancement is held up through political views. Since that's a serious charge -- more serious to me than just finding out that most my colleagues vote Democrat -- we should examine what they did. All of the following come from the article.

...we examined the correlation between quality of academic affiliation (the dependent variable) and three measures of ideological orientation � left-right self-identification, political party identification, and the ideology index.

...An academic achievement index was constructed from items measuring the number of refereed journal articles, chapters in academic books, books authored or co-authored, service on editorial boards of academic journals, attendance at international meetings of one�s discipline, and proportion of time spent on research.

...There are various emblems of individual success among academics, ranging from monetary compensation to awards to chaired professorships. Perhaps the most
significant single indicator of the academic status hierarchy is the quality of the college or university with which an individual is affiliated. We can construct an institutional quality index by combining the Carnegie Foundation for the Advancement of Teaching classification with the well-known US News & World Report rankings of universities and colleges.

...Both the ideology index and party affiliation, when entered into multiple regression analyses, independently predict the quality of a subject's institutional affiliation. As we would expect, academic achievement matters the most in determining the quality of schools in which faculty teach. But ideology is the second most powerful predictor in Model I (beta=.09, p<.001), accounting for more than one-fifth as much variation in quality of institutional affiliation as does achievement (beta=.39, p<.001). That is, more liberal responses to the attitude questions predict a significantly higher quality of institutional affiliation, after controlling for scholarly achievement.

Second, religiosity is negatively related to quality of institutional affiliation among practicing Christians (beta=-.06, p<.05), but not among Jews. The other variable that is a statistically significant contributor to the equation is gender: Being female is a negative predictor of institutional quality (beta=-.07, p<.01). None of the other potential sources of discrimination for which we have measures is significantly related to the dependent variable. Overall, this regression model explains just under 20% of the variation in the quality of schools in which faculty teach. This analysis confirms the expected impact of achievement on professional status, but it also suggests that ideology plays an independent role. In effect, the ideological orientations of professors are about one-fifth as important as their professional achievements in determining the quality of the school that hires and retains or promotes them. In addition to conservatives, our analysis finds that women and religiously observant Christians are disadvantaged in their placement in the institutional hierarchy, after taking their professional achievements into account.

I could chip away on some items in the study: What constitutes a good institution varies some by field of study. Harvey Mudd College, for example, is a top 20 program in the sciences and engineering, but teaching social sciences there is simply service work and not a particularly desirable post. (I probably am in trouble for that line, since I'm visiting Claremont this weekend.) I'm sure liberal readers will wish to make hay with the result that females are underrepresented in academia as well, but I'm not sure if it's a demand or a supply problem. And that might be true for Republicans as well: Bright Republicans with terminal degrees may be disproportionately drawn to the private sector and away from academia.

I don't consider the results proof positive of much of anything. But as Zywicki notes,

Second, no one has provided any evidence that contradicts the central findings of these studies, whether Klein's or the apparent conclusions of the new study. I'm sure that advocates of the status quo will find something to pick at in the new study as well--but if the findings of these studies are fundamentally flawed, at some point wouldn't someone find something to the contrary? If the evidence was otherwise mixed, then nitpicking at particular studies is one thing, but when the evidence begins to accumulate, at some point it seems like nitpicking is somewhat unresponsive to the underlying issue.

If there is evidence out there that shows a libertarian/conservative takeover of academia, I haven't seen it.

Me neither. Lucky for us, Kurtz notes that the AAUP's representative says "a number of studies show the core values that students bring into the university are not very much altered by being in college." Thank God.