If you ask professors about their politics, they’ll say one thing.
But if you use a complex algorithm to predict their politics based on their social-media interactions — as a recent study did — it’ll say another.
By scraping the accounts of more than 4,000 faculty members at over 500 institutions, a forthcoming paper based on the study says that the professoriate’s political persuasions are more diverse than previous survey-based research would suggest. The paper, which will be published in The Review of Higher Education, a peer-reviewed journal, also points to polarization across the political spectrum, arguing that professors’ true beliefs are more extreme and varied than widely thought.
The findings come as many conservative policymakers have sought to rein in a perceived left-leaning bias in academe, often drumming up fear over suspected liberal indoctrination. The authors of the study argue that those claims are exaggerations, and hope the paper will give academics some “firepower” to push back on those characterizations.
While conservative faculty members remain a minority, the study finds far more of them than previous research did, with over 13 percent categorized as strongly right-leaning. A major survey-based study in 2013 found that around 9 percent of professors identified as strongly conservative.
Some of the new study’s other findings, though, back up long-established trends: Tenured faculty members are more conservative than junior scholars are, and there are major ideological disparities between disciplines. Business professors and economists, for example, are often very conservative; the humanities and some sociology subfields, meanwhile, are “decidedly liberal.”
The larger share of conservative scholars found in the study could be a product of its contemporary methodology, rather than actual changes in beliefs, said its lead author, Nicholas Havey, who began the research when he was getting his Ph.D. in computational social science at the University of California at Los Angeles. Havey is now the American Association of Colleges of Nursing’s director of institutional research.
Previous studies of faculty politics employed surveys in which professors categorized or described themselves. This study argues that those results could be inaccurate because of people’s warped perceptions of their own politics or social pressures.
An older professor “might think they’re liberal compared to a goalpost they set 30 years prior, but that old goalpost now makes them far-right conservative,” Havey said. Or many professors may have misleadingly labeled themselves as moderate because “no one wants to think of themselves as inherently radical.” The 2013 study found around half of professors identified as moderate; this one says only about 15 percent actually are.
‘No One Was Watching’
The study, which is based on data collected from 2021 to 2022, attempted to bypass people’s inaccurate self-images with an algorithm that scored their politics by evaluating their likes, posts, followings, and interactions on the website formerly known as Twitter. Professors created that data set “thinking no one was watching,” leading to more accurate results, said Havey, who wrote the study with Xiong Her, a fellow Ph.D. candidate at UCLA.
When talking about the paper at a recent conference and among colleagues, Havey said he enjoyed asking people about their politics, then looking them up in the data. Some ended up “pissed off” when he revealed that “they want to identify as something they have never been and aren’t.”
The method isn’t new. The algorithm was originally developed in other studies that compared its predictions to users’ voter registrations, and found it accurate.
The application to professors creates an “amazing snapshot” of today’s faculty leanings but doesn’t prove there is open debate on campuses, said Alexander Arnold, director of research at the Heterodox Academy, a nonprofit that advocates for viewpoint diversity in academe.
There’s a lot of concern from the right about “the distribution of opinions” in higher education, he said, but the “root problem” is actually whether there are mechanisms and policies in place to allow expression of all views. That’s a matter of college leaders’ creating the right conditions for free inquiry, Arnold said.
Havey considered conducting a similar study of administrators, he said, but the method has become unfeasible. The data in the study were collected before Twitter was purchased by the billionaire Elon Musk and renamed X. Since then, academics’ access to the sort of data needed to perform such studies has been greatly limited. There has also been an exodus of academics from the site.
“Talk about stifling inquiry,” Arnold said of Musk’s effect. “It really is too bad.”
Still, Havey said, the study could be useful in countering a narrative that colleges have become hostile to conservative faculty members, if it ever reaches the intended audience.
“I don’t think the people making a fuss about things are going to care,” he said. “It’s not necessarily the common practice of Republican lawmakers who are fired up about higher education to believe in data.”
