How Campus Groupthink Distorts Teaching and Research

Recent studies show how ideology can shade all corners of the academy, even the findings of quantitative researchers

If you’re like me, you’ve grown used to looking across America’s colleges and universities and seeing bastions of progressive groupthink, where dissenters fear to speak up and out-of-step scholars carefully navigate in the shadows.

Yawn. You’ve heard this litany many times. You’ve probably also heard it downplayed.

Higher ed mandarins mostly dismiss such concerns, as in last fall’s much-celebrated Chronicle of Higher Education essay, “Viewpoint Diversity Is a MAGA Plot.” They insist, sometimes at great length, that complaints about a left-wing monoculture are nonsense.

Even if there are occasional outbreaks of groupthink, say these apologists, what’s the big deal? Scholars are professionals. Their ideological priors tell you nothing about their disciplinary perspectives or methodological orientations, and they’ve learned to separate their personal views from their academic endeavors.

Well.

While I’m happy to stipulate that partisan affiliation or ideology are not direct proxies for academic perspectives, I think it’s pretty clear they do tend to reflect fundamental (if difficult-to-measure) differences in values, worldviews, and interests.

Given that, does the ideological homogeneity of the professoriate actually matter? For a very long time, the higher ed blob managed the neat trick of convincing everyone it doesn’t (aided mightily by the remarkable lack of inquiry into this inconvenient topic). Of late, however, the evidence that this homogeneity does matter has started to accumulate. It may even infect something as seemingly clinical and “objective” as econometric modeling. More about that in a minute.

Take teaching. Last summer, Claremont McKenna’s Jon Shields and two colleagues dove into the 27 million syllabi collected by the Open Syllabus Project to examine the texts college students are assigned to read on controversial topics like racial bias in the criminal justice system and the ethics of abortion. The most widely assigned texts were celebrated left-leaning works like Michelle Alexander’s The New Jim Crow. But the team was more interested in how frequently such canonical progressive works were paired with “their most important intellectual critics.” The answer? Rarely. Nine times out of ten, no works are assigned that offer a competing perspective. As they explain, “It seems that professors generally insulate their students from the wider intellectual disagreements . . . That is the academic norm, at least in the cases we studied.”

OK, so it happens in teaching. But does such bias also influence research? Well, the Progressive Policy Institute’s Richard Kahlenberg and Lief Lin recently examined every article published in The American Quarterly, the flagship journal of the American Studies Association, between 2022 and 2024. They found “a one-sided and unrelentingly negative portrait of the U.S.” Eighty percent of articles were explicitly critical, luxuriating in America’s “racism, imperialism, classism, sexism, xenophobia, homophobia and transphobia.” They could find not a single positive article. While Kahlenberg and Lin denounced President Trump’s push “to erase the negative components of American history,” they’d hoped that “scholars who study” American “history, literature, and culture” would offer a more “nuanced” take. Not so much. It seems scholarly discipline is no match for gut-level feeling.

All right, so this sort of bias might color qualitative scholarship, but it doesn’t skew “serious” quantitative research, right? On that count, a new study in Science Advances is telling. In January, George Borjas of Harvard University and Nate Breznau of the German Institute for Adult Education published the paper “Ideological Bias in the Production of Research Findings.” They queried 71 teams of researchers (158 scholars total) about their personal views on immigration policy. They then provided the group with the same data set and asked, “Does immigration affect public support for social welfare programs?” The punchline: “Pro-immigration researchers estimated more positive impacts of immigration on public support for social programs, while anti-immigration teams estimated more negative impacts.”

Let’s break this down a bit, because it gets to the core of whether concerns about ideological homogeneity are valid or just, you know, “a MAGA plot.” Borjas and Breznau explain the dilemma:

As the production of research findings does not take place in an observed experimental setting, it is difficult to isolate the potential role of ideological bias. The problem is further confounded because ideological bias can enter the research process in many ways at different stages, including the framing of the hypothesis and the design of the research methodology.

With that in mind, Borjas and Breznau asked whether ideological priors might influence research findings when scholars are interpreting an already-collected data set. It’s pretty easy to imagine how bias might sneak into data collection, survey questions, or a case study, but that didn’t happen here; the data already existed. This is about how ideological priors might creep into something as technical as regression specification.

Borjas and Breznau asked: Did ideology matter when teams of quantitative scholars, most of them sociologists and political scientists, used econometric tools to analyze pre-existing data? This is exactly the case where higher ed impresarios painstakingly explain that scholarly professionalism and good ol’ disciplinary standards overcome any tinge of crude ideological bias.

Well, it sure looks like ideology can influence findings even when researchers believe they are conducting their work carefully and professionally. This can happen, Borjas and Breznau note, because analytic results are always sensitive to tedious “specification” questions like whether to enter immigration into a regression “as a stock or a flow” and whether to use “multilevel modeling to account for variation in country-year units.” (The researchers are nodding along; everyone else is rolling their eyes.)

Researchers must make judgment calls on such questions, and those decisions help shape the teams’ findings—yielding “wildly different estimates” of immigration’s impact. Researchers who have intuitions about the “right” answer can wind up subtly (and even unintentionally) putting a thumb on the scale.

The point is not that right-wing researchers did a better (or worse) job than their left-wing peers in crafting the models but that researchers wound up with models that tended to reflect their prior beliefs. This is why a monolithic culture, left or right, is vulnerable to slant and confirmation bias. And that’s why ideological diversity is healthy for academic truth-seeking.

This post originally appeared on the blog Old School with Rick Hess at Education Next.