A Critique of the New UW Madison Faculty Survey

A new report by Alex Tahk, director of the Tommy G. Thompson Center on Public Leadership at the University of Wisconsin at Madison, reveals the results of an important survey of UW Madison faculty. But there are serious problems with the survey questions, and we need to be careful not to adopt Tahk’s claims about “ideological imbalance and its consequences” uncritically.

There is a huge problem of response bias in this survey. The survey of 2,400 faculty got 633 replies, which is a good response rate. But that doesn’t remove bias from the pool. There are two groups most likely to not respond to this survey. First, faculty with a wide range of political views who don’t see any problems with bias and feel indifferent to academic freedom may be less likely to respond to a survey explicitly about this topic, undercounting the number of moderates and conservatives while overrepresenting faculty on the far right who feel oppressed. Second, faculty with far-left views who are deeply concerned about academic freedom might boycott a survey from a center created by the Legislature and named for former Republican governor Tommy Thompson because they distrust it as a right-wing trap. As a result, the findings of this survey are likely to be distorted.

Conservatives who feel silenced and have experienced consequences for their speech are likely to be overrepresented because of the indifferent conservatives who ignored the survey, while liberals who have been silenced are likely to be underrepresented because those faculty might boycott the survey due to its source.

These potential biases make it difficult to trust the validity of the results, especially data such as this: “While liberal faculty are more likely to express controversial views, conservative faculty who do express views report experiencing institutional consequences—such as warnings from administrators—at substantially higher rates than liberal faculty who express views.”

That bias in the results is no fault of the center or its director who led the survey, but instead is a consequence of political distrust or indifference about the topic. But it is a bias that should have been acknowledged in the survey results and one that readers should understand when interpreting any of the data. This is not likely to be a representative sample.

Beyond the problems with who responded to the survey, there are numerous conservative biases in the questions that enhance the flaws of this survey and the author’s interpretation of the data.

The study’s author is constantly claiming that UW Madison faculty are excessively liberal and conservatives face bias, such as this key finding: “UW–Madison faculty are substantially more liberal than the general public and Americans with doctoral degrees.” The study shows that about 50 percent of the UW Madison faculty are “liberal” or “extremely liberal,” exactly the same as the general public who hold Ph.D.s or Ed.D.s. UW Madison faculty are slightly more likely to be extremely liberal, slightly liberal or moderate, and less likely to be conservative, but it’s not a dramatic difference.

And it’s easy to explain. Some people have Ph.D.s or Ed.D.s but do not currently work in academia, such as scientists. (The overwhelming majority of new Ph.D.s—77 percent—are in science and technical fields, a much higher number than the allocation of professors at UW Madison.) Many people with Ed.D.s are K–12 administrators. And retired professors have Ph.D.s but are no longer in the academic workplace. All three of these groups are more likely to be conservative than a typical college professor today. UW Madison professors are a substantially different market from the average person with a Ph.D. or Ed.D., and therefore even the small differences found here do not provide proof of a liberal bias in hiring.

Much of the survey relates to highly unusual hypothetical questions about a faculty candidate announcing irrelevant political beliefs during a job interview. But this speculation is not evidence that normal faculty hiring utilizes ideological discrimination or investigates the political views of faculty candidates. Consider the example of immigration. The liberal statement supports “a path to citizenship,” while the conservative statement says to “deport those who are here illegally.” Since students at the university could be Dreamers, some might interpret this as a question about whether they would hire someone who openly announces that they would take action to deport their own students. Not surprisingly, this question resulted in by far the largest gap in treatment between conservative and liberal views.

It’s awful that even a small minority of faculty are willing to say that they might be less likely to hire a candidate for expressing political views they don’t like, and we need to oppose political bias of any kind in hiring. But this survey, with its biased terms and unrepresentative sample, doesn’t allow us to measure how serious this problem is.

In an example cited as evidence of bias against Christians, faculty were asked, “If a job candidate for a faculty position in your academic unit expressed that they would be uncomfortable mentoring students who are practicing Muslims, would this make you more or less likely to support hiring that candidate?”

For half of the surveys, “practicing Muslims” was changed to “Evangelical Christians.” Seventy-five percent of faculty objected to the anti-Christian candidate, compared to 85 percent who opposed the anti-Muslim candidate, which was taken as proof that “not all faculty would treat the type of religious background as irrelevant.” But the survey question was flawed. It should have maintained neutrality by asking about “practicing Christians” alongside “practicing Muslims.” Instead, it used the politicized term “Evangelical,” associated with a radical minority of Christians, which makes it impossible to have any valid comparison of the results. If faculty were asked a similarly loaded term about Muslims and someone biased against “radical Islamists,” it’s possible that there would have been no religious differences in the results.

One of the most interesting findings from this survey is that an overwhelming majority of liberals considered it “very” or “extremely” important to include students and faculty with conservative views—equal to what moderates wanted. The openness of liberal faculty to conservative views should be something to celebrate.

Yet even this finding was twisted by the survey authors into a negative: “Faculty generally rated inclusion of both conservatives and underrepresented racial groups as important, but with a significant gap: 92 percent rated inclusion of underrepresented racial groups as ‘very’ or ‘extremely’ important, compared to 79 percent for conservatives.” But this is a biased comparison of completely different kinds of questions. A math class doesn’t need to include the “views” of conservatives or underrepresented minorities. But it does need to be inclusive of people who are conservatives and minorities.

However, the survey asked about students and faculty “with conservative views” and “from underrepresented racial groups,” making it impossible to compare the responses to these two questions. If the survey had asked about including the “views” of unrepresented minorities, as it did with conservatives, then it might be proper to compare these answers, but any attempt to twist this response as one showing faculty bias against conservatives is simply false.