As an organization, Heterodox Academy is growing increasingly focused on highlighting not just problems, and what is going wrong – but also solutions, and what’s going right, who is doing it well, and how. Consequently, many of us were thrilled reading Jeffrey Sachs’ recent Niskanen essayhighlighting positive trends on campus (although an implication of his argument — that the 2015-7 blowups were tied to broader polarization around the election — is that we may be fast approaching another wave of unrest). I dove into Sachs’ latest Washington Post essay in the same spirit, eager for good news.
Therein, Sachs argues that the self-censorship problem on universities seems to be overblown — and is largely a product of (mis)perceptions among conservative students that they will be punished by their professors and fellow students should they contradict the prevailing consensus. Despite clear evidence that conservative students are deeply concerned about being graded more harshly by their professors for their political views (leading them to self-censor), he asserts there isn’t strong evidence that political discrimination in grading actually occurs.
I think this latter point is important – and as far as I can tell, it is true. Some have (rightfully, to my mind) quibbled with some of the studies he cited, and/ or the way they were deployed – but none of these critiques have really undermined his core claim that there is a dearth of positive empirical evidence that grading discrimination actually does occur. Of course, an absence of evidence is not necessarily evidence of absence. However, it is a fundamental aspect of our legal and moral thinking in Western liberal democracies that the burden of proof lies with the people leveling a charge (such as alleging pervasive political discrimination in grading).
Prima facie, it seems pretty logical to me that political grading discrimination would not occur, even in an environment where there is strong evidence of political discrimination in other quarters, towards other members of the academic community. That will be the topic of a forthcoming essay.
Here, I want to focus on a different claim Sachs made in his piece. Citing a recent survey by the Cato Institute he posits:
“Reports of self-censorship on campus are not much more prevalent than reports of self-censorship among the general public. About 58 [sic] percent of Americans admit they stop themselves from saying what they believe to avoid giving offense…”
Yet the Cato data actually suggest students are (5 percentage-points) more likely than the general public to agree that current political climate prevents them from saying things they believe “because others may find them offensive.” In fact, they are significantly less willing to share their views on virtually any issue than the general public. This effect is particularly pronounced with regards to crime, guns, poverty, inequality, policing, immigration, foreign policy — for which the relative comfort gap approaches or exceeds 20 percentage points:
That is, in the very institutions that are supposed to be places for understanding and formulating responses to social problems, people feel especially uncomfortable discussing pressing social issues. Indeed, they were less comfortable discussing even technocratic and scientific issues like health care and the environment than the general public. Perhaps the most striking, students were less comfortable talking about education than those who are outside the academy. This is a sorry state of affairs — and the Cato data show that most students agree. A majority of students felt that:
Political correctness has stifled conversations we need to have as a society (62%)
People call others racist/ sexist to avoid debate with them rather than because they deserve it (53%)
Colleges are not doing enough to protect free speech (51%) or expose students to a wide variety of viewpoints or perspectives (51%).
Yet the same students who feel more constrained than the general public in their expression, and who feel widespread discontent with the relative lack of freedom on campus, also tend to be far more censorious than the general public:
What is most striking about this data is how uniform the higher support for censorship is. Students are more supportive of laws against disparaging LGBTQ people, immigrants and racial or ethnic minorities on the one hand (views associated with the left) – but also more supportive of making it illegal to disrespect police or military service members, disparage Christianity, or call for violent protest (impulses associated with conservativism).
Literally the only item for which the general public was more supportive of censorship than students related to burning the American flag – but even here, students were slightly more likely to support the draconian measure of stripping citizenship away from flag burners (40% of students supported, as compared 38% of the broader public)!
Beyond legal censorship, students were also more supportive of denying campus platforms to speakers across the board. Once again, they are more supportive than the general public of denying campus platforms to those who disparage immigrants, minorities, and LGBTQ Americans — and also whites, Christians or law enforcement. Disturbingly, the only kind of speaker students are relatively more supportive of granting a platform to are people advocating violent protest (indeed, students are also 10 percentage-points more likely than non-students to say that it is morally acceptable to punch a Nazi).
Here, a critic might object that of course students would be less likely than most to support providing controversial speakers a platform on campus. After all, students are on campus all the time; many students literally live there. Isn’t it reasonable to want to avoid needless disruption in one’s own community? Perhaps the real test might be to see whether the general public would support hosting these speakers in their own communities, and compare that to students’ willingness to host different speakers on campus. Yet even here, students tend to be more supportive of denying platforms on campus to controversial speakers than non-students are to host these speakers in their local communities:
Although, in the abstract, a slight majority (52%) of college students recognize a distinction between allowing a controversial speaker a platform v. tacitly endorsing their views – this distinction quickly collapses when applied to more specific cases. A plurality (49%) of college students believe that supporting someone’s right to say racist things is as bad as being racist oneself – a difference of 7 percentage points as compared to the general public. 44% of college students believe that allowing a racist to speak would be tantamount to that university endorsing racism (14 percentage points higher than non-students).
But as with the other examples, it’s not just racism or left-aligned views that generate this effect. For instance, students were also slightly more likely than the general public to believe that providing a platform to an anti-American speaker would be the same as the college endorsing those views.
In the same paragraph where Sachs cited the CATO report, he highlighted that private-sector employers also restrict speech. He is right to point that out, and it is another important point that could use more attention in the media than it gets (especially relative to campus blowups). But disturbingly, students are also significantly more supportive than most of employers punishing workers for speech, even for speech that occurs outside the workplace (and of course, policing speech entails surveilling it). For instance, 64% of students support employers sanctioning employees for things they say on social media, as compared to 44% of the general public.
We can see, from the very data Sachs cited, that students seem to feel substantially less free to express themselves than general public. Yet while they chafe under these perceived constraints and express frustration about them, students are also tend to be significantly more censorious than the general public.
How can we square these findings with the datapoint Sachs cites from FIRE that the vast majority (87%) of students feel generally comfortable expressing themselves on campus? Simple: just as people are more likely to support free speech in abstract than with regards to specific controversial issues, or specific groups they dislike – students are also more likely to declare comfort expressing themselves in general than with respect to particular issues. As we saw in the Cato data, this holds with virtually any specific issue.
Indeed, issue by issue, across the board, students felt substantially less comfortable expressing themselves than the general public. They were also far more likely to support censorship on a range of issues – irrespective of the political valence of those issues. This suggests that while there doesseem to be an element of political bias to censorship (margins tended to be much larger for issues with a left valence) — it seems like students are far less comfortable with controversy of any kind than the general public, and more likely than the general public to try to suppress controversy of virtually any kind through censorship of self and others.
It is unclear whether these dynamics are a product of selection effects (i.e. more censorious people just happen to be more likely to go to college), cohort effects (i.e. contemporary students are more censorious than previous generations – my take on this question here, here), or whether they are a product of the education process itself (i.e. education renders people more likely to support censorship than they otherwise would have been – my take on that question here).
Regardless, rather than serving as bastions for free exchange of ideas and the exploration of controversial topics, the state of discourse on campus seems substantially less free than that of the broader society. At least, that’s what the Cato data suggest.
I’ll note in closing that Heterodox Academy is very committed both to measuring this phenomenon in more detail, and to providing others with tools to do the same. Towards this end, we developed our Campus Expression Survey (CES). We have fielded a couple nationally-representative samples, and will be sharing some of that data soon. But in the meantime, if you are a faculty member or administrator hoping to understand the dynamics within your own classroom or institution, instructions for how to deploy the CES are available here (we also have a version of the CES for high-school students).