On August 3rd, Remi Adekoya, Tom Simpson and I released a 120-page Policy Exchange report on Academic Freedom in the UK. The report received bipartisan support. On the Left, Ruth Smeeth, an ex-Labour member of Parliament, wrote the foreword and Ruth Kelly, another former Labour MP, backed the report. Lord Sumption, a former Supreme Court judge, and Trevor Phillips, an ex-Equalities czar, rounded out the list of those writing endorsements.
A range of broadsheets from across the political spectrum, from the Telegraph and Times to the Guardian, had good things to say about it, noting the significant level of political bias in academia in which left- and right-wing academics discriminate against each other at relatively similar rates.
The glaring outlier to the positive coverage was the academic activist Left, who laid down a barrage of fire on Twitter in an attempt to divert attention from the unmistakeable story jumping out of the data. New evidence I have collected since replicates precisely the same pattern. Before addressing the critics, however, let’s revisit the findings.
The report was based on a survey of 820 current and retired academics and found as follows:
No-platforming and attempts to “cancel” academics through dismissal campaigns are rising but rare. Their threat to academic freedom comes largely from their chilling effects on academics who might consider challenging orthodoxy. As a result, the few conservative or gender-critical scholars writing on race, gender, and sexuality tend to stay well away from progressive red lines to avoid challenging sacred progressive values.
Political discrimination is a much bigger problem for academic freedom than no-platforming. One in three academics in the survey said they would discriminate against someone who supported the Leave side in the 2016 EU Referendum when hiring for a position, rising to nearly four in 10 among active social sciences and humanities (SSH) scholars. Just 38 percent of academics would feel comfortable sitting next to a gender-critical scholar at lunch.
Having said this, these findings also reveal considerable tolerance: Two in three academics, including a majority of leftists, would not politically discriminate against a Leaver.
Academics don’t discriminate more than other educated professionals, and the Right discriminates as much as the Left, but the fact the Left outnumbers the Right 6:1 (9:1 among current SSH staff) means that conservatives and Leavers experience a far higher discriminatory effect than the left-liberal majority. On a four-person hiring panel, a Leaver faces an 80 percent chance of discrimination.
There is also social pressure. Just half of academics would feel comfortable sitting next to a Leaver, though there is also intra-progressive friction: Only half of centrist Labour, Lib Dem, and Green academics are comfortable sitting next to a supporter of the leftist ex-Labour leader Jeremy Corbyn (by contrast 95 percent of “very left” Labour academics feel comfortable doing so).
Conservative, Leave-supporting, and gender-critical feminist scholars are highly aware of this climate of political discrimination. For instance, a third of conservative academics, rising to a half among conservative SSH academics currently in post, say they self-censor out of concern for their careers. Among leftist academics, the most commonly-stated reason for self-censoring was fear of running afoul of orthodoxy on trans issues.
One in two right-leaning academics say there is a hostile climate for their beliefs in their department (the share is less than half as large among the far Left). Only 18 percent of Leave-supporting academics say a Leaver would be comfortable expressing their Brexit views to colleagues, and most of their Remain colleagues concur: Just three in 10 SSH Remain academics say a Leaver would be comfortable expressing their views to colleagues.
The result of these chilling effects is that conservative and gender-critical researchers conceal their views and write papers which, by staying within the bounds set by progressives, allow them to get published, win grants and be hired. As a US study of the content of legal scholarship showed, registered Republican academics act prudently by concealing their beliefs, researching uncontroversial topics or sticking to technocratic opinions, while progressive scholars openly espouse their beliefs in their work. Cass Sunstein adds that non-career motivations, such as enjoying a cordial relationship with colleagues and a pleasant workplace, also reinforce conformity in organisations. The result, in the social sciences and humanities, is that difficult questions don’t get asked and orthodoxy goes unchallenged, damaging the scientific enterprise.
While we have shown in a previous report that there is little evidence of academics influencing the political views of their students, and while discrimination by lecturers against conservative students is not a serious problem, the prevailing leftward slant of academia makes it less likely that academics will challenge the ideological conformity that currently exists among students in SSH disciplines.
Most academics, including left academics, do not support campaigns to cancel dissenting academics. We tested five hypothetical scenarios involving controversial research touching on questions of empire, family, diversity, and immigration, and found that only about one in 10 academics backed any given dismissal campaign. Overall, most left-wing academics don’t support cancelling controversial scholars. Intolerant leftists really are a minority on campus.
Universities are not in a position to solve the problem on their own. Whereas threats to liberalism typically emerge from the state, there are different scenarios in which the state needs to step in to protect individual liberty—from mobs, illiberal pressure groups, or corrupt institutions. The radical intersectional Left is partly based inside universities, among activist faculty and students. They can call on added support online to coordinate Twitter mobbings, open letters, formal complaints to universities, and protests agitating for academics to be fired or no-platformed. Inside the university, the leftward skew of opinion means that few are willing to speak up against progressive activists.
This allows radicals to leverage the progressive moral code, wielding influence far beyond their numbers. Since chilling effects fall on a minority of less than 10 percent of scholars, most staff and students don’t see a problem. Like anti-black discrimination on southern US campuses in the early 1960s, the pressure is felt by a minority, and the university cannot, or will not, defend them. In these cases, we argued in our paper, the government must intervene to compel universities to protect the liberty of persecuted individuals.
While the full spectrum of media, from the Spectator to the Guardian, acknowledged the importance of our findings, academic Left activists entered denial mode. In fact, no sooner had the ink dried on our report than they moved in to discredit it by changing the conversation from discussing results to zeroing in on methodological minutiae.
In doing so, they faced a difficult problem. We had consciously replicated the questions and methodology of three previous studies (Inbar and Lammers 2012; Honeycutt and Freberg 2017; Peters et al 2020). These used samples consisting mainly of, respectively, American psychologists, American STEM scholars, and European philosophers.
Our methods and findings mirrored those of previous studies, showing that left-wing and right-wing academics discriminate against each other in roughly equal measure, but that the Left greatly outnumbers the Right among the professoriate. Around one in three Leave-supporting academics in Britain face discrimination and about half of conservative scholars said they experienced a hostile climate for their beliefs, replicating results from existing studies.
One question previous scholars raised was whether the number of academics who would discriminate was greater than the share who would openly admit it. To get at this, we used a technique known as a list experiment which found that in Britain, the share who would discriminate (32 percent) was three times larger than the proportion willing to admit it (10 percent). But even for unconcealed questions, we found—in line with previous work—a significant number of academics willing to discriminate against Leavers and the Right.
Unable to contest our core findings, opponents scrambled to create a smokescreen. One activist professor of statistics interpreted a passage in the report to suggest that we claimed our sample was more representative than a political opinion poll, calling this a “howler” and hoping this would discredit the findings. However, we simply drew attention to the fact that our sample was a similarly valid exercise to an opinion poll, something evident from our three-to-four percent margin of error, similar to a standard poll. Furthermore, in the context of our core results, the entire discussion was a distraction.
Another attack concentrated on our sample size of 820, claiming the number of conservatives included in the sample was too small to be statistically significant. Emboldened by these “expert” assertions, activist scholars who lacked statistical expertise jumped on the bandwagon, convincing each other and friendly journalists that the study was flawed, having been “debunked” by their “expert” comrades. What emerged was a self-referential castle built on empirical sand. Indeed, the claim that our sample is too small is flat out nonsense.
On the face of it, 820 may seem like a modest number given the 215,000 academics in the UK. But if you think about it, so is a typical opinion poll of 1,000 for a Britain of 67 million. Funny how none of our adversaries pointed out the “howler” of trying to pretend that a small number in a survey can’t accurately represent a much larger population. They also zeroed in on the relatively small number of conservative academics in the sample, suggesting that their high level of reported hostility and self-censorship arose by chance. By focusing only on conservative academics who reported hostility, they could trumpet a small number of respondents and claim this was too limited to draw inferences. But this, too, is a “howler” because the quantities we compared consisted of the ratio of conservatives reporting hostility versus no hostility with the same ratio among leftists and centrists, a much larger number. Together these utilise most of our 820 respondents making the results statistically significant, with large effect sizes. The equivalent mistake would be to assert, on the basis of a small number of Lib Dem Leavers in an opinion poll, that you can’t claim Lib Dems are less likely than Conservative voters to have voted for Brexit.
Since 40 percent of our sample were retired, with younger and minority staff under-represented, activist scholars argued our survey could not be representative. However, the reality is that younger staff in the survey leaned, if anything, more towards political discrimination than older staff. Including retired academics—whose average age is 70 and who often play a role in their departments after retirement—acted as a conservative brake on our findings. Moreover, controlling for age, there was no difference in the views of current and retired academics. The insinuation that the retired academics in our sample were not true professors and lecturers can’t account for the magical coincidence that they somehow managed to provide the same highly unusual answers as practicing scholars. Perhaps they think YouGov used professional impersonators?
Samples of academics can’t be sourced the usual way because few profs will show up in a random draw from the population. Previous studies have therefore focused on academic associations or mailing lists, achieving a response rate of no more than 26 percent. Their sample size has ranged from 618 to 794 academics. Ours is not only larger but of superior quality because we drew on YouGov’s Profiles panel, the largest of its kind in the world. With over 500,000 people answering surveys, there happen to be more than 1,000 current or former academics on the panel. Nothing comparable is possible in America, for instance. Since these respondents are paid, and don’t choose to answer surveys because of their interest in a subject, they are sampled rather than self-selected. We got 61–76 per cent of the academics on Profiles, making our study less vulnerable than previous research to the charge of being influenced by people with a motivation to take the survey. This gives us a strong claim to representativeness, which is further bolstered by the fact our results line up with those from previous studies.
None of the above prevented our detractors—many of whom have no grounding in statistics—from attacking myself, and our work, as “shoddy.” If you can’t contest the findings, you have to play it like an anti-vaxxer and manipulate impressions or engage in ad hominems. The use of red herrings to create a diversion doesn’t appear to have influenced the wider conversation, but academic activists bought into their own propaganda that our sample was too small or flawed, our methods too “shoddy,” to prove that political discrimination and chilling effects are widespread in British academia. If you repeat a comforting fairy tale often enough it can sound like the truth.
It’s always possible that our results could be in the five percent tail of unrepresentative studies that arise by chance. Perhaps academics aren’t to the left of the population and don’t politically discriminate. Maybe there is no hostile climate for conservatives. The problem is that study after study is telling us otherwise. In fact, there isn’t a single survey going in the other direction. This is bad news for our critics. Since our report was released, I have undertaken a further set of studies, one of UK-based professionals outside academia, another of North American academics. The results, along with those of previous studies, are presented in Table 1.
The Table shows the three prior studies, followed by ours (with separate columns for all academics and current social science/humanities academics), and the two new surveys I’ve fielded. The final column contrasts the academic surveys with a sample of degree-holding non-academics in full-time jobs.
The share of academics who lean left is between 71 and 83 percent across the first six columns, with just 4–16 percent conservative. The share of faculty who openly admit they would discriminate against a right-leaning grant application ranges between 18 and 35 percent, depending in part on variations in the number of response categories. Our UK data are well within this range. In all studies, a majority or near-majority of right-leaning academics experience a hostile climate for their beliefs in their workplace. To claim otherwise, without evidence, is pure fantasy.
Let’s take a closer look at my recent survey of North American social science and humanities (SSH) academics which replicated the Policy Exchange/YouGov questions. As an opt-in survey, it has less of a claim to representativeness, but its findings comport with those of other studies. The concealed method shows that 42 percent of this sample of North American SSH academics would discriminate against a Trump supporter for a job, slightly above the 37 percent of active UK SSH academics that our Policy Exchange study revealed would discriminate against a Leaver for a job. Nineteen percent of the North American sample openly admit they would discriminate against a right-leaning grant application, not far off the 24 per cent of active UK SSH academics that said they would do so.
78 percent of Trump-supporting SSH academics in North America said they experienced a hostile climate for their beliefs, higher than the still-substantial 50 percent for Leave-supporters we uncovered for the UK. Finally, just 15 percent of North American academics said a Trump supporter would be comfortable expressing this belief to a colleague while 88 percent felt a Biden supporter would. This 15–88 ratio compares to a somewhat more modest 30–86 ratio among UK SSH academics for an equivalent question about whether Leavers and Remainers would feel comfortable expressing their beliefs to colleagues. While anti-conservative discrimination and hostility appear somewhat higher in North America, the broad trends reinforce those we find for the UK.
All told, these studies tell a consistent story in which political discrimination strongly impacts conservatives, creating a hostile climate for their beliefs. Imagine denying results like these if five consecutive studies reported that one in three academics would discriminate against a black applicant, or that half of women reported a hostile climate against them in their department.
We can also compare the results to the British non-academic graduate employees in the last column of Table 1, which I sampled using a separate questionnaire on the Prolific survey platform. Most work in private businesses, government, schools and hospitals. The share of non-academic professionals saying a Leaver would be comfortable expressing their views is considerably higher, at 54 percent, than among current academics (30 percent). The proportion of right-wingers in the non-academic study is only slightly larger than in academia due to the young and liberal tilt of Prolific respondents, but, tellingly, just 25 percent of right-leaning professionals reported a hostile climate for their beliefs at work, much lower than in academia, where about half of active right-wing SSH academics say they face hostility.
Nevertheless, these findings concern differences in structural effects, not individual prejudice. While a smaller share of UK non-academic professionals than academics said they would discriminate against a Leaver, this is mainly to do with the smaller share of leftists outside academia rather than academics being more politically prejudiced. As Figure 1 shows, when you compare identical categories of academics and non-academics, levels of discrimination against Leave are similar. For instance, while 58 percent of self-described activists in academia would discriminate against a Leaver for a job, 55 percent of activists outside academia would do so as well.
Source: Adekoya et.al. 2020, N=820; Prolific Academic 2020, N=458 (list only run on half sample).
This gets to our core point that political discrimination is more acceptable in society than other forms of bias. What is distinct about social science and humanities academia is not scholars’ propensity to discriminate, but rather the greater leftward skew and the fact academics’ views are more transparent in their work. This accounts for the anti-conservative and anti-gender critical discrimination that is having such an enormous impact on the academic freedom of these targeted groups.
To be sure, a serious problem of political discrimination also exists outside academia wherever one finds the same combination of ideological homogeneity and belief transparency. This is especially the case in the arts, culture and media sectors. A recent survey of arts employees revealed that eight in 10 acknowledged a chilling effect directed against Leavers and the Right. A recent US survey found that among those with postgraduate degrees, 60 percent of Republicans and almost half of Independents said they were “worried about losing your job or losing out on job opportunities if your political views became known” compared to 25 percent among Democrats. Even for regular degree holders, 40 percent of Republicans and Independents were worried. As in academia, political discrimination leads to widespread self-censorship, a form of John Stuart Mill’s “despotism of custom” that curtails expressive freedom and hampers performance.
Activist scholars want people to focus on their display of molehills at the edge of the terrain rather than the 800-pound gorilla of self-censorship in the middle of it. These activists dominate on social media but don’t actually represent the majority of faculty.
In our Policy Exchange paper, we called for a new Academic Freedom Bill in England that would create a post of Director of Academic Freedom on the Office for Students, the English universities regulator. By providing the Director with ombudsman powers to hear complaints from those whose freedom has been violated and fine universities that do not obey the law, this would provide a counterweight to the powerful activist network inside universities that discriminates against conservatives and gender-critical feminists, curtailing their freedom.
If our experience is any guide, government action will be resisted by this network, who will recycle the well-worn myth that critics are blowing a few rare no-platformings out of proportion; academic freedom is safe; and the whole discussion is a right-wing plot. After all, this is how the world looks from the perspective of people who don’t know any Leave-supporting or gender-critical academics. If you are swimming with the current, how can you understand how it feels to swim against it? Academic activists claim to speak on behalf of their profession but in fact only represent a vocal minority.
Let’s hope the British government has the vision and political will to see beyond these self-appointed representatives to the fair-minded majority of scholars who would welcome better protection for academic freedom and viewpoint diversity on campus.
Eric Kaufmann is professor of politics at Birkbeck College, University of London, and author of Whiteshift: Immigration, Populism and the Future of White Majorities (Penguin/Abrams, 2019). Follow him on Twitter @epkaufm.