Abstract: Six essays in March asserted that “there is no campus free speech crisis,” as Acadia U. political scientist Jeffrey Sachs put it in the Twitter thread that launched the wave of skepticism. In our first post responding to the skeptics, we argued that they went wrong by basing their case primarily on GSS data about the Millennial generation. We explained why the debate hinges not on Millennials but on the generation after them—iGen, or Gen Z—who began replacing Millennials in college in 2013. In this post we draw on five other datasets to show that there are reasons for concern about the speech climate on campus, and there are reasons to think that it is changing since 2015. We address three questions: 1) Is the speech climate (i.e., willingness to speak up) worsening on college campuses, overall, in recent years? We show that it is. 2) Is there a “politically correct” range of viewpoints on campus? We show that there is. 3) Which side of the spectrum is the bigger threat to free speech on campus? We show that students on the left and right used to be similar in their desire to “disinvite” speakers or shout them down, but since 2013 the right has used those tactics much less often while the left has used them much more often. In conclusion, the skeptics are right to demand evidence for claims about change, but wrong to say that there is no such evidence.

Click here for a downloadable and printable PDF of this post.

This is the second post of a multi-part series exploring the issues raised by the skeptics. See also:

  • Part 1 (laying out the debate and giving evidence of changes in averages)
  • Part 3 (on intolerance and threat)


Several essays in recent weeks have asserted that concerns about free speech on college campuses are nothing but a moral panic. The titles speak for themselves:  There Is No Campus Free Speech Crisis (Jeffrey Sachs, in a long Twitter thread); Everything we think about the political correctness debate is wrong (Matt Yglesias, at Vox); The ‘campus free speech crisis’ is a myth. Here are the facts (Sachs, again, at The Monkey Cage). More recently, Aaron Hanlon has weighed in with Are liberal college students creating a free speech crisis? Not according to data (at NBC news); Andrew Hartman argued that People always think students are hostile to speech. They never really are (at The Washington Post); and Mari Uyehara bemoaned The Free Speech Grifters (at GQ). President Trump has joined their ranks, stating that the great majority of college students want free speech, and the problem is limited to a few campuses that get a lot of publicity.

In our first post responding to the skeptics, we showed that the skeptics support their skepticism primarily by relying on data about the Millennial generation (those born 1982-1994). The skeptics are correct that Millennials are not much different than previous generations when asked about free speech issues. We also argued that this debate has nothing to do with Millennials; it is about CURRENT college students, who are not Millennials. By the fall of 2015, most college students (especially at elite four year schools) were members of iGen, the “Internet generation” (sometimes called “Gen Z”),  which begins around birth year 1995, and which first arrived at college around 2013.

We noted that the new attitudes about speech––including the idea that speech can be violence (even when it includes no threat), and corresponding requests for safe spaces and trigger warnings––only began to appear on select campuses around 2013 or 2014, and we noted that these ideas only became widely known after the wave of student protests that began at the tail end of 2015. Therefore, we pointed out, it is unlikely that nationally representative samples, drawing on students in America’s 4,700 institutions of higher education, could have picked up any changes before 2015, when colleges were still full of Millennials who had never heard of trigger warnings and microaggressions. We proposed that the best way to evaluate whether or not things have changed on campus is to examine data collected on current college students in 2016 or later, and compare it to data on current college students from 2014 and before.

When we performed such comparisons, we found some evidence that in fact things are changing. There is not yet much data available to make direct comparisons, but the GSS does show a change for the little bit of iGen data that it has (see figure 1 in post 1), and the larger Knight study showed a change just from 2016 to 2017. In this post we do a much deeper dive. We present far more data on current college students and we assess whether the campus climate has changed in the last few years with regard to speaking up and sharing one’s views.

The key question is this: are students and professors today more reluctant than they were a few years ago to share their views or to question dominant views? If so, then there is a climate or culture problem on campuses where that change has occurred. We note that the overall climate can change rapidly even if there has been no change in average attitudes about speech. All that needs to happen is that a small group of students begins imposing social costs on those who say things they don’t like, while at the same time college administrators do nothing to stop them. (For a fuller explanation, see this essay by Lee Jussim, or this one by Nassim Taleb, whose title explains the key point: The most intolerant wins: The dictatorship of the small minority.) If college students are more likely to report the feeling of “walking on eggshells” in the years after 2015 than they did in the years before 2015, then there has been a change in the campus culture, even if the average student’s support for free speech has not changed.

Indeed, there has been a proliferation of reports about exactly this feeling, from students (see e.g., herehere, and here), and also from professors (see e.g., herehere, and here). It is helpful to have an intuitive sense of what we mean by “climate” or “culture.” We quote from an essay by a student at Smith College, published in September, 2016, titled: Walking on Eggshells- How Political Correctness is Changing the Campus Dynamic

During my first days at Smith, I witnessed countless conversations that consisted of one person telling the other that their opinion was wrong. The word “offensive” was almost always included in the reasoning. Within a few short weeks, members of my freshman class had quickly assimilated to this new way of non-thinking. They could soon detect a politically incorrect view and call the person out on their “mistake.” I began to voice my opinion less often to avoid being berated and judged by a community that claims to represent the free expression of ideas. I learned, along with every other student, to walk on eggshells for fear that I may say something “offensive.” That is the social norm here.

Our challenge in this post is to assess whether this student’s perceptions are just an anecdote––is the problem unique to Smith College? Is the student even correct about Smith College? Or might the problem be happening at enough schools that we can pick up signs of it in national surveys of students? (We know of no dataset that would allow us to address the perceptions of professors.)

We structure this post around three questions:

Question 1: Is the speech climate (i.e., willingness to speak up) worsening on college campuses, overall, in recent years? (Or, are the skeptics correct that “college campuses are more tolerant than ever?“)

Question 2: Is there a “politically correct” range of viewpoints on campus? (Or, are the skeptics correct that “everything we think about the political correctness debate is wrong”).

Question 3:Which side of the spectrum is the bigger threat to free speech on campus? (The skeptics claim that “the majority of successful disinvites came from the right, not the left.”)

QUESTION 1: Is the speech climate worsening on college campuses, overall, in recent years?

To address this question, we first ask whether the students themselves think something is changing. A 2017 survey conducted by Cato/YouGov, on Free Speech and Tolerance (Total N = 2,300, including an oversample of 769 current college students and college graduates), asked: “Do you think that recent student protests and cancellations of controversial speakers on college campuses are isolated incidents, or are they part of a broader pattern of how college students respond to controversial ideas?”

Of the current 4-year college students in the survey, 79% responded that they thought recent campus events are part of a broader pattern of how college students respond to controversial ideas. (The percentage is nearly identical for college graduates, at 81%).

The perception that something is changing is further supported in the Knight Foundation/Gallup survey (N = 3,014 current college students), that asked a direct question about the speech climate on campus: “The climate on my campus prevents some people from saying things they believe because others might find them offensive.” Figure 1 shows that majorities at the four kinds of institutions surveyed agree with that statement, and that the number increased just between 2016 and 2017.

Figure 1. Knight Foundation 2017 data on campus climate. Most students agree that the campus climate sometimes inhibits speech, and agreement is rising.

Data from the Cato/YouGov Free Speech and Tolerance 2017 survey suggests that current 4-year college students, compared to college graduates, hold different views about free speech and expression.  Figure 2 shows that current college students were more likely to agree that “colleges have an obligation to protect students from offensive speech and ideas that could create a difficult learning environment.”  They also were likely to agree that “people who don’t respect others don’t deserve the right to free speech,” and that “supporting someone’s right to say racist things is as bad as holding racist views yourself.”

Figure 2: Cato/YouGov Free Speech and Tolerance 2017 data on opinions about free and hate among current 4-year college students and college graduates. Current students are more in favor of restrictions than are college graduates overall.

To be clear, this is not an “apples to apples” comparison; it is possible that once these iGen students graduate, their attitudes will change and they will look just like the older “college graduates” in Figure 2. But the fact that most of the green bars in Figure 2 are higher than the yellow bars is, on its face, incompatible with the skeptics claim that each generation is getting more and more permissive and tolerant about speech.

Conclusion about Question 1: The majority of current American college students a) perceive that the climate on their campus prevents some people from speaking up, b) perceive this to be part of a “broad pattern of how college students respond to controversial ideas,” and c) perceive the problem to be worse in 2017 than they did in 2016. Additional data from a sample of current 4-year college students suggest that they are more supportive of arguments commonly given for restricting speech, compared to those who have already graduated from college.  It is difficult to see how these findings can be reconciled with the skeptics’ claim that there is no problem and that”college campuses are more tolerant than ever.

Question 2: Is there a “politically correct” range of viewpoints on campus, or are all equally free to speak?

Yglesias claims that “Everything we think about the political correctness debate is wrong.” One of the most widespread things people think about speech on campus is that there is, at many schools, a culture of “political correctness” in which students on the left feel far more free to speak up and share their views than do students on the right, who are more likely to feel that they must self-censor. Is this the case? The evidence we review below points to three substantive conclusions.

A) Students generally think that conservatives are less free to “openly express their views.”

The Knight Foundation survey asked students: “On your college’s campus, do you think members of each of the following groups are, or are not, able to freely and openly express their views?” Two of the groups asked about were “political liberals” and “political conservatives.” Overall, 93% of college students believe that political liberals are free to express their views, whereas only 69% of college students believe that political conservatives are free to express their views. Figure 3 shows that this difference in perception shows up in all groups surveyed. It is not only conservatives who believe this.

Figure 3. Knight Foundation 2017 data on free expression. In all groups, more people think liberals are free to speak than think conservatives are free to speak

B) Conservative students say they are more likely to self-censor in class.

The FIRE/YouGov 2017 data (N = 1,250 current undergraduate students) supports the finding shown in Figure 3. As shown in Figure 4, a little more than half of all students report having stopped themselves from sharing their ideas and opinions in class, but the number is higher for the two conservative bars (average of  61%) than the two liberal bars (average of 53%). The numbers are lower when students are asked about stopping themselves from sharing outside of class but there too, conservatives are more likely to say yes.

Figure 4. FIRE/YouGov 2017 data on sharing ideas and opinions. Students who are “very conservative” are the most likely to say they self censor.

The FIRE/Yougov survey asked a very general question about willingness to share ideas. There are many reasons why one would stop oneself, most having nothing to do with politics or political correctness. (For example, a student might simply be afraid of being wrong, or looking foolish.) The Campus Expression Survey, from Heterodox Academy, asks specifically about class discussions of politically valenced topics and finds much bigger group differences, as you can see in Figure 5. Conservatives are much more reluctant to speak up in class when the discussion topic is related to race, gender, or politics, but there is no difference in reluctance when discussing a non-controversial topic. These data are not from a nationally representative dataset; they reflect the 1,078 current US college students who found their way to  YourMorals.org and completed the survey, so the overall means should not be taken to be accurate reflections of national means. Nonetheless, differences between self-identified liberals (progressives) and conservatives on surveys at the YourMorals site generally mirror differences between liberals and conservatives found in nationally representative datasets.

Figure 5. Heterodox Academy Campus Expression Survey, 2017, on discussing controversial issues in “a small class of 20 to 30 students.” Most conservative students say they would be “somewhat reluctant” or “very reluctant” to share their views in class on politically controversial issues. Most liberals do not say that.

Taken together, Figures 4 and 5 indicate that politically conservative students are more likely to report self-censoring in class, especially about politically charged topics, further validating the common belief among students (shown in Figure 3) that conservatives are less free than liberals to openly express their views on campus.

C) Conservative students self-censor because they fear negative reactions from their professors and from their fellow students.

Reasons given for not speaking up vary as a function of political ideology. FIRE/YouGov’s 2017 survey offered eight possible reasons why students might stop themselves from sharing ideas or opinions in class.  Figure 6 presents six of these eight options; three of them (on the right side of the figure) are about negative responses from the professor. Conservatives were more than twice as likely as liberals to select those reasons. For the two fears about classmates, it’s a mixed bag: very liberal students were most afraid of being judged by their classmates, while very conservative students were most afraid of offending their classmates. [The two items not depicted were “I thought I might be incorrect or mistaken” and, “Other.”  Very liberal (57%) and liberal (66%) students were more likely to select the former item, than very conservative (30%) and conservative students (43%)].

Figure 6. FIRE/YouGov 2017 data on sharing ideas in the classroom. Conservative students are more concerned than liberals about reactions from the professor.

Data from the Heterodox Academy Campus Expression Survey, collected in 2017, supports and extends the findings in Figure 6. The CES asks about discussions of race, gender, and politics more generally. We show the results for discussions of politics, but the pattern was nearly identical for race and gender discussions. Figure 7 shows that, once again, conservative students are more than twice as likely to report concerns about the professor’s reactions to their views, compared to liberal students. Figure 7 also shows a greater concern among conservative students than we saw in Figure 6 about negative reactions from peers.

Figure 7. Heterodox Academy Campus Expression Survey, 2017 data, on concerns about consequences of sharing one’s views on controversial political issues in the classroom. Conservative students are much more concerned than liberal students about negative reactions from professors and from other students.

Conclusion about Question 2:Three surveys provide evidence that there is a “politically correct” range of viewpoints on American college campuses, taken as a whole. We have shown that A) Students think that conservatives are less free to openly express their views; B) conservative students say that they are more likely to self-censor in class, and C) conservatives refrain from speaking up because they–more than progressive students– fear negative reactions from their professors as well as from their fellow students. The data we have presented do not allow us to say whether these problems are worse after 2015 than they were before 2015. However, it is difficult to see how these three findings can be reconciled with the skeptics’ claims that there is no problem on campus and that “everything we think about the political correctness debate is wrong

Question 3:  Which side of the spectrum is the bigger threat to free speech on campus?

In the past year, there have been many articles by academics asserting that the right is a bigger threat to free speech on campus than is the left (see herehere, and here). We agree with the claim made in these articles that many off-campus right-wing groups and individuals are using harassing and intimidating techniques. Heterodox Academy and its members have criticized those groups and techniques as well (see here, here, and here). But does the increasing aggressiveness of off-campus right-wing groups mean that the on-campus left is not a problem, and is not becoming increasingly aggressive too in its efforts to shut down disliked speech? Yglesias, for example, dismisses the charge that liberal college students are more willing these days to shut down speakers. He draws on GSS data to assert that “there does not appear to be any such trend — and … public desire to stymie free expression is concentrated in the working class and targeted primarily at Muslims.”

Hanlon recently wrote that “A Heterodox Academy analysis of the FIRE disinvitation data shows that the most successful attempts to shut down speakers have come from right-leaning groups shutting down speech with which they don’t agree, but this hasn’t stopped pundits and politicians from seeing the student left as the gravest threat to free speech.” Mari Uyehara expands on this claim, writing: “from 2000 to 2017, there were anywhere from six to 35 self-reported disinvitation attempts annually and 40 percent of them came from the right, while Heterodox Academy, an organization devoted to increasing viewpoint diversity, finds that the majority of successful disinvites came from the right, not the left.”

Hanlon and Uyehara’s claims are based on a strained reading of one point in a post written by Stevens. Let us look more closely at the FIRE disinvitation database.

Figure 8 depicts the number of disinvitation attempts by source (on-campus in solid lines, or off-campus in dotted lines) and by ideology relative to the speaker (from the left of the speaker in blue, or from the right of the speaker in red). This dataset includes all disinvitation attempts (successful or unsuccessful) as well as all of the actions that FIRE has classified as substantial event disruptions, defined as: “Heckler’s vetoes,” in which students or faculty persistently disrupt or entirely prevent the speakers’ ability to speak, illustrated by the case of Ray Kelly at Brown University. These incidents are labeled as “substantial event disruption.” The gray line in Figure 8 demarcates roughly when the first members of iGen began to turn 18 and arrive at college.  (Note that 2016 had the unique phenomenon of Milo Yiannopoulos, for whom there were 13 disinvitation attempts, plus three more in 2017. Nobody else was even close; Charles Murray and Ben Shapiro had a total of five each, spread across the two years, and after them, nobody else had more than two. We have therefore removed Yiannopoulos’s disinvitations from the blue line, so that readers can see that this trend is not just a result of his activity. To show where the line would have been had we left him in, we have added asterisks for 2016 and 2017.)

Figure 8. FIRE Disinvitation Database indicates an increase in disinvitation attempts from on-campus and from the left of the speaker since iGen began to turn 18 and enroll in college. (Asterisks for 2016 and 2017 show where the blue line would be if it included Milo Yiannopoulos’s 16 disinvitations.)

First, let us examine the two dotted lines, showing attempts made by off-campus groups. As you can see, there are almost no such attempts from groups to the left of the speaker (meaning: groups offended that the speaker seems too far to the right). The dotted red line, in contrast, shows a few such attempts coming from the right of the speaker; an average of 1.67 per year. These are mostly cases in which a Christian organization or church protested a speaker’s pro-choice stance.

Clearly, the action is in the efforts by on-campus groups–the two solid lines. Figure 8 shows that left and right were equally active up through 2008. It is only in 2009 that a gap opens up, a gap that grew even larger once iGen began to arrive on campus in 2013.

Let us dig deeper. Figure 8 depicts all the events in the FIRE database, however, not all of the on-campus disinvitation attempts involved students.  Some of them were faculty-led campaigns, and others represented administrative decisions. There were also a few instances where students and faculty both engaged in a disinvitation campaign against a speaker.

Since our main interest is in how current college students (and members of iGen more broadly) may differ from previous generations, we went through each of the 370 entries in the FIRE Disinvitation Database and, using the notes provided and the linked to media sources, attempted to identify whether or not students led the disinvitation campaign (either by themselves or in tandem with faculty, other staff members, or alumni).  Of the 370 entries, we were able to identify whether or not students led the disinvitation campaign for 364 of them. In total, students led or worked in tandem on 236 disinvitation campaigns, 194 from the left of the speaker and 42 from the right of the speaker.

Now we can re-cast the data from Figure 8 into a form that allows us to address our central question most directly: Which side of the spectrum is more willing to use illiberal tactics? And has there been any change in student behavior since 2015? We can remove off-campus attempts from the analysis, as they are typically not led by current college students.  Calculating the average number of student-led disinvitation/disruption events per year, for all years before 2015, and comparing them to the two years after 2015 gives us figure 9.

Figure 9. FIRE Disinvitation Database indicates an increase in average disinvitation attempts per year from students who are to the left of the speaker after 2015, compared to before 2015. (Asterisk shows where the blue line would be if it included Milo Yiannopoulos.)

As you can see, the main thing that has changed on campus since 2015, with regard to disinvitations and disruptions, is a roughly 320% increase in disinvitation/disruption efforts by the left, with no change at all from the right.

But what should we make of the claim, made by Hanlon and Uyehara, that the majority of successful attempts came from the right? That claim seems to have been based on one line in our report from early 2017. Here is the quotation from Stevens’ post that they seem to be drawing on:

Speaker disinvitation attempts have a higher success rate when they come from the right of the speaker (54.64%) than when they come from the left of the speaker (32.89%).

Note that this is a statement about the rate, not the total number. It is true that if you just look at the percentage of attempts that are successful in the entire database, the right’s number is higher. But the left has so many more attempts that the totals are approximately the same, with neither side getting a “majority.” The table in the blog post shows a total of 50 successful disinvitations from the left (representing 42% of all successful disinvitations) compared to 53 successful ones from the right (representing 45% of all successful disinvitations; the remaining 8% is accounted for by attempts that are not classifiable by ideology). A skeptic might note that in absolute terms the right is still higher (53 compared to 50), but this ignores the question of trend, which is the crux of the current debate. The left and right used to be roughly comparable in their efforts to shut down speakers.  Since 2015, however, all the action and all the growth has been on the left, as you can see in the next two figures.

When we look only at the cases of successful disinvitations (i.e., removing the unsuccessful disinvitation attempts from Figures 8 and 9, and the attempts that resulted in an event that was successfully disrupted by protesters) we get Figure 10. The picture looks more or less the same as Figure 9 (albeit with lower totals on the Y axis).

Figure 10. FIRE Disinvitation Database, limited to average number per year of successful disinvitations. (Asterisk shows where the blue line would be if it included Milo Yiannopoulos.)

Let’s filter down to the most aggressive and illiberal tactic of all, the “substantial event disruption.” FIRE defines these as: “Heckler’s vetoes,” in which students or faculty persistently disrupt or entirely prevent the speaker’s ability to speak (e.g., attempted shout-downspulling of fire alarms, and blocking entry to the speaking event; see also this example case provided by FIRE).  This now also removes the disinvitation campaigns that succeeded in achieving a cancellation of the speaker’s appearance, and only includes speaking events that were disrupted. These are cases where the speaker traveled to campus, the audience showed up to hear the speaker, and a group of students decided to prevent the audience from hearing the speaker. Figure 11 shows that, once again, the left and right were not so different in the years up through 2014, when successful disruptions were very rare: about one per year from students on the left, and zero per year from students on the right. (There is only one from students on the right in the database, against Chris Hedges, in 2003).  But in the two years after 2015, things are very different. There are now five or six substantial event disruptions per year, all from the left.

Figure 11. FIRE Disinvitation Database. Substantial event disruptions were quite rare before 2015. The average number per year initiated by students to the left of the speaker has increased since then. (Asterisk shows where the blue line would be if it included Milo Yiannopoulos.)

The skeptics may say that “five or six” is a trivial number in a large country with thousands of universities, but the damage that these events can do to property, reputation, and fundraising efforts can be enormous. Each one may have many downstream effects on students, professors and administrators, not just at the school where it happens, but at other schools too. This may be why 70% of college presidents say they are “somewhat concerned” or “very concerned” about violence and student safety when “managing efforts between inclusion and free speech,” as reported in a new survey of college presidents just released this week by the American Council on Education. This fear is consistent with our point about how quickly campus dynamics can change in response to a high profile event, even in the absence of a change in the average attitudes of students.

Figures 8 through 11 plot actual behavior. They show that left and right were not very different before 2015, but since 2015, disinvitations and disruptions have increased greatly, and now come almost exclusively from the left. This change in behavior is consistent with the finding of a partisan difference in attitudes toward shouting down speakers. The 2017 Knight Foundation survey asked students if they thought that a variety of behaviors were always, sometimes, or never acceptable. One of the behaviors was “shouting down speakers or trying to prevent them from talking.” Hardly anyone chose “always acceptable,” so we have merged that response with “sometimes acceptable” and plotted the results in Figure 12. As you can see, large majorities of Republicans and Independents said that shouting down a speaker was “never acceptable.” Democrats, however, were evenly split; only 49% said it was “never acceptable.”

Figure 12. Knight Foundation 2017 data on the acceptability of shouting down a speaker to prevent them from talking.

Conclusion about Question 3: The FIRE disinvitation database shows that students on the left and right were not very different in their tendency to push for speaker disinvitations in the years before 2015, and particularly before 2009. But there is a strong and clear trend in the data: in the last few years, students on the right have largely stopped trying to disinvite speakers, while students on the left have greatly increased such efforts to disinvite, and, failing that, to disrupt. The Knight Foundation survey shows that activists on the left who use these tactics can now count on substantial support from their fellow students on the left, half of whom say it is sometimes acceptable to shout down speakers. These findings are incompatible with Yglesias’ claim that “there does not appear to be any such trend” toward illiberalism on the left, and with Sachs’ claim that “The ‘campus free speech crisis’ is a myth.” These findings also contradict the claims of Hanlon and Uyehara, who drew on a single data point in a Heterodox Academy post to support their claim that students on the right are a bigger threat to free speech on campus than are students on the left. That argument could have been made, with regard to successful disinvitations only, before 2015, but figures 8 through 11 show that things have changed dramatically since 2015.


In On Liberty, John Stuart Mill argued that people get complacent, intellectually, when they are surrounded by those who agree with them: “Both teachers and learners go to sleep at their post, as soon as there is no enemy in the field.” Mill also wrote that

The fatal tendency of mankind to leave off thinking about a thing when it is no longer doubtful, is the cause of half their errors. A contemporary author has well spoken of “the deep slumber of a decided opinion.”

The skeptics were reacting to a wave of commentary among pundits who took it as an established fact that something is changing on American college campuses. The skeptics were right to demand evidence beyond the anecdotes, and they were right to turn to nationally representative surveys for that evidence. We thank them for waking us all up and making us refine our arguments and check our empirical claims. We agree with Yglesias that “The PC debate would benefit from more facts and rigor.”

We have tried in this post to add facts and rigor. Our overall conclusion is that the speech climate is changing on many American college campuses in the years after 2015, compared to the years before. Even if many campuses are unaffected, the changes are large enough to be picked up in nationally representative surveys. We have shown evidence supporting three specific claims:

1) The majority of current American college students perceive that the climate on their campus prevents some people from speaking up, and they perceive this to be more true in 2017 than they did in 2016.

2) Left-wing views are more socially acceptable (“politically correct”) than right wing views, as shown by the findings that students across the spectrum agree that conservatives are less free to speak openly; conservatives say that they self-censor more than progressives; and conservatives are more concerned than progressives about suffering negative consequences from their professors and from fellow students if they were to share their views openly.

3) Before 2009, students on the left and right were not very different in their tendencies to try to disinvite speakers they found offensive. But after 2009 and especially after 2013, the left and right began to diverge; students on the left began to attempt many more disinvitations, to succeed more often in those disinvitations, and to disrupt more of the talks that they failed to stop by disinvitation.

Our analysis is, of course, not definitive. We expect that some of the skeptics will point out potential flaws in our analyses or our logic. Participants on each side of the debate will be influenced by motivated reasoning and the confirmation bias (as we surely were). But, by the magic of viewpoint diversity operating within a community of scholars governed by norms of civility, each side’s motivated exertions will cancel out the confirmation biases of the other, and, over time, we will converge on the truth. Or, at least, we’ll get nearer to it than we are today. (4)

* * * * *

To see why we at Heterodox Academy members care so much about viewpoint diversity, please download your free copy of our new book All Minus One: John Stuart Mill’s Ideas on Free Speech Illustrated. Or watch our three minute video animation:


  1. To read Sachs’ response to this post, at the Niskanen Center, see here.
  2. Why are there no error bars? In this post, as in all of the graphs shown in the pieces by the skeptics, there are no error bars. There are two reasons for this.  First, we do not have access to the original survey data in most cases; we are all drawing on cross-tabs published by the survey outlet.  Second, with the exception of the FIRE disinvitation database, all of the datasets are relatively large, which means almost any difference you can see will be statistically significant.  However, statistical significance matters less than absolute size of the difference.  Many differences are statistically significant yet so small in magnitude as to be unimportant. Like Sachs and the other skeptics, we are interpreting patterns in the data with respect to the patterns we’d expect if various hypotheses were true. None of us would point to a change of three percentage points in responses to a particular survey question and claim it as support, even if it was a statistically significant change.
  3. Our original post noted that we excluded disinvitation efforts against Milo Yiannopoulos, because his “explicit goal” was to provoke disinvitations and disruptions. Some readers and commenters noted that we had no proof that that was his goal, so on April 16 we removed that sentence and updated the graphs in part 3 to show where the blue lines would have been had we included the 16 disinvitation attempts against Yiannopoulos.
  4. This is the second post of a multi-part series exploring the issues raised by the skeptics. See also Part 1 (laying out the debate and giving evidence of changes in averages), and Part 3 (on intolerance and threat)