Three weeks ago I wrote about a recent survey of attitudes on college campuses regarding free speech. I and many other bloggers used the new data as an opportunity to make a few skeptical points.

First, the data does not support the popular narrative that there is a free speech “crisis” on college campuses. The long term trends show that support for free speech is increasing, and that college education and being liberal both correlate with more support for free speech. These trends directly contradict the standard narrative that liberal college professors have “run amok” with their political correctness.

In response Sean Stevens and Jonathan Haidt wrote a couple of articles arguing that the skeptics were wrong on this issue. To be as fair as possible, I do think they have one small point to make, but overall I think they are tilting at a straw man of their own making. I also think they are making the exact kind of errors of biased interpretation that they are accusing the skeptics of making.

The legitimate point they make is that while the long term trends are positive toward free speech, recent data suggests that the current generation (iGen) entering college may be reversing that trend. At least, we should consider this recent data in formulating any opinions about the current state of affairs.

However, their attempt to use this data to support the narrative that there is a free speech crisis brewing on college campuses is not compelling, and fraught with biased or at least narrow interpretations of the data.

Their major straw man is this – what I was saying (and I think most of the other writers were saying) is that the narrative of a free speech “crisis” is not supported by the data. It is just that – a narrative, used for political purposes. The data tells a far more complex and nuanced picture, and the major trends don’t support the narrative.

Saying that there isn’t a “crisis” does not translate into – there is absolutely no concern, or no data showing anything worrisome. I am always concerned about the topic of free speech, and feel that overall I would like to see more  informed support for this fundamental right. But we are not at crisis levels, and there does not appear to be any liberal conspiracy to attack free speech or the First Amendment.

Let’s take another look at the data, with the idea that survey data are always complicated to interpret. At best we can make inferences from survey data, which (as I pointed out in my original article) are fraught with potential confounding factors. I pointed out possible confounds, but Stevens and Haidt I felt tended to ignore or gloss over confounding factors and argued as if the surveys they were relying on were direct measures of support for free speech.

They rely heavily on a 2017 survey by the Cato institute. When searching on the topic, most of the top results were references to this article, and supporting the “free speech crisis” narrative. I hasten to point out that the Cato institute is a libertarian think tank, with a clear ideological dog in the race. But even putting that aside, even their data does not support the narrative.

Stevens and Haidt point out:

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.”

First, this is a bit of cherry picking on their part. When I looked at that survey I was struck by the fact that the authors were not making a big case for a difference between Republicans and Democrats – because the data don’t support that conclusion. For example, the survey found that:

51% of strong liberals say it’s “morally acceptable” to punch Nazis.
• 53% of Republicans favor stripping U.S. citizenship from people who burn the American
flag.
• 51% of Democrats support a law that requires Americans use transgender people’s
preferred gender pronouns.
• 47% of Republicans favor bans on building new mosques.
• 58% of Democrats say employers should punish employees for offensive Facebook posts.
• 65% of Republicans say NFL players should be fired if they refuse to stand for the national anthem.

You will notice that the numbers are not different between the two political parties overall, and the group that took that most anti-free speech position were Republicans regarding NFL players. Throughout the survey, this overall pattern is very strong. Support for the alleged free speech position hovers around 50%, and varies with the topic and political party, with the two parties being roughly equal (although, if anything, hot button issues for Republicans tend to have the highest number taking the alleged anti-free speech position).

There was enough data in this survey, however, to support whatever narrative you had. But take a look at the numbers for yourself, try to look at them from an unbiased perspective. If you knew nothing about American politics, could you look at this data and say which political party more strongly supports free speech? I don’t think so.

I also said “allegedly” above to make my second point – it is very difficult to interpret surveys like this. People may answer questions about – “would you support a speaker with belief X to speak at your college” from different perspectives. They may simply interpret the question as asking about their support for that position. I am against racism, therefore I should not support a racist speaker. They may not be considering the free speech issue. You could further argue, that is part of the point, but we don’t really know how that individual balances freedom of speech and other priorities like respect and tolerance from this one question.

Also, and perhaps more importantly, you can be for free speech but still think that colleges have a responsibility not to lend their name and credibility to speakers who are intellectually bankrupt. In other words – opposition to certain speakers may be about the academic quality of those speakers, and not the offensiveness of their positions. The Cato Institute survey, however, assumed they were markers for tolerance of offensive speech only.

I, for example, would not support a creationist being invited to give a speech at a university so that they could spread pseudoscience, and then pad their resume with the name of that university, and use their speaker fee to further their career trashing science.

In other words – there is a quality control issue here that appears to be entirely ignored by the Cato survey.

Let’s take a look at the recent trends that Stevens and Haidt think are so important to this debate. First, even if I grant the data they are using, that current college students on surveys meant to measure their support for free speech are showing a reversal in the longer trend toward increasing support for free speech, there are many problems that they ignore. This data, by necessity, is preliminary and short term. We will need to follow the data for at least several more years before we can declare a trend.

There are many factors that can be causing short term trends in the long term data. This is like pointing to a couple years of cooling as if it contradicts the long term trend of warming.

But perhaps even more important, there are so many possible confounding factors here. To name just a few, it is possible that what we are seeing is increased polarization on both ends, not a trend on the liberal side. There is independent data to support the conclusion that politics have become more polarized overall.

There are also lots of trends from social media, which has disrupted our normal political discourse in this country. Are we seeing an echochamber effect, for example.

Also, one might reasonably argue that the end of the political spectrum that has become more radicalized in recent years is the right. Especially very recently, I would not be the first to observe that some people on the right feel more free to express racist, nationalist, homophobic, and generally intolerant views. We may simply be seeing in the current college generation a reaction to this backlash against the long term trend of globalization and pluralism.

In fact, speakers on the right have deliberately tried to get themselves invited to college campuses in order to give radical speeches, in order to provoke a reaction from the left. They are essentially creating the narrative they wish to complain about.

I also think that Stevens and Haidt are biased in their interpretation of the data. They point out that the average number of successful disinvitations of speakers in the US from the liberal side (liberals disinviting a conservative speaker) has increased – from 2 to 7 per year. That is the total per year for all the colleges in the US. Responding to the obvious point they write:

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.

That is a lot of hand-waving. Again, a few extra events where a speaker was disinvited could easily be entirely an artifact of the strategy used on the right to provoke getting disinvited as a publicity stunt. And quoting a survey saying college presidents are concerned about safety (of course they are) doesn’t do much to support their point.

If anything this data supports the conclusion that there is no crisis. We are talking about a few events that are blown out of proportion, to create a false impression that such events are widespread. There are 5,300 colleges in the US, so 7 disinvitations is 0.13%. Sorry, that hardly seems like a crisis.

The bottom line is that the data is still very different than what most people with the “free speech crisis” narrative would naively believe. The overall trends are still good, the actual events that are grabbing the headlines are rare and often engineered as theater, and this is not a liberal phenomenon (even when looking at data collected by a libertarian think tank).

What I think is happening is that recent social changes brought about by social media, by increased awareness of continued sexism and racism in our society, increased polarization, and of populist pushback against the long term trends of pluralism, tolerance, and globalism are all having an effect. This is prompting a very healthy discussion about the balancing of freedom of speech, tolerance, respect, and academic quality.

In this discussion freedom of speech still has wide and overall growing support. I hope we will come out the other end with even more and deeper support for free speech, and I think we will. There will be fires, there will be extremists, and I will be there to help put them out and defend what I think is the core freedom of our democracy.

But I still find hysteria about a free speech “crisis” to be nothing but political theater.