A Brief Note on Intellectual Sloth; or, Are 70% of U.S. Faculty Really Adjuncts?

Call me old-fashioned, but it makes me sad when people rely on shitty data to support their arguments. It makes me even sadder when they minimize, dismiss, or get angry when confronted with the shittiness of the data. These experiences highlight that, for many people, what the data actually are doesn’t really matter. All that matters is that there are some data, somewhere, that they can point to in support of their position.

A sloth studying. From a Facebook page that has not been updated since 2013.

I use the rate of adjunct instructors in U.S. higher education as an example, but this same phenomenon applies to many different topics, so I encourage you to read on even if you are none too interested in the intricacies of U.S. higher education.

Before going any further, I want to be clear about something, which some of you will nevertheless ignore or even downright misrepresent: The reliance on adjunct instructors is bad. It’s bad for students, it’s bad for institutions, and most importantly, it’s bad for the people in those roles. Adjunct instructors are generally not paid well, not provided benefits, not guaranteed employment from term to term, and not accorded much respect by their colleagues or supervisors. This is bad. Other bad things about the U.S. higher education scene include the growing legions of administrators who draw large salaries and the major declines in state funding for public institutions. All very bad¹.

The way we know about these trends is not just intuition or through observing what is happening in our local environments. We know about this stuff because we collect, analyze, and disseminate data on the topics. We do that because we think the data matter. And if we think the data matter, we should also think it matters that they are as accurate as possible.

Yesterday I came across this post on Bluesky:

Source: https://bsky.app/profile/tomrooney.bsky.social/post/3l3uqi2d3ge2x

70% of all U.S. faculty are adjuncts! 70% of all faculty are making close to minimum wage!

Best I could tell, these claims went totally unchecked and were readily accepted. C’mon. This is the kind of number that, when you see it, you should immediately question its accuracy. If you come across a statistic that you consider “mind-blowing,” you may want to at least look into whether or not it is true before sharing it with others.

The source for this figure seems to be from the American Association of University Professors (AAUP). On their website, its states the following:

“Contingent positions that are ineligible for tenure now account for nearly 70 percent of all instructional staff appointments in American higher education, including 49 percent part time and 19 percent full-time non-tenure-track.”

Wait, the statement references “contingent positions that are ineligible for tenure.” It does not say “adjuncts.” What exactly are contingent positions not eligible for tenure?

Well, it says right in the quote that it includes the 19% who are full-time non-tenure track positions. These positions are certainly sometimes low-paid positions with no benefits or guarantees of employment, but not always. Many institutions have “teaching professor” tracks that are well-paid, include benefits, and come with multi-year contracts, so this is a noisy category that is difficult to understand.

You may say I am nitpicking with that one, but the 70% also includes the 20% that are graduate student instructors. Sorry, graduate students are not faculty! We definitely can—and should—discuss how graduate student instructors are underpaid and poorly treated, but we should not include them in faculty statistics².

The more detailed breakdown is readily available on the AAUP website. This is the default figure that is shown on the landing page (other than I switched from headcount to percentage):

Source: AAUP Department of Research and Public Policy, https://data.aaup.org/academic-workforce/

The figure clearly shows that in 2022 the actual number of part-time faculty, what people generally think of when referring to “adjuncts” is about 38%, not 70%. The 70% figure seems to come from adding together part-time, full-time non-tenure track, and graduate students. (Not really the point of the post, but I am struck by how stable these numbers are across 20 years.)

I had wrongly assumed that these data only pertained to instructional staff, but as Dave Vanness pointed out, the graduate student data also include those who have research appointments and likely are not teaching³. Removing them brings the part-time rate up to 42% from 38%. Still not even close to 70%.

These are the kinds of details you can get by clicking on the “drilldown” tab on the AAUP page for the data, and I encourage interested folks to poke around. Importantly, there is a great deal of variability among institutional types, and among schools within those types. For example, the situation is much more dire at community colleges, which tend to be ignored in discussions of higher education, relative to the R1 universities that dominate our attention.

Is 42% still bad? Yes! Is 42% the same as 70%? No! It is simply false to say that 70% of U.S. faculty are adjuncts making nearly minimum wage, and if we are going to share such statistics—especially in such a dramatic way—we should ensure they are correct. Several folks did not take too kindly to me pointing this out on Bluesky, and that brings me back to what is really the main point of this post: there are people out there who want to have data to support their views, but don’t really care if the data are accurate or not. What’s more, questioning the accuracy of the data is equivalent to minimizing the issue.

For me, this is some of the worst behavior of the supposedly scientific, empirically-minded person. What seems to matter is that there are some data out there, somewhere, that they can point to that support their position and allow them to argue that their view is based in evidence. This is the kind of thing that we see quite frequently in research on the negative implications of social media and tech use for well-being. A lot of the data are just bad and don’t provide even modest evidence for the claims, but pointing to some data—any data—gives the arguments extra force and credibility.

If you want to argue for some policy based on moral, ethical, and rational grounds, or because you just think it’s a good idea, then go on ahead. But don’t say that the policy is backed by science^TM just because you have a crapheap of data you can point to.

This whole issue reminds me of Geoffrey Pullum’s classic takedown of the myth of the “Eskimo words for snow⁴.” If you have never read it, do so now. It is a quick, entertaining, insightful read that is well worth your time. I assign it for the first day of my undergraduate research methods class because, as Pullum describes in the Appendix, the essay is fundamentally about intellectual sloth:

The tragedy is not that so many people got the facts wildly wrong; it is that in the mentally lazy and anti-intellectual world we live in today, hardly anyone cares enough to think about trying to determine what the facts are. (Pullum, 1991, p. 171)

1. I’m sure there are many other things that are bad, and I apologize if I left out the issue that you are particularly passionate about.

2. A lot of folks will say that 20% graduate instructors is too high. That may be so, I don’t really know what an optimal number is. But a lot of (the same) folks will also say we don’t provide enough training to prepare graduate students to teach. Doing so does requires that they actually teach, not just take classes in pedagogy.

3. Of course, some have both types of appointments at the same time.

4. I include the following note when referring people to this essay: The term “Eskimo” is still widely used but is generally considered offensive, or at the very least outdated, as it is a label that was developed and applied by people who are outside of the relevant communities. It is preferable to use the labels that are used by the specific communities in question, generally the Inuit and Yup’ik peoples. The term is sometimes still used by Inuit and Yup’ik peoples to describe themselves, but should not be used by outsiders. This is similar to the use of “Indian” in the U.S., which is often used by those are Indigenous or Native American, but should not be used by those who are not. The great Eskimo vocabulary hoax takes on an issue that is related to the cultural homogenization and exotification that is wrapped up in the label “Eskimo.” Thus, despite the outdated terminology, I continue to find it an appropriate reading to assign and ask that you do so while being mindful and critical about the labels we use to describe people, especially those from marginalized backgrounds.