“Trigger warnings” are of course cautionary statements given out, usually by college professors, before they present disturbing material to students. The intent is to prevent those who might have been traumatized by that subject or a related one (usually people with PTSD: post-traumatic stress disorder) from being re-traumatized.
My view of such warnings is that material that might be disturbing to everyone should be preceded by a caution (e.g., an ISIS beheading video, other gory stuff or crimes like rape), and that the professor should announce to the class at the beginning of the semester that if anyone has trigger issues, they should come to see the professor and get a list of course material that they might find offensive. But I also think that students should still be responsible for mastering “triggering” material presented in class, that students with trigger issues should be seeing a therapist, and that, in the long run, trigger warnings aren’t helpful in curing the individual of their phobias (psychologists think that one must be exposed to material to get over being triggered by it).
Further, trigger warnings might serve to keep the traumatized student in a status of perpetual victimhood. Finally, the list of stuff that has been deemed potentially triggering is so long that it’s impossible to give warnings in advance about every possible cause of anxiety. Here, for instance, is a list from a sympathetic intersectional website:

One simply can’t warn people about all that stuff in advance!
A new paper, which has been given publicity by lots of right-wing websites, piqued my interest, as it purported to show, according to those sites, that trigger warnings don’t work. Unfortunately, it shows nothing of the sort—only that trigger warnings can temporarily increase one’s sense of vulnerability in non-traumatized people exposed to disturbing prose.
Here’s the paper (click on screenshot to get to article, free pdf is here, and the reference is below). The paper is by Benjamin Bellett, Payton Jones, and Richard McNally, all at Harvard University’s Department of Psychology, and it’s in a reputable journal: Behavior Therapy and Experimental Psychiatry.

The title is cute, but what did the researchers do? They exposed 270 people recruited on the internet to prose passages considered either neutral, mildly distressing, or markedly distressing. These weren’t mostly college students, as the median age of subjects was 37 (see paper for other characteristics). Here’s what the authors say about the passages:
To simulate an academic setting, we chose passages from world literature that commonly appear in high school or college courses. Each passage was standardized in word length, and passage exposures were set to a minimum of 20 s before participants were allowed to continue to the next screen. Transparent attention checks based on the passages’ content assessed whether participants were attentively reading the passages (see supplementary materials S2 for an example of a content check question). We used three types of passages. Neutral passages were devoid of disturbing content (e.g. a character description from Herman Melville’s Moby-Dick). Mildly distressing passages concerned themes of violence, injury, or death, but lacked graphic details (e.g. a description of a battle from James Bradley’s Flags of Our Fathers). Markedly distressing passages contained graphic descriptions of violence, injury, or death (e.g. the murder scene from Fyodor Dostoevsky’s Crime and Punishment). See supplementary materials (S1) for a sample passage from each category.
Subjects with PTSD or who said they had trauma issues were excluded from the study. The authors also surveyed the subjects’ demographics (sex, race, ethnicity, political stance etc.) and attitudes, like whether they believed in advance that words could cause harm and whether they thought trigger warnings should be used (80% said yes in advance). Participants were also asked if they had any history of psychiatric disorders.
They then exposed the subjects to three mildly distressing passages to get a baseline anxiety level. After that, they read ten passages in random order: five were markedly distressing and five were neutral. Half the subjects were given a trigger warning before reading the distressing passages (“The passage you are about to read contains disturbing content and may trigger an anxiety response, especially in those who have a history of trauma”) and half were not. The subjects were then assessed for various psychological variables, and multiple regression analyses were use to tease out the effects of single factors.
Here are the most important results:
- Students who got trigger warnings saw themselves as “more vulnerable to suffering persistent negative emotional effects in the event of experiencing trauma”. That is, they say themselves as having become more easily traumatized. But this effect was small: the increase in level of vulnerability was only 5.2% and the probability that this was true under the null hypothesis of no effect was less than 0.05 but not much lower. That’s not considered a very significant effect, nor is it a large one.
- More students who got trigger warnings believed afterwards that “trauma survivors would suffer persistent emotional effects” than did the controls who didn’t get warnings. But again this effect was small (5.4% increase in strength of that belief), and p<0.05, again not a hugely significant result. They also didn’t correct statistically for multiple comparisons, which would probably make these effects nonsignificant.
These two results are the basis of media reports that trigger warnings were harmful, but of course you can see the problems: not college students, self-report, small and likely nonsignificant effects. Hardly the stuff of headlines. Here are a couple other results:
- Participants who believed in advance that words can cause harm had a significantly higher increase in anxiety from receiving trigger warnings than the un-warned controls. Again the effect is barely significant (p , 0.05), although the result makes sense.
- Factors that made people who got trigger warnings even more likely to see themselves as more vulnerable included being a woman, a member of a racial minority, a liberal, a younger person, and, especially strongly, one with a psychiatric diagnosis that did not include PTSD. These all conform to our expectations or to previous results; the psychiatric diagnosis effect was especially significant (p < 0.001, but not significant for the second test of assessing the vulnerability of other people).
This study, I think, says very little about whether trigger warnings work, for those warnings are used to prevent people with PTSD or diagnosed trauma from being re-traumatized without warning. Those kind of subjects weren’t used in this study. Further, the effects were small, so they don’t even convince me that “trigger warnings don’t work and can even be harmful”. To be fair, the authors list a number of problems with the study at the end of their paper, including the fact that they used reading passages and not visual images or representations. All they can really conclude is that “Trigger warnings do not appear to be conducive to resilience as measured by any of our metrics. . . and may present nuanced threats to selective domains of psychological resilience.”
So this is a start, but what we really want to know is whether trigger warnings are helpful to traumatized individuals, and whether they can contribute to de-traumatizing them over the long run. There’s simply not much data on this issue, though the authors mention a poster—not a published paper—by Bruce (reference below) suggesting that physiological markers of anxiety in traumatized individuals were heightened after presentation of a trigger warning compared to “no warning” controls. That, too, suggests that trigger warnings might not be helpful.
So caveat lector. While the headlines may be comforting to those who don’t like trigger warnings, the data in the paper don’t show that these warnings don’t work—at least when they’re used on traumatized individuals, as is their purpose. And headlines like the one below, from the right-wing website The College Fix, are simply misleading (click on screenshot to read the piece):

________
Bellet, B. W., P. J. Jones, and R. J. McNally. 2018. Trigger warning: Empirical evidence ahead. Journal of Behavior Therapy and Experimental Psychiatry, in press.
Poster said to show negative effects of trigger warnings on traumatized individuals:
M.J. Bruce. 2017. Predictors of trigger warning use: Avoidance or asserting accommodation needs? Poster presented at the annual meeting of the international society of traumatic stress studies, Chicago, IL (2017, November)