Content Warnings, also referred to as “Trigger Warnings” or “Content Notes”, have been highly controversial in academic settings over the past decade. However, much of the still influential early commentary on the topic occurred prior to any empirical data that could inform the debate (Ellison, 2016, AAUP, 2014; Lukianoff & Haidt, 2015). Partly because of this initial scarcity, researchers in the past 5 years have taken to the task of investigating how potentially triggering material impacts students and whether warning those students makes any difference. The goal of this paper is to concisely review this work and present new data relevant to those with sexual assault histories.
To date, the research has been diverse in its objectives and goals but can be thought of as trying to achieve three major objectives: 1) How distressed do students get when reading potentially triggering material, 2) Does providing trigger warnings make a difference in how students respond to material, and 3) Does trauma relevant material exacerbate symptoms in those with PTSD? While there is still work to do, the field is starting to move towards some consensus on all accounts.
How distressed are students by triggering material?
In general, students report being somewhat upset when reading or seeing disturbing content (Sanson, Strange, & Gerry, 2019; Kimble, Flack, Koide, Bennion, Brenneman, & Meyersburg, 2021; Kimble, Koide, & Flack, 2022). This effect is not entirely surprising as there is a long history of research indicating that potentially triggering material in the form of movies, text, and pictures will reliably produce PTSD-like symptoms including physiological reactivity (Woodward & Beck 2017; Holmes & Bourne, 2008; Pole, 2006). Sanson and colleagues (2019) have published the most comprehensive recent work in this regard reporting on six experiments which included students and non-students in their samples. While it was not the primary question in their analyses, Sanson et al. (2019) concluded that participants demonstrated more negative affect and PTSD-like intrusions and avoidance when shown “more negative material” as opposed to “less negative material.” In some of their experiments these were fictional passages while in other experiments they were films. Importantly, while students were more distressed by the negative material, this material did not seem to disproportionately affect those with trauma histories. In 2021, Kimble et al. looked at student responses across four institutions to a passage from literature that included sexual and physical assault content. Students reported moderate distress, endorsing subjective units of distress (SUDs) scores between 5 and 6 on an eleven-point SUDs scale. This contrasts to SUDS scores that averaged below 3 two days later. Similar to the work of Sanson et al. (2019), distress was not higher in those with relevant trauma histories. These findings were replicated in a follow up study by Kimble and colleagues (2022) in which undergraduates again reported more distress to a passage containing sexual and physical assault as compared to a control passage. Once again, those with relevant trauma histories did not report more distress. In summary, when presented with potentially triggering material, students will be briefly and moderately distressed, but this seems to happen to all students regardless of whether they have a trauma history or not.
Do Content Warnings work?
There is further growing consensus that this distress is not affected by content warnings, either positively or negatively. At this point, a number of studies have investigated the effects of warning students of potentially upsetting material (Boysen, 2021; Kimble et al. 2022; Bellet, Jones, & McNally, 2018; Bridgland, Green, Oulton, & Takarangi, 2019; Bruce & Roberts 2020; Bridgland, Barnard, & Takarangi, 2022; Jones, Bellet, & McNally 2020), and the findings have been mostly mixed with respect to whether trigger warnings affect how students respond to the material. In many cases no effect was found but Bruce and Roberts (2020) reported that warning may affect comprehension, and Jones et al. (2020) found that warnings increased the centrality of trauma to one’s identity. Bridgland and Takarangi (2021) found that warnings did not affect a number of study outcomes but did result in significantly higher Impact of Event Scale Scores (IES: Weiss, 2007) 2 weeks later. These fairly specific findings bear replication, but there is increasing evidence that students demonstrate more anticipatory anxiety prior to distressing content when they are warned that the material might be disturbing as compared to conditions in which they are not warned (Bellet et al., 2020, Bridgland et al., 2019; Bridgland & Takarangi, 2021). The data suggest that content warnings might not make students respond more (or less) anxiously to the actual material, but they do seem to make them more anxious in anticipation of that material.
One might think that this anticipatory anxiety might lead to widespread avoidance of potentially triggering material, yet there is very little evidence that this is the case. Across a range of studies, participants who are warned do not appear to use that warning to avoid the material, and there is some evidence that it may produce further engagement (Bridgland & Takarangi 2022; Bruce & Roberts 2020; Kimble et al., 2021; Kimble et al., 2022). This is not to say that avoidance is nonexistent, as Kimble and colleagues (2021; 2022) have found that approximately 3–5% of students read an alternative passage rather than a potentially triggering passage, but these rates do not differ based on trauma histories. Bridgland and Takarangi (2022) found that the option to “cover” disturbing pictures only occurred on 3.75% of the trials and did not differ based on whether those pictures had been warned or not.
Is PTSD exacerbated by exposure to triggering stimuli?
While there are some reasons to be concerned about college students in general, content warnings are primarily intended to assist those with relevant trauma histories and ongoing psychological concerns. Historically the concern has been that individuals with PTSD will be dysregulated by exposure to the material and experience an increase in their symptoms. This requires a longitudinal design and Kimble and colleagues (2021; 2022) have completed two studies to date with the ability to assess this. In both studies, the researcher had students read a passage from Toni Morrison’s The Bluest Eye which contained a depiction of both physical and sexual assault. Prior to doing the reading, the students completed the PTSD Checklist for the DSM-5 (PCL-5: Weathers et al., 2013) inquiring about their PTSD symptoms in the past 2 weeks. Two weeks later, the students took the PCL-5 a second time, again inquiring about their symptoms in the past 2 weeks. Neither study found an increase in PTSD symptoms over a 2 week period even in those who would likely qualify for a PTSD diagnosis. Their PCL-5 scores actually went down from Time 1 to Time 2 just as they did for those who did not qualify for a provisional PTSD diagnosis (See Figs. 1 and 2). While those with a provisional PTSD diagnosis reported more distress (as measured by the SUDS) and more PTSD symptoms across all time points in the study, there was no indication that their symptoms were made worse by participating in the study and reading the difficult passage.
The current study
The current study had a number of goals designed to extend this work. One goal was to replicate the previous research but using a passage that was explicitly sexual assault related from Jon Krakauer’s Missoula. This passage would be closer to the experience of a female undergraduate student, particularly in comparison to the previous research using Toni Morrison’s novel set in the 1930s. The Missoula passage is a detailed account of the events before, after, and during a real assault that occurred on a college campus.
Secondly, most previous assessments of trauma in this line of research, including the previous work by our lab, have used the Life Events Checklist (LEC: Weathers et al., 2013). The LEC has 17 items that cover a wide range of possible traumas. Only two of the questions are related to sexual assault or unwanted sexual experiences. To implement an improved assessment of sexual assault, this study used Mary Koss’s Sexual Experiences Survey Revised (SES-SFR: Koss et al., 2007). The SES-SFR uses explicitly consent-based and behavioral language in all items (i.e., “Someone put their penis into my vagina or butt, or someone inserted fingers or objects without my consent”). This has been demonstrated to improve the validity and reliability of sexual assault assessment. Positive endorsements of such items are consistent with most legal and campus-based definitions of sexual assault. The short form of the SES contains questions about seven types of sexual assault: 1) non-consensual non-invasive sexual contact/touching, attempted non-consensual 2) oral, 3) vaginal, and 4) anal assault and completed non-consensual 5) oral, 6) vaginal, and 7) anal assault.
Finally, a signicant limitation to the research to date has been a lack of qualitative responses or analysis at the individual level. The prevalence of a student being significantly triggered by course material, while emotionally significant for that individual, would not necessarily be a common occurrence. Triggering would not occur in those students who had never been traumatized or those who had been traumatized but from a different type of experience (i.e., natural disaster, motor vehicle accident). Even those with a corresponding trauma might not be triggered as they may have never developed PTSD or have successfully engaged in treatment and no longer find such material triggering. For all of the above reasons, the majority of students will never be triggered by difficult class material. However, there could be individuals who are significantly affected by potentially triggering material but whose experiences are missed in research designs that only investigate triggering using group differences in broad samples. The experiences of those few individuals could easily be subsumed in the broader average. They could even potentially be removed as outliers. This study included qualitative responses from the participants about how the reading affected them. Specifically, they were asked to comment on how their score changed over time and why they believed this was the case. While there may be no group differences, there may be a subset of individuals whose scores do increase and they attribute it to the reading.