Are trigger warnings – or “content notes” – a pedagogical technique for academics to use as they see fit, on the basis of professional intuition and tacit judgement? Or, where they are being used, should academics be able to identify the rationale for their use?
Earlier this year, CAF launched a systematic Freedom of Information (FOI) inquiry across the Russell Group to examine how trigger warnings are used in teaching – and, crucially, what evidential basis, if any, institutions rely on when doing so.
The context was the commencement, on 1 August 2025, of the Higher Education (Freedom of Speech) Act 2023 (HEFSA). As operationalised through the Office for Students’ Regulatory Advice 24 (RA24), certain decisions relating to teaching and pedagogical practice may now have to be considered through the lens of the statutory duty to secure freedom of speech within the law. Where such decisions may restrict lawful speech, RA24 expects universities to ask whether “reasonably practicable steps” can be taken to secure that speech; and, where restriction is contemplated, whether that restriction is legally authorised, proportionate and no more intrusive than necessary.
On trigger warnings, the OfS adopts a cautious position. Rather than banning them outright, it warns that any standing requirement for their use may deter academics from exposing students to controversial material. The guidance therefore acknowledges that although “there may be occasions” when trigger warnings may be advisable, this is only where they will “be helpful to enable students to access material, if there is evidence that they are in fact helpful”.
This is, admittedly, a somewhat unusual way of framing the issue. Trigger warnings are still commonly defended in therapeutic language as devices intended to help students with histories of trauma prepare for material that might provoke severe psychological distress. Yet their use often drifts far beyond that rationale, functioning at best as a warning that something may be vaguely difficult to read and, at worst, as a vehicle for lecturers to virtue signal that a text contains attitudes, assumptions or representations that they regard as morally or politically problematic.
This concept creep is not hard to spot. How else to explain the University of the Highlands and Islands attaching a trigger warning to Ernest Hemingway’s The Old Man and the Sea because it contains “graphic fishing scenes”; Chester University warning students that J.K. Rowling’s first Harry Potter novel may prompt “difficult conversations” about gender, race, sexuality, class and identity; Glasgow flagging the same book for its “outdated attitudes”; the University of Sheffield attaching trigger warnings to biblical accounts of the crucifixion; or the University of Nottingham warning that Chaucer’s Canterbury Tales contain “expressions of Christian faith”?
Viewed through the lens of a statutory free speech regime, the OfS’s formulation therefore becomes easier to understand. Trigger warnings are being held to the standard by which they were originally justified – i.e., do they actually help students engage with difficult material? – as opposed to the tacit logic by which they increasingly appear to be adopted, subtly encouraging students to swerve engagement altogether or, worse, acting as a hint as to the interpretation a lecturer expects to see reflected in subsequently assessed work.
In this sense, RA24 gives regulatory expression to a debate already underway in the academic literature. While some scholars continue to defend trigger warnings as useful pedagogical tools, a growing body of empirical research suggests they are frequently ineffective, may increase anticipatory anxiety and can operate less as neutral safeguards than as interpretive cues, curtailing the analytic possibilities students might otherwise have explored in and as part of their academic studies.
So if universities are indeed using, recommending or requiring them, it seems eminently reasonable to ask what evidence they rely on. It is that question that CAF’s FOI investigation set out to explore.
Across the Russell Group, 22 of the 24 universities responded to our first round of FOI requests. Of those respondents, approximately two-thirds (68 per cent) acknowledged that trigger warnings are used in teaching, confirming that this is not a fringe practice.
But what evidence supports their use?
Nearly half of respondents (46 per cent) have issued central or faculty-level guidance or recommendations that encourage or normalise trigger warnings. These documents typically assume that content notes are beneficial and focus not on whether they work, but on implementation: when warnings should be used, how they should be worded, where they should appear, and who should be responsible for providing them.
Yet a striking number of universities – including some that had issued such guidance – simultaneously told us that they held no recorded evidence of effectiveness. Among them were Glasgow, Nottingham, Newcastle, Queen’s University Belfast, Warwick, UCL and Bristol. So despite normalising their use as a legitimate pedagogical practice, these universities were unable to identify any recorded evidence showing why – or if – trigger warnings help students access academic material.
Some institutions (18 per cent) have even publicly defended content notes after cases involving modules run by their Schools and departments appeared in the press, describing them as a “standard academic tool” or claiming that particular approaches have been “shown to be useful”, while telling us under FOI that they held no recorded evidence supporting those claims.
Beyond guidance and governance materials, none of the responding universities that acknowledged using trigger warnings provided evaluative evidence demonstrating that they help students access academic material. There were no internal evaluations, pilot studies, effectiveness assessments, outcome measures, evidence reviews or external studies relied upon as a basis for adoption.
A small minority – around 9 per cent – pointed to materials they appeared to treat as evidence. But as with documentation disclosed under the heading of “guidance”, this consisted of student consultations, workshop exercises, interview material and implementation resources that always began from the premise that trigger warnings are useful.
Methodologically, these materials may tell us something about how trigger warnings are perceived. They are not, however, the same thing as evidence that trigger warnings improve learning, comprehension or access to course material.
Exeter’s first-round disclosure is a case in point, directing us towards a student-led “creative and collaborative” project, later written up in Times Higher Education, that assumed content warnings are needed in education and focused chiefly on how they should be implemented.
At a “Scrabble centred creative station”, participants highlighted the need to use “phrases such as ‘everyone’ and ‘compassion’”, while “emphasising the need for inclusive language”. One recommendation was the use of the term “content warnings” rather than “trigger warnings” in order “to avoid assumptions and violent connotations and help staff foster an inclusive environment”.
A “sticky note forum in which people could leave their thoughts and feelings on why content warnings mattered to them” produced a further recommendation: “Provide warnings in advance, allowing individuals to mentally prepare and make informed decisions”.
Another source disclosed by Exeter was an institutionally produced podcast, one of the trigger-warning governance “takeaways” of which was: “Implement content warnings based on clear institutional policies rather than leaving it to individual lecturers’ discretion.”
Alongside this, Exeter disclosed Faculty of Humanities, Arts and Social Sciences guidance telling staff that “it is advisable to include a ‘content notice’ when content that may be deemed sensitive is being taught and discussed in class, or used in teaching and learning materials”, before listing the following topics:
“Race / ethnicity; gender; sexuality; sex; sexual relations; nudity; domestic violence; abortion; sexual abuse; drugs; drug abuse; alcoholism; ableism; disability; religious belief; political belief; warfare; Empire and Imperialism; slavery; violence and criminality; race relations; racial conflict; youth; old age; class and social background; poverty; deprivation; income inequality; homelessness; unemployment; etc.”
“Etc”? After a list so expansive, one is left wondering what meaningful boundary remains between material that requires a trigger warning and material that does not.
When asked in a follow-up FOI whether it held any formal evaluation showing that content warnings help students access teaching material, Exeter answered: “No – there hasn’t been any formal evaluation.” Absent such an evaluation, the Faculty’s rationale for concluding that it was “advisable” to include trigger warnings when teaching or discussing such a wide range of topics is not immediately obvious.
That is precisely why CAF has now sent follow-up FOIs to every institution where trigger warnings were acknowledged to be in use but no evidential basis was identified. The aim is to establish whether evidence exists below the level of central university policy, a possibility left open by the necessarily high-level nature of our first-round requests. We therefore directed our follow-up requests to the places where teaching decisions are most likely to be made: departments, Schools and degree programmes.
Although a number of responses are still outstanding, Bristol is illustrative of the early returns. In its first-round response, the University acknowledged that trigger warnings “are used at the discretion of unit leaders”, explaining that “teaching staff make informed decisions based on their knowledge and experience”.
This is, in essence, an appeal to what has traditionally been known as academic judgement: the idea that certain decisions about teaching, course content and disciplinary method properly belong to academic experts rather than external reviewers, regulators or the judiciary.
We therefore went back to Bristol and focused on two departments – English and History – where media reports indicated that trigger warnings had been used, and asked three questions: did either department hold guidance relating to trigger warnings; did either hold evidence as to their efficacy; and was there any recorded indication that limitations, contested effectiveness or potential unintended effects had been considered?
In both departments, on every point, the answer was “no”.
To be sure, that does not establish that trigger warnings are ineffective, nor that the academics using them are acting unreasonably. What it does suggest is that, in these departments, the University’s earlier reference to the “knowledge and experience” of its academic staff may not correspond to any recorded evidential basis.
All of which brings us back to RA24’s suggestion that trigger warnings may be justified where there is evidence that they are “in fact helpful” in enabling students to access material.
In the Bristol departments we tested, the rationale for using trigger warnings may exist as a matter of professional judgement or tacit knowledge. But if so, it leaves no recorded evidential trace. Put simply, trigger warnings are being used, but no-one can tell us why.
CAF is continuing to analyse second-round FOI responses from across the Russell Group. We will publish a fuller account once those responses have been returned.
Support CAF’s research, advocacy and casework defending academic freedom and free speech rights for staff and students across UK higher education.