Trigger warning: Empirical evidence ahead

Article (PDF Available)inJournal of Behavior Therapy and Experimental Psychiatry 61 · July 2018with 757 Reads
DOI: 10.1016/j.jbtep.2018.07.002
Abstract
Background and objectives: Trigger warnings notify people of the distress that written, audiovisual, or other material may evoke, and were initially used to provide for the needs of those with posttraumatic stress disorder (PTSD). Since their inception, trigger warnings have become more widely applied throughout contemporary culture, sparking intense controversy in academia and beyond. Some argue that they empower vulnerable individuals by allowing them to psychologically prepare for or avoid disturbing content, whereas others argue that such warnings undermine resilience to stress and increase vulnerability to psychopathology while constraining academic freedom. The objective of our experiment was to investigate the psychological effects of issuing trigger warnings. Methods: We randomly assigned online participants to receive (n = 133) or not receive (n = 137) trigger warnings prior to reading literary passages that varied in potentially disturbing content. Results: Participants in the trigger warning group believed themselves and people in general to be more emotionally vulnerable if they were to experience trauma. Participants receiving warnings reported greater anxiety in response to reading potentially distressing passages, but only if they believed that words can cause harm. Warnings did not affect participants' implicit self-identification as vulnerable, or subsequent anxiety response to less distressing content. Limitations: The sample included only non-traumatized participants; the observed effects may differ for a traumatized population. Conclusions: Trigger warnings may inadvertently undermine some aspects of emotional resilience. Further research is needed on the generalizability of our findings, especially to collegiate populations and to those with trauma histories.
Figures - uploaded by Benjamin Bellet
Author content
All content in this area was uploaded by Benjamin Bellet
Accepted Manuscript
Trigger warning: Empirical evidence ahead
Benjamin W. Bellet, Payton J. Jones, Richard J. McNally
PII: S0005-7916(18)30113-7
DOI: 10.1016/j.jbtep.2018.07.002
Reference: BTEP 1405
To appear in: Journal of Behavior Therapy and Experimental Psychiatry
Received Date: 25 April 2018
Revised Date: 20 July 2018
Accepted Date: 26 July 2018
Please cite this article as: Bellet, B.W., Jones, P.J., McNally, R.J., Trigger warning: Empirical
evidence ahead, Journal of Behavior Therapy and Experimental Psychiatry (2018), doi: 10.1016/
j.jbtep.2018.07.002.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
Running Head: TRIGGER WARNINGS 1
Trigger Warning: Empirical Evidence Ahead
Benjamin W. Bellet
1
, Payton J. Jones
1
, and Richard J. McNally
1
1
Harvard University, Cambridge, MA
Author Note
Benjamin W. Bellet, Department of Psychology, Harvard University; Payton J. Jones,
Department of Psychology, Harvard University; Richard J. McNally, Department of Psychology,
Harvard University.
Correspondence concerning this article should be addressed to Benjamin W. Bellet,
Department of Psychology, Harvard University, 33 Kirkland St., Cambridge, MA 02138. E-mail:
bbellet@g.harvard.edu.
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS 2
Abstract
Background and Objectives: Trigger warnings notify people of the distress that written,
audiovisual, or other material may evoke, and were initially used to provide for the needs of
those with posttraumatic stress disorder (PTSD). Since their inception, trigger warnings have
become more widely applied throughout contemporary culture, sparking intense controversy in
academia and beyond. Some argue that they empower vulnerable individuals by allowing them
to psychologically prepare for or avoid disturbing content, whereas others argue that such
warnings undermine resilience to stress and increase vulnerability to psychopathology while
constraining academic freedom. The objective of our experiment was to investigate the
psychological effects of issuing trigger warnings.
Methods: We randomly assigned online participants to receive (n = 133) or not receive (n = 137)
trigger warnings prior to reading literary passages that varied in potentially disturbing content.
Results: Participants in the trigger warning group believed themselves and people in general to
be more emotionally vulnerable if they were to experience trauma. Participants receiving
warnings reported greater anxiety in response to reading potentially distressing passages, but
only if they believed that words can cause harm. Warnings did not affect participants’ implicit
self-identification as vulnerable, or subsequent anxiety response to less distressing content.
Limitations: The sample included only non-traumatized participants; the observed effects may
differ for a traumatized population.
Conclusions: Trigger warnings may inadvertently undermine some aspects of emotional
resilience. Further research is needed on the generalizability of our findings, especially to
collegiate populations and to those with trauma histories.
Keywords: trigger warning, anxiety, PTSD, resilience, vulnerability
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS 3
Trigger Warning: Empirical Evidence Ahead
Is it better to warn people about potentially distressing material, or allow them to deal
with it on their own terms? Trigger warnings and other protective measures implemented at
institutions of higher learning, such as safe spaces and the dis-invitation of potentially offensive
speakers, have become the subject of contentious, widespread debate (Wilson, 2015). In the
classroom, a trigger warning is the practice of “teachers offering prior notification of an
educational topic so that students may prepare for or avoid distress that is automatically evoked
by that topic due to clinical mental health problems” (Boysen, 2017, p. 164). Much support for
trigger warnings arises from the desire to provide students with posttraumatic stress disorder
(PTSD) and other disadvantaged groups with an inclusive, level academic playing field (Carter,
2015; Stokes, 2015). However, others believe that trigger warnings hamper free academic
inquiry and “coddle” students by sheltering them from any stressful material they may encounter
(Lukianoff & Haidt, 2015), thereby undermining their preparation for the “real world” beyond
the campus gates.
The use of trigger warnings is supported by evidence that individuals with PTSD can
experience painful recollections of trauma in response to reminders of their experience
(American Psychiatric Association, 2013); trigger warnings may help those with PTSD to choose
the time and place of their exposure to reminders, or psychologically brace for them (Boysen,
2017). However, trigger warnings may encourage avoidance of cues related to trauma (McNally,
2014). Avoidance runs counter to the aims of prolonged exposure (PE) therapy, the most
efficacious treatment for PTSD (Institute of Medicine, 2008). PE encourages systematic
exposure to triggers, enabling patients to habituate to them and regain functioning. Conversely,
avoidance of triggers may diminish distress in the short term, but worsens symptom severity in
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS 4
the long term (Rosenthal, Hall, Palm, Batten & Follette, 2005). Further, receiving trigger
warnings about trauma-related cues may enhance the centrality of traumatic events to survivors’
identities (McNally, 2014), reminding them to view material through the lens of trauma.
Regarding trauma as central to one’s identity is associated with severity of PTSD symptoms
(Berntsen & Rubin, 2007; Boelen, 2012; Robinaugh & McNally, 2011). Clearly, the question of
whether trigger warnings help or harm trauma survivors has been the subject of much spirited
debate, with plausible arguments on both sides of the aisle.
However, the use of trigger warnings has spread beyond efforts to accommodate only
trauma survivors; trigger warnings have been used more broadly to shield members of other
disadvantaged groups from a wide range of content, including depictions of classism and
privilege (Boysen, 2017; Lukianoff & Haidt, 2015). Further, trigger warnings have become
normative in settings other than academia, such as online discussion groups (Wyatt, 2016). The
question of whether trigger warnings are beneficial or harmful for trauma survivors is an
important one. However, because trigger warnings are now applied to a broad range of content in
many different settings, another important question is whether they foster attitudes that
undermine resilience in people who have not – or not yet – experienced trauma. Despite the
timeliness and importance of this question, experimental research has remained silent on the
subject.
The areas of concern about how trigger warnings affect trauma survivors, such as
avoidance behaviors and trauma centrality, are distinct from those of interest in trauma-naïve
individuals. One such area of concern for those not yet traumatized is whether trigger warnings
increase individuals’ vulnerability to psychopathology, i.e. developing PTSD in the event of
exposure to trauma. Although trauma is common, PTSD is rare (Breslau & Kessler, 2001;
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS 5
McNally, 2014). Experiencing some symptoms of PTSD in the immediate aftermath of a
traumatic event is common, but symptoms rarely persist (Rothbaum, Foa, Riggs, Murdock, &
Walsh, 1992). Trauma survivors who appraise acute symptoms negatively are at heightened risk
for PTSD (Dunmore, Clark, & Ehlers, 2001; Ehring, Ehlers, & Glucksman, 2006). Trigger
warnings suggest that trauma survivors will have difficulty with content encountered in daily
life, and may lead people to believe that they are likely to develop PTSD should they encounter
trauma, causing them to iatrogenically catastrophize acute posttraumatic symptoms. Further,
receiving constant reminders of potential emotional harm may contribute to perceptions of
heightened vulnerability, fostering a maladaptive self-identification as a victim (Wyatt, 2016).
Similarly, trigger warnings may also change the way that people think about others’
vulnerability in the wake of trauma. Trigger warnings may raise awareness of the difficulties of
people suffering from PTSD. However, they may also create the impression that the experience
of trauma always renders survivors emotionally incapacitated. In reality, most trauma survivors
are resilient and show few symptoms of PTSD after an initial period of adjustment (Breslau &
Kessler, 2001). The perception of trauma survivors as dysregulated victims may contribute to
negative stigma concerning the very individuals trigger warnings are intended to protect.
Trigger warnings may also ironically increase acute anxiety by producing an expectation
of negative consequences. Indeed, nocebo effects (detrimental effects produced by negative
expectations) are an established phenomenon in psychological research (e.g. Barsky, Saintfort,
Rogers, & Borus, 2002). Research provides some support for a nocebo effect of trigger warnings
(Bruce, 2017a) indicating that physiological markers of anxiety are heightened in the presence of
trigger warnings in comparison to “PG-13” warning and “no warning” conditions. Such an
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS 6
effect may be exacerbated for individuals who already harbor the belief that exposure to
offensive words or other media can cause long-lasting emotional harm.
On the other hand, perception of control over stressors reduces stress reactions
(Thompson, 1981), and predictable stressors are less distressing than unpredictable ones (Grupe
& Nitschke, 2013; Mineka & Kihlstrom, 1978). Distressing physiological sensations produce
more anxiety when they violate expectations (Telch, Harrington, Smits, & Powers, 2011).
Therefore, trigger warnings may enable people who choose to view the material to brace
themselves for disturbing content without being surprised and dysregulated by its presentation.
Alternatively, trigger-warning accustomed individuals may develop the implicit assumption that
offensive content can always be anticipated, rendering even relatively innocuous content viewed
without a warning surprising and more fearful (the cognitive equivalent of Lukianoff and Haidt’s
“coddling” hypothesis). Such an effect may be exacerbated for individuals who are already have
high expectations of controllability and predictability in their daily lives.
The Current Study
Taken together, some research suggests that trigger warnings could be conducive to
better emotional functioning and lower anxiety levels, whereas other research indicates that they
may be anxiogenic and generative of risk for developing PTSD in the event of trauma. Despite
these equally plausible hypotheses (and the spirited political debate surrounding trigger
warnings), there is a dearth of research on trigger warnings’ impact on resilience factors in the
non-traumatized population.
Aims. Working within the tradition of experimental psychopathology, we sought to
determine whether (and in what way) trigger warnings affect resilience variables specific to
those who have not yet experienced potentially traumatic events. We also explored other
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS 7
demographic characteristics that may influence these resilience variables, and examined the
reasons that individuals might support the use of trigger warnings, apart from their psychological
reactions to them.
To achieve these aims, we recruited participants who had not experienced canonical
traumatic events. We restricted our sample to trauma-naïve individuals because we wanted to
examine how trigger warnings affect aspects of resilience specific to those who have not yet
been traumatized (e.g., perceived emotional vulnerability in the event of experiencing trauma),
which are distinct from those that concern traumatized individuals (e.g., encouraging avoidance
behaviors). We had participants read distressing passages from world literature either with
trigger warnings (experimental condition) or without trigger warnings (control condition) prior to
reporting their anxiety levels after each passage. Participants then completed measures
addressing perceptions of vulnerability in themselves and others. To test whether trigger
warnings affect subsequent emotional reactivity to less distressing content, we included
moderately distressing passages without a trigger warning at the end of the study. We also
wanted to assess traits that may influence one’s anxiety response to a trigger warning.
Accordingly, we measured participants’ strength of belief that words can harm people, enabling
us to test whether it affects anxiety in response to potentially distressing material preceded by a
trigger warning. We also measured participants’ assumptions about how controllable and
predictable the world is to test whether such beliefs increase anxiety provoked by less distressing
material not preceded by trigger warnings.
Research Questions. Due to different sources of indirect evidence suggesting that trigger
warnings may be either detrimental or helpful to resilience, and the lack of empirical data on this
topic, we formed research questions about whether trigger warning use would influence
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS 8
resilience variables, rather than making a priori hypotheses as to the direction of such effects.
Accordingly, we tested whether trigger warnings would (Q1) affect participants’ perceptions of
their posttraumatic vulnerability, (Q2) affect participants’ overall degree of implicit
identification as “vulnerable” versus “resilient”, and (Q3) affect participants’ perceptions of
others’ posttraumatic vulnerability. We also tested whether trigger warnings would (Q4) affect
immediate anxiety response to potentially distressing material, and whether the belief that words
can cause harm might amplify an anxiety response. We also examined whether (Q5) trigger
warnings would affect subsequent anxiety response to less distressing material, and whether
stronger beliefs in the world’s controllability and predictability might amplify this anxiety
response.
Method
Participants
Participants were recruited on Amazon’s Mechanical Turk (MTurk; Berinsky, Huber, &
Lenz, 2012), then read and acknowledged an institutionally approved informed consent form. A
single-item screening question excluded individuals who had experienced a canonical stressor
(e.g., rape, natural disaster) qualifying for Criterion A of the PTSD diagnosis in DSM-5
(American Psychiatric Association, 2013). Three hundred participants completed the study. Four
participants were excluded from all analyses because they reported having received a diagnosis
of PTSD despite denying exposure to canonical traumatic stressors. An additional 26
participants were excluded because they answered content-based attention check questions
incorrectly, indicating inattentive responding. This left 270 participants, 133 in the Trigger
Warning condition, and 137 in the No Warning condition.
Materials
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS 9
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 seconds 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.
To ensure that our passages elicited levels of anxiety consistent with their categories, we
conducted a pilot study involving 50 participants on MTurk to norm each passage’s anxiogenic
properties. Forty candidate passages were included. Means and inter-quartile ranges (IQRs) of
anxiety response were calculated for each passage. Passages with the lowest means and IQRs
that did not extend above the grand mean were designated neutral. Passages with means closest
to the grand mean and IQRs within the grand IQR were designated mildly distressing. Passages
with the highest means and IQRs that extended above the grand IQR were designated markedly
distressing.
Measures
Perceived Posttraumatic Vulnerability Scale-Self (PPVS-S). The PPVS-S is a 19-item
questionnaire which assesses belief in the likelihood of long-term adverse emotional effects of
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
10
trauma exposure. These perceived vulnerabilities include developing a mental disorder, being
unable to effectively regulate emotions, or functional disability. Participants are asked to imagine
themselves experiencing a hypothetical traumatic event, and to indicate their level of
endorsement for each statement concerning its effects (e.g. I would lose my grip on reality.) on a
100-point scale (1 = disagree, 100 = agree). These responses are averaged for a composite
score. Higher scores indicate stronger belief in vulnerability. The PPVS-S displayed excellent
internal consistency in our sample (α = .95). All measures devised for this experiment are in the
supplementary materials, S2.
Perceived Posttraumatic Vulnerability Scale-Other (PPVS-O). Analogous in format
to the PPVS-S, the PPVS-O assesses the degree to which individuals believe that trauma
survivors are vulnerable to long-term negative emotional events. Participants are asked to
imagine a hypothetical “average” person experiencing a traumatic event, and indicate their level
of endorsement for each statement in reference to this person (e.g. He/she would feel isolated
and alone.) on a 100-point scale (1 = disagree, 100 = agree). These responses are averaged for a
composite score. Higher scores indicate stronger beliefs that trauma survivors will experience
persistent and debilitating negative emotional effects. The PPVS-O displayed excellent internal
consistency in our sample (α = .96).
Implicit Association Test (IAT), Vulnerability vs. Resilience.
The IAT (Greenwald,
McGhee, & Schwartz, 1998) measures the degree to which a participant implicitly associates
concepts with each other. Response latencies when sorting items between different categories
formed by concept pairs determines which concepts are more strongly associated with each other
(Greenwald, Poehlman, Uhlmann, & Banaji, 2009). The IAT has good internal consistency and
test-retest reliability (Bosson, Swann, & Pennebaker, 2000) and convergent validity in its
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
11
agreement with measures of explicit preferences (Greenwald et al., 1998). Our IAT assessed
strength of implicit association between the self (i.e., me versus not me) and vulnerability (i.e.,
vulnerable resilient). More positive d-scores on the IAT indicate a greater implicit association
of the self with the resilient attribute versus the opposite configuration; more negative scores
indicate a greater association of the self with the vulnerable attribute.
The IAT had adequate
reliability in our sample (α = .85). While calculating these d-scores, we identified 25 participants
whose response latencies were implausibly fast, signifying invalid responding. Indeed, several
participants reported difficulties with the IAT arising from online connectivity glitches. We
eliminated these 25 participants from analyses involving the IAT. We IAT was created using,
iatgen, an open-source IAT builder for online surveys, and was analyzed using the iatgen
package for R (Carpenter et al., 2017).
Words-Can-Harm Scale (WCHS). The WCHS (see supplementary materials, S1) is a
10-item scale that assesses the degree to which an individual believes that exposure to offensive
words has the potential to cause serious harm to themselves or other people. Participants
indicated their level of endorsement for each statement (e.g. I could be traumatized without ever
being touched, just through someone's hurtful words) on a 100-point scale (1 = disagree, 100 =
agree). Responses were averaged for a composite score. Higher scores indicate stronger beliefs
that words can harm people. The WCHS displayed excellent internal consistency in our sample
(α = .92).
World Assumptions Scale (WAS). The WAS (Janoff-Bulman, 1989) is a measure that
assesses a participant’s degree of belief in different underlying assumptions about the world and
themselves. For the purposes of our experiment, we used only the 3 subscales from this measure
that pertain to our proposed moderator, which deals with controlling and predicting stressful
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
12
events (Controllability, Randomness, and Self-Controllability Subscale), a total of 12 items.
Participants indicate their level of agreement with statements about their underlying assumptions
(Through our actions we can prevent bad things from happening to us) on a scale from 1
(strongly disagree) to 7 (strongly agree). These subscales were averaged (with the Randomness
Subscale reverse-scored) to create a composite score reflecting perceptions of the world’s
controllability and predictability. Higher scores indicate stronger beliefs that one’s world is
predictable and controllable. Our controllability/predictability scale had good internal
consistency (α = .83). The subscales of the WAS have satisfactory convergent validity in their
sensitivity to whether individuals have experienced trauma (Janoff-Bulman, 1989), and correlate
with posttraumatic symptom severity (Elklit, Shevlin, Solomon, & Dekel, 2007), indicating that
different life events can affect global assumptions about the world.
Trigger Warning Attitudes Assessment (TWAA). The TWAA (see supplementary
materials, S1) is a two-item scale that assesses attitudes toward trigger warnings. First,
participants receive a short definition of trigger warnings, and are asked “Do you think that
trigger warnings should be used?” If participants agree with this statement, they are then asked
“Why do you think that trigger warnings should be used?” Participants view a list of potential
reasons for trigger warning use (e.g. protection of vulnerable populations, fairness, psychological
harm) and are asked to select all that apply. An “other” category is also provided, and
participants can add reasons not listed.
Demographics Questionnaire. This questionnaire asked for non-identifying information
on participants’ backgrounds. The questionnaire assessed gender, self-reported race and
ethnicity, and age. Religiosity was assessed using a 5-point Likert scale (1 = not religious. 5 =
extremely Religious), as was political orientation (1 = very liberal, 5 = very conservative).
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
13
Psychiatric History Questionnaire. This questionnaire asked for participants’ history of
psychiatric diagnoses, providing them with a list of potential diagnoses as well as an “other”
option that allowed them to add any not included on the questionnaire.
Procedure
After undergoing institutional review and receiving approval, our online experiment was
posted as a Human Intelligence Task (HIT) on MTurk. The HIT description indicated that our
survey involved reading and providing feedback on passages from literature. The consent form
also mentioned that the readings would “cover a diverse range of emotional and dramatic
content.” After providing informed consent, participants were screened for exposure to
traumatic events. We excluded those reporting trauma exposure, and randomly assigned the
others to either the No Warning or Trigger Warning condition.
Participants in both conditions then read three mildly distressing passages in random
order. After each passage, they used slider bar scales ranging from 0 (not at all) to 100 (very
much) to rate their response on the following measures: sad, happy, afraid, anxious, angry,
content, disgusted, degree of unpleasant emotion overall, and degree of anticipated long-term
negative emotion. The target emotion was anxiety; the other items were fillers included to
diminish demand effects. The average of these three passages’ anxiety responses served as the
baseline anxiety response for each participant.
Next, participants read another series of 10 passages in random order. Five were neutral,
and the other five were markedly distressing. In the Trigger Warning condition, each of the
markedly distressing passages was preceded by a trigger warning screen which had to be
acknowledged by clicking a radio button, i.e. TRIGGER WARNING: The passage you are about
to read contains disturbing content and may trigger an anxiety response, especially in those who
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
14
have a history of trauma. (Although we screened out individuals who experienced events likely
to constitute Criterion A traumas, we included the phrase concerning trauma victims because it
unmistakably qualifies the statement as a trigger warning.) The No Warning condition
participants viewed a screen that indicated they were about to view the next passage, which was
also acknowledged by clicking a radio button. Participants rated the intensity of their reactions
after each markedly distressing passage; the difference between the average of these anxiety
ratings and the baseline average anxiety rating constituted the “immediate anxiety change” for
each participant.
After completion of condition-specific passage presentations, participants read three more
mildly distressing passages and rated the intensity of their reactions. The difference between the
average of these anxiety responses and baseline anxiety responses constituted the “follow-up
anxiety change” score for each participant. Next, participants completed measures presented in
random order that tapped the outcomes of perceived vulnerability to posttraumatic symptoms for
themselves (PPVS-S) and others (PPVS-O) and a measure of implicit self-identification as
vulnerable versus resilient (IAT). Participants also completed measures that tapped constructs
hypothesized to moderate the relationship between trigger warning presentation and immediate
anxiety change (Words Can Harm Scale; WCHS) and trigger warning presentation and follow-up
anxiety change (Controllability Scale derived from the World Assumptions Scale; WAS).
Participants also responded to the demographics questionnaire, the psychiatric diagnosis history
questionnaire, and the TWAA. Finally, participants were provided with a debriefing form which
explained the deception implemented and purpose of the experiment. The debriefing form also
informed participants that “the graphic nature of some of the passages may have caused you
discomfort or emotional distress. Such feelings, although unpleasant, usually subside fairly
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
15
quickly.” The form provided information on available free resources for any persistent distress
that participants might have experienced. Participants were each compensated $3.00 for
completing the entire survey based on a projected completion time of no more than 1 hour, a rate
of payment consistent with ethical guidelines for crowdsourced remuneration (Chandler &
Shapiro, 2016).
Planned Analyses
Our sample size (N = 270) provided sufficient power (1 – β error probability = .96) to
detect a small effect size (f
2
= .10) in our planned interaction analyses, which had the greatest
number of possible predictors of all our analyses (maximum possible predictors = 9). We based
our sample size requirements on a small effect, as there is no precedent for experiments
involving trigger warnings. We first planned to examine demographic characteristics to
determine whether participants had been effectively randomized to condition. As an exploratory
analysis, we also planned to determine the reasons why participants might favor the use of
trigger warnings. Next, we planned to conduct bivariate analyses between our demographic and
outcome variables in order to determine which characteristics of the sample should be controlled
for in the main analyses. For our main analyses, we planned to conduct multiple regressions to
determine the effects of trigger warnings on each outcome variable while controlling for relevant
demographic characteristics. Following the suggestion of an anonymous peer reviewer, we also
conducted uncontrolled regression analyses in order to account for the possibility of statistical
overcontrol (Meehl, 1971). We also planned follow-up analyses to examine our proposed
moderated relationships between trigger warnings and anxiety changes using regression-based
interaction detections and simple slopes analyses. See supplementary materials (S3) for all R
code used in our analyses.
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
16
Results
Sample Characteristics
The sample contained a majority of females (n = 156, 57.8%), and the mean age was 37
years old (SD = 12.4 years). Race was predominantly Caucasian (n = 191, 70.7%), with 9.6%
African American (n = 26) and 9.3% Asian/Pacific Islander (n = 25) participants. Ethnicity was
predominantly non-Hispanic (n = 250, 92.6%), and political orientation was predominantly at
least “somewhat liberal” (n = 146, 54.0%). The majority of participants identified as at least
“somewhat religious” (n =156, 57.8%). A total of 42 participants (15.6%) endorsed one or more
psychiatric diagnoses other than PTSD. A majority of participants (n = 216, 80.0%) believed
that trigger warnings should be used. Of these, a large majority based their belief on the need to
protect psychologically vulnerable populations (such as those with PTSD) (n = 192, 88.9%),
with roughly half believing that protection should be afforded to any minority group member (n
= 109, 50.5%) or to people in general (n = 112, 51.9%). When measured as a continuous
variable, political orientation differed by condition (r
pb
= .13, p < .05), indicating that the Trigger
Warning condition participants were slightly more conservative than those in the No Warning
condition. Therefore, political orientation was included as a covariate in all regression analyses
(both controlled and uncontrolled).
Bivariate Associations
Bivariate associations between demographic characteristics and outcome variables appear
in Table 1. Women, racial minorities, liberals, younger individuals, and those with at least one
psychiatric diagnosis perceived themselves as more vulnerable to persistent negative emotional
effects in the event of trauma than did men, Caucasians, conservatives, older participants, and
those without a psychiatric diagnosis. Liberals, younger individuals, and racial minorities
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
17
perceived trauma survivors in general as more vulnerable. Accordingly, we included these
variables in those outcomes’ controlled regression analyses.
Table 1
Bivariate Correlations between Demographic Variables and Outcome Variables (N = 270)
IAC FAC
PPVS
-S PPVS-O
IAT
Gender
a
,b
-.10 -.01
-.20** -.05 .11
Race
a,
c
-.03 -.01
.13*
.16**
-.02
Ethnicity
a,
-.02 -.04
.00
.07
-.06
Psychiatric Diagnostic Status
a,
d
-.06 -.10
.24***
.10
-.09
Religiosity -.04 -.02
-.01
-.01
-.06
Political Orientation -.03 -.01
-.14*
-.16**
.03
Age
-.08 .09
-.18**
-.20**
.07
Note. IAC = Immediate Anxiety Change, FAC = Follow-Up Anxiety Change, PPVS-S =
Perceived Posttraumatic Vulnerability Scale – Self, PPVS-O = Perceived Posttraumatic
Vulnerability Scale – Other, IAT = Vulnerable/Resilient IAT d-score.
a
Correlation coefficients depicted for these dichotomous variables are point-biserial correlations
(r
pb
).
b
Dichotomized as female = 0, male = 1.
c
Dichotomized as non-minority = 0, minority = 1.
d
Dichotomized as no diagnosis = 0, at least one diagnosis = 1.
e
n = 245 due to 25 IAT scores
identified as invalid.
***p < .001, **p < .01, *p < .05
Multiple Regression Analyses
Table 2 shows the results of the multiple regressions for each outcome variable, with relevant
demographic characteristics entered as control variables and condition (Trigger Warning or No
Warning) entered as predictors. Relative to participants who received no trigger warnings, those
receiving them perceived themselves as more vulnerable to suffering persistent negative
emotional effects in the event of experiencing trauma (i.e., a 5.2% increase in the strength of this
belief, B = 5.17, t(263) = 2.12, p < .05). The results of the uncontrolled analysis for this outcome
were similar; B = 5.48, t(263) = 2.13, p < .05. Relative to participants who received no warnings,
those who received trigger warnings had stronger beliefs that trauma survivors would suffer
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
18
persistent negative emotional effects, (i.e., a 5.4% increase in the strength of this belief, B = 5.38,
t(265) = 2.35, p < .05). The results of the uncontrolled analysis for this outcome were similar; B
= 5.09, t(263) = 2.19, p < .05). No significant effect of condition was found for participants’
implicit identification of self with the attributes of vulnerable versus resilient. The groups with
and without trigger warnings did not differ in their immediate anxiety change in response to
“markedly distressing” content during the experimental paradigm. Similarly, exposure to trigger
warnings did not display a global effect on participants’ follow-up anxiety change.
Table 2
Multiple Regression Analyses of the Effect of Condition on Outcome Variables, Controlling for
Relevant Demographic Characteristics (N = 270)
Outcome Variable
Predictor
PPVS-S
PPVS-O
IAT
IAC
FAC
B SE B B SE B B SE B B SE B B SE B
Condition
b
5.16* 2.43 5.38* 2.29 .02 .05 .98 1.93 -3.59
2.01
R
2
.17** .09** .00 .00 .01
Note. PPVS-S = Perceived Posttraumatic Vulnerability Scale – Self, PPVS-O = Perceived
Posttraumatic Vulnerability Scale – Other, IAT = Vulnerable/Resilient IAT d-score, IAC =
Immediate Anxiety Change, FAC = Follow-Up Anxiety Change.
a
n = 245 due to 25 IAT scores identified as invalid.
b
Control predictors entered included gender
(PPVS-S), race (PPVS-S, PPVS-O), psychiatric diagnostic status (PPVS-S), age (PPVS-S,
PPVS-O), and political orientation (all outcomes).
**p < .001 *p < .05
Moderation Analyses
For our moderation analyses, we entered political orientation, condition, moderator
variable scores, and the cross-product of condition and the moderator variable scores as
independent variables in a multiple regression predicting the outcome of interest. If the cross-
product’s coefficient was significant, we considered it support for the presence of an interaction.
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
19
Table 3 shows the results of these interaction detections. For significant interactions, we
conducted a simple slopes analysis that tested for the conditional effect of the predictor variable
on the outcome variable at 1 SD above and below the mean of the moderator variable.
Table 3
Interaction Detection Analyses for the Prediction of Anxiety Change Variables (N = 270)
Model 1: Immediate Anxiety Change F(4, 265) = 4.67
R
2
= .07**
Variable B SE
Condition -12.52* 5.07
WCHS Score -.01 .06
Condition × WCHS Score .24**
.09
Model 2: Follow-Up Anxiety Change F(4, 265) = 2.79 R
2
= .04*
Variable B SE
Condition
.27 11.27
Controllability Score .90* .41
Condition × Controllability Score -.19 .63
Note. WCHS = Words Can Harm Scale. Political orientation was controlled for in both analyses.
**p < .01, *p < .05
Words Can Harm Belief. The analyses suggest that trigger warnings increase acute
anxiety to the extent that participants believe that words can cause harm. A simple slopes
analysis indicated that for participants who do not have a strong belief that words can cause harm
(M
WCHS
– 1 SD), receiving a trigger warning does not significantly increase anxiety from
baseline (B = -4.57, t(265) = -1.73, ns). However, if participants have a strong belief that words
can harm (M
WCHS
+ 1 SD), trigger warnings significantly increase anxiety from baseline (B =
5.90, t(265) = 2.20, p < .05). Figure 1 depicts this moderating function.
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
20
Figure 1. Simple slopes of condition predicting change in immediate anxiety response from
baseline at high (M + 1 SD) and low (M – 1 SD) values of the belief that words can harm.
Assumptions of Controllability. This analysis indicated that assumptions about the
world’s controllability do not change the relation between trigger warnings and subsequent
anxiety change, as the cross-product term was nonsignificant. However, assumptions of
controllability did display a small but significant main effect on follow-up anxiety change (B =
.90, t(265) = 2.16, p < .05), such that higher controllability beliefs increased anxiety change by a
very small percentage.
Discussion
This study is the first to examine the effects of trigger warnings on individual resilience
factors via a randomized controlled experiment. Our results indicate that trigger warnings affect
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
21
some specific domains of resilience relevant to trauma-naïve individuals, but seem to matter less
for other domains. We will now address each of our questions and discuss implications for
resilience to stress and trauma.
Perceived Vulnerability of the Self (Q1, Q2)
Trigger warnings increased people’s perceived risk of suffering long-term debilitating
emotional harm (such as PTSD) in the wake of a traumatic event (Q1). This effect, albeit small,
is notable. Beliefs about the self are generally quite stable (Church et al., 2012); a significant
change based on such a small manipulation is somewhat surprising. Trigger warnings may
increase perceptions of self-vulnerability by sending an implicit message about the long-term
harm caused by trauma; extensive exposure to trigger warnings may amplify this effect. This
result has implications for resilience, as pathogenic appraisal of one’s emotional reactions to
stressors increases risk for PTSD (Dunmore, Clark, & Ehlers, 2001; Ehring, Ehlers, &
Glucksman, 2006). Importantly, the effect of trigger warnings on perceptions of vulnerability
appear to apply only to explicit beliefs regarding resilience to traumatic events; trigger warnings
did not significantly affect implicit identification of the self as resilient versus vulnerable (Q2).
Perceived Posttraumatic Vulnerability of Others (Q3)
Our results also indicate that trigger warnings enforce a “soft stigma” concerning trauma
survivors, implying their inability to function as other people can. This effect was also small, but
may be additive over the long-term. This finding suggests trigger warnings may have unintended
yet potentially deleterious consequences for those they aim to protect.
Anxiety Response to Potentially Distressing Material (Q4)
Trigger warnings did not affect anxiety responses to potentially distressing material in
general. However, trigger warnings may foster a self-fulfilling prophecy (Merton, 1948) that
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
22
increases anxiety for those individuals who believe that words can harm them. Trigger warnings
themselves do not appear to generate the belief that words can harm, as the strength of this belief
was not significantly related to condition. Rather, trigger warnings may confirm this belief in
those who already harbor it. Hence, such warnings may increase acute anxiety by fostering an
expectancy of harm (Barsky, Saintfort, Rogers, & Borus, 2002; Reiss & McNally, 1985).
Subsequent Anxiety Response to Less Distressing Material (Q5)
Trigger warnings did not affect reactivity to mildly distressing material viewed without a
warning, indicating that their anxiogenic effects are limited to immediate reactions for a specific
subset of people. Additionally, assumptions that one’s world is controllable and predictable do
not appear to affect this relationship, failing to support the notion that trigger warnings
exacerbate an expectancy of predictability that sensitizes people to less severe unexpected
stressors (the “coddling” hypothesis; Lukianoff & Haidt, 2015).
Relationships between Resilience Factors and Demographic Variables
To some extent, our outcome variables correlated in expected directions with
demographic variables. For example, perceived vulnerability to posttraumatic impairment is
associated with demographic factors positively associated for risk of PTSD among people
exposed to trauma, such as being female (Tolin & Foa, 2006) and having pre-existing psychiatric
disorders (Breslau, Davis, Andreski, & Peterson, 1991). Additionally, racial minority status and
younger age were associated with higher levels of perceived vulnerability. In this case, it is
possible that these variables are proxies for more meaningful third variables. For example, the
relation between age and perceived risk for impairment could signify a cohort effect; younger
participants perceiving greater risk may result from their upbringing within an especially
protective cultural moment (Lukianoff & Haidt, 2015).
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
23
Trigger Warning Attitudes
A large majority of participants supported the use of trigger warnings, independently of
whether they were randomized to the trigger warning condition or to the control condition. A
considerable proportion of these participants believed that trigger warnings are needed not only
by those with PTSD and other psychological vulnerabilities, but also by minority groups and
people in general. These results suggest that trigger warnings are viewed by many people as
applicable to a much broader range of concerns than those of accommodating people with PTSD,
as others have noted (Boysen, 2017; Lukianoff & Haidt, 2015).
Limitations and Future Directions
One limitation of our study is that we had to devise novel and hitherto untested measures
to assess perceived vulnerability to posttraumatic impairment (PPVS-S, PPVS-O), belief that
words can harm (WCHS), and attitudes about trigger warnings (TWAA). However, their internal
consistencies were high, and several correlated in expected directions with demographic
variables. Another limitation was the use of self-report. As with any online crowdsourced study,
the validity of responses can be difficult to determine (e.g., an accurate report of a participant’s
psychiatric history, unverified by medical records). Yet our use of specific prompts (e.g., having
participants choose from a list of potential diagnoses rather than just asking a simple yes or no
question concerning one’s diagnostic history) and content-based attention checks mitigate the
effects of this limitation.
Do our findings merely reflect demand effects? Perhaps participants in the trigger
warning condition reported themselves and others as vulnerable to posttraumatic impairment
after exposure to the explicit warnings embodied in trigger warnings. Perhaps they merely tried
to satisfy presumptive experimental expectations rather than conveying their actual beliefs about
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
24
posttraumatic vulnerability. To guard against such demand effects, we did not directly ask them
whether they thought they would develop PTSD following trauma. Rather, we asked them to
imagine themselves surviving an attempted murder (the words “trauma” or “PTSD” were not
mentioned) and then asked them to rate the likely severity of specific symptoms they believed
they might experience thereafter. Additionally, trigger warnings did not have a main effect on
immediate anxiety responses to passages, but rather only displayed an effect for those who
strongly believed that words can harm. This result indicates that that trigger warnings are
achieving their effects by exacerbating specific iatrogenic beliefs about the likelihood of harm
rather than by activating participants’ desire to be good research subjects.
This study used the written word as stimuli, rather than in vivo stress inductions or the
use of film or images. The use of literary passages as stimuli is a strength of this study, as the
written word is ubiquitous in educational settings -- the center of the trigger warning debate.
However, the effortful engagement required in order for the written word to induce an emotional
response may have limited the size of our effects when compared to the use of more vivid media.
Future research should examine whether the effects generalize to other types of stimuli.
Our study used participants from a crowdsourcing website, as we were primarily
interested in the effects of trigger warnings in the general population. It is unclear whether our
findings generalize to an exclusively collegiate population. However, the MTurk population is
more demographically diverse than the typical undergraduate population (Chandler & Shapiro,
2016; Mason & Suri, 2012), and hence our findings may generalize broadly to American society.
Our study emphasized pre-traumatic resilience, and it remains unclear whether our results
pertain to traumatized individuals. Nevertheless, Bruce’s (2017a) research indicates that some
of our effects may apply to a population with PTSD. She found that physiological markers of
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
25
anxiety were heightened after the presentation of a trigger warning when compared to “PG-13”
and “no warning” conditions, and that this effect was significantly larger for those with more
severe PTSD symptoms. More broadly, concerns about trigger warnings as they apply to
trauma-naïve individuals are different from the resilience factors at issue in trauma survivors,
such as iatrogenically encouraging avoidance of trauma-related cues, and reinforcing the
centrality of trauma to individuals’ identities (McNally, 2014).
Research to date has lent some plausibility to such concerns, showing significant positive
cross-sectional associations between amount of trigger warning use and trauma centrality (Bruce;
2017a, b), and between degree of trigger warning use and avoidance behavior (Bruce, 2017b).
Researchers should address trauma survivor-specific concerns about trigger warnings with
experimental tests to clarify these issues.
Conclusion
Taken together, our findings provide a preliminary look at the effects of trigger warnings
on pre-traumatic resilience variables as they apply to the general population, and a step forward
in answering the question of whether trigger warnings help or harm. Trigger warnings do not
appear to be conducive to resilience as measured by any of our metrics. Rather, our findings
indicate that trigger warnings may present nuanced threats to selective domains of psychological
resilience. Such consequences are limited to perceived vulnerability to emotional harm, which
may increase risk for developing PTSD in the event of trauma, and disability-related stigma
around trauma survivors. However, this effect does not apply to implicit self-identification
regarding vulnerability. Trigger warnings do not appear to affect sensitivity to distressing
material in general, but may increase immediate anxiety response for a subset of individuals
whose beliefs predispose them to such a response. These findings do not form the basis for
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
26
immediate policy changes regarding the use of trigger warnings without subsequent replication,
as effect sizes were small. This may partly be due to our use of literary passages that require
effortful engagement to induce emotional response, and to our use of a non-traumatized sample.
Finally, although many of our subjects were of the same cohort as college students, replication in
an entirely collegiate sample is warranted, as this population is especially likely to experience
exposure to trigger warnings.
Declaration of Interest and Funding
None of the authors has conflicts of interest to report. This work was supported by
research support funds from the Department of Psychology at Harvard University.
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
27
References
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders
(5
th
ed.). Washington, DC: Author.
Barsky, A. J., Saintfort, R., Rogers, M. P., & Borus, J. F. (2002). Nonspecific medication side
effects and the nocebo phenomenon. Journal of the American Medical Association, 287,
622–627. http://dx.doi.org/10.1001/jama.287.5.622
Berinsky, A. J., Huber, G. A. & Lenz, G. S. (2012). Evaluating online labor markets for
experimental research: Amazon.com’s Mechanical Turk. Political Analysis, 20(3), 351–
368. http://dx.doi.org/10.1093/pan/mpr057
Berntsen, D. & Rubin, D. C. (2007). When a trauma becomes a key to identity: Enhanced
integration of trauma memories predicts posttraumatic stress disorder symptoms. Applied
Cognitive Psychology, 21, 417–431. doi: 10.1002/acp.1290
Boelen, P. (2012) A prospective examination of the association between the centrality of a loss
and post-loss psychopathology. Journal of Affective Disorders, 137, 117-124.
doi:10.1016/j.jad.2011.12.004
Bosson, J. K., Swann, W. B., & Pennebaker, J. W. (2000). Stalking the perfect measure of
implicit self-esteem: The blind men and the elephant revisited? Journal of Personality
and Social Psychology, 79, 79-101.
Boysen, G. A. (2017). Evidence-based answers to questions about trigger warnings for clinically-
based distress: A review for teachers. Scholarship of Teaching and Learning in
Psychology, 3(2), 163-177. http://dx.doi.org/10.1037/stl0000084
Bradley, J. (2000). Flags of Our Fathers. New York, NY: Bantam Dell.
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
28
Breslau, N., Davis, G. C., Andreski, P., & Peterson, E. (1991). Traumatic events and
posttraumatic stress disorder in an urban population of young adults. Archives of General
Psychiatry, 48, 216-222. http://dx.doi.org/10.1001/archpsyc.1991.01810270028003
Breslau, N., & Kessler, R. C. (2001). The stressor criterion in DSM-IV posttraumatic stress
disorder: an empirical investigation. Biological psychiatry, 50(9), 699-704.
https://doi.org/10.1016/S0006-3223(01)01167-2
Bruce, M. J. (2017a, May). Does trauma centrality predict trigger warning use? Physiological
responses to using a trigger warning. Poster presented at the annual meeting of the
Midwest Psychological Association, Chicago, IL.
Bruce, M. J. (2017b, November). 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.
Carpenter, T., Pogacar, R., Pullig, C., Kouril, M., LaBouff, J., Aguilar, S., … Chakroff, A.
(2017, August 7). Building and analyzing Implicit Association Tests for online surveys:
A tutorial and open-source tool. Retrieved from osf.io/preprints/psyarxiv/6xdyj.
Carter, A. M. (2015). Teaching with trauma: Trigger warnings, feminism, and disability
pedagogy. Disability Studies Quarterly, 35, 9.
Chandler, J. & Shapiro, D.N. (2016). Conducting clinical research using crowdsourced
convenience samples. Annual Review of Clinical Psychology, 12, 53-81.
Church, T. A., Alvarez, J. M., Katigbak, M. S., Mastor, K. A., Cabrera, H. F., Tanaka-Matsumi,
J.,… Buchanan, A. L. (2012). Self-concept consistency and short-term stability in eight
cultures. Journal of Research in Personality, 46(5), 556-570.
Dostoevsky, F. (1917). Crime and Punishment. New York, NY: P.F. Collier and Son
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
29
Dunmore, E., Clark, D. M., & Ehlers, A. (2001). A prospective investigation of the role of
cognitive factors in persistent posttraumatic stress disorder (PTSD) after physical or
sexual assault. Behaviour Research and Therapy, 39, 1063-1084.
https://doi.org/10.1016/S0005-7967(00)00088-7
Ehring, T., Ehlers, A., & Glucksman, E. (2006). Contribution of cognitive factors to the
prediction of post-traumatic stress disorder, phobia, and depression after motor vehicle
accidents. Behaviour Research and Therapy, 44, 1699-1716.
https://doi.org/10.1016/j.brat.2005.11.013
Elklit, A., Shevlin, M., Solomon, Z. and Dekel, R. (2007). Factor structure and concurrent
validity of the world assumptions scale. Journal of Traumatic Stress, 20, 291–301.
doi:10.1002/jts.20203
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. (1998). Measuring individual differences in
implicit cognition: the implicit association test. Journal of personality and social
psychology, 74(6), 1464-1480.
Greenwald, A., Poehlman, T., Uhlmann, E., & Banaji, M. (2009). Understanding and Using the
Implicit Association Test: III. Meta-Analysis of Predictive Validity. Journal of
Personality and Social Psychology, 97(1), 17-41.
Grupe, D. W., & Nitschke, J. B. (2013). Uncertainty and anticipation in anxiety: An integrated
neurobiological and psychological perspective. Nature Reviews Neuroscience, 14, 488–
501. http://dx.doi.org/10.1038/nrn3524
Institute of Medicine of the National Academies (2008). Treatment of posttraumatic stress
disorder: An assessment of the evidence. Washington, DC: The National Academies
Press.
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
30
Janoff-Bulman, R. (1989). Assumptive worlds and the stress of traumatic events: Applications of
the schema construct. Social Cognition, 7, 113-136.
Lukianoff, G., & Haidt, J. (2015, September). The coddling of the American Mind. The Atlantic.
Retrieved from https://www.theatlantic.com/magazine/archive/2015/09/the-coddling-of-
the-american-mind/399356/
Mason, W., & Suri, S. (2012). Conducting behavioral research on Amazon’s Mechanical
Turk. Behavior Research Methods, 44, 1–23. https://doi-org.ezp-
prod1.hul.harvard.edu/10.3758/s13428-011-0124-6
McNally, R. J. (2014, May 20). Hazards ahead: The problem with trigger warnings, according to
the research. Pacific Standard. Retrieved from https://psmag.com/education/hazards-
ahead-problem-trigger-warnings-according-research-81946
McNally, R. J. (2016, September 13). If you need a trigger warning, you need P.T.S.D.
treatment. New York Times. Retrieved from
https://www.nytimes.com/roomfordebate/2016/09/13/do-trigger-warnings-work/if-you-
need-a-trigger-warning-you-need-ptsd-treatment
Meehl, P. E. (1971). High school yearbooks: A reply to Schwarz. Journal of Abnormal
Psychology, 77, 143-148. doi: 10.1037/h0030750
Melville, H. (1967). Moby-Dick. H. Hayford & H. Parker (Eds.). New York, NY: Norton
Merton, R. K. (1948). The self-fulfilling prophecy. Antioch Review, 8, 1930210.
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
31
Reiss, S., & McNally, R. J. (1985). Expectancy model of fear. In S. Reiss & R. R. Bootzin
(Eds.), Theoretical issues in behavior therapy (pp. 107–121). San Diego, CA: Academic
Press.
Robinaugh, D. J., & McNally, R. J. (2011). Trauma centrality and PTSD symptom severity in
adult survivors of childhood sexual abuse. Journal of Traumatic Stress, 24(4), 483-486.
http://dx.doi.org/10.1002/jts.20656
Rosenthal, M. Z., Hall, M. L., Palm, K. M., Batten, S. V., & Follette, V. M. (2005). Chronic
avoidance helps explain the relationship between severity of childhood sexual abuse and
psychological distress. Journal of Child Sexual Abuse, 14, 25-41.
http://dx.doi.org/10.1300/J070v14n04_02
Rothbaum, B. O., Foa, E. B., Riggs, D. S., Murdock, T., & Walsh, W. (1992). A prospective
examination of post
traumatic stress disorder in rape victims. Journal of Traumatic
stress, 5(3), 455-475. http://dx.doi.org/10.1002/jts.2490050309
Stokes, M. (2014). In defense of trigger warnings. Chronicle of Higher Education. Retrieved
from http://chronicle.com/blogs/conversation/2014/05/29/in-defense-of-trigger-warnings/
Telch, M. J., Harrington, P. J., Smits, J. A., & Powers, M. B. (2011). Unexpected arousal,
anxiety sensitivity, and their interaction on CO2-induced panic: Further evidence for the
context-sensitivity vulnerability model. Journal of Anxiety Disorders, 25, 645-653.
https://doi.org/10.1016/j.janxdis.2011.02.005
Thompson, S. C. (1981). Will it hurt less if I can control it? A complex answer to a simple
question. Psychological Bulletin, 90, 89–101. http://dx.doi.org/10.1037/0033-
2909.90.1.89
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
TRIGGER WARNINGS
32
Tolin, D. F., & Foa, E. B. (2006). Sex differences in trauma and posttraumatic stress disorder: A
quantitative review of 25 years of research. Psychological Bulletin, 132, 959-992.
https://doi.org/10.1037/0033-2909.132.6.959
Wilson, R. (2015, September 14). Students’ requests for trigger warnings grow more varied.
Chronicle of Higher Education. Retrieved from http://chronicle.com/article/Students-
Requests-for/233043
Wyatt, W. (2016). The ethics of trigger warnings. Teaching Ethics, 16, 17–35.
http://dx.doi.org/10.5840/tej201632427
MANUS CRIP T
ACCEP TED
ACCEPTED MANUSCRIPT
Highlights
Trigger warnings increase peoples’ perceived emotional vulnerability to trauma.
Trigger warnings increase peoples’ belief that trauma survivors are vulnerable.
Trigger warnings increase anxiety to written material perceived as harmful.
  • Article
    Trigger warnings are messages alerting people to content containing themes that could cause distressing emotional reactions. Advocates claim that warnings allow people to prepare themselves and subsequently reduce negative reactions toward content, while critics insist warnings may increase negative interpretations. Here, we investigated (a) the emotional impact of viewing a warning message, (b) if a warning message would increase or decrease participants' negative evaluations of a set of ambiguous photos, and (c) how participants evaluated overall study participation. We meta-analyzed the results of 5 experiments (N = 1,600) conducted online, and found that trigger warnings did not cause participants to interpret the photos in a more negative manner than participants who were unwarned. However, warned participants experienced a negative anticipatory period prior to photo viewing that did little to mitigate subsequent negative reactions. (PsycINFO Database Record (c) 2019 APA, all rights reserved).
  • Article
    In 2014, a US college created a policy requiring faculty to provide trigger warnings for students. This spurred a heated debate across North America regarding the need for and efficacy of trigger warnings in classes. A content analysis of comment responses (over 1500) to 20 articles on the topic of trigger warnings from two higher education news journals (Inside Higher Ed; Chronicle of Higher Education) was performed. The trends, quality, and nature of the comments were categorized and opinions were gleaned in an effort to understand common arguments for or against the implementation of trigger warnings in the classroom. Findings against trigger warnings included concerns about academic freedom, infantalization of students, and unfair responsibility for professors; whereas findings for included promotion of positive pedagogical values, recognition of human courtesy, and supporting student mental health. This study helps to inform pedagogical practices and further research on best practices in higher education.
  • Article
    Students are requesting and professors issuing trigger warnings—content warnings cautioning that college course material may cause distress. Trigger warnings are meant to alleviate distress students may otherwise experience, but multiple lines of research suggest trigger warnings could either increase or decrease symptoms of distress. We examined how these theories translate to this applied situation. Across six experiments, we gave some college students and Internet users a trigger warning but not others, exposed everyone to one of a variety of negative materials, then measured symptoms of distress. To better estimate trigger warnings’ effects, we conducted mini meta-analyses on our data, revealing trigger warnings had trivial effects—people reported similar levels of negative affect, intrusions, and avoidance regardless of whether they had received a trigger warning. Moreover, these patterns were similar among people with a history of trauma. These results suggest a trigger warning is neither meaningfully helpful nor harmful.
  • Article
    Recent interest in the implicit self-esteem construct has led to the creation and use of several new assessment tools whose psychometric properties have not been fully explored. In this article, the authors investigated the reliability and validity of seven implicit self-esteem measures. The different implicit measures did not correlate with each other, and they correlated only weakly with measures of explicit self-esteem. Only some of the implicit measures demonstrated good test–retest reliabilities, and overall, the implicit measures were limited in their ability to predict our criterion variables. Finally, there was some evidence that implicit self-esteem measures are sensitive to context. The implications of these findings for the future of implicit self-esteem research are discussed.
  • Article
    According to the popular press, students have been increasingly demanding warnings before being exposed to potentially distressing classroom material. The validity of these types of trigger warnings has been a topic of vigorous debate. Based on a review of posttraumatic stress disorder (PTSD) research and closely related topics, this article answers questions that teachers might ask about the validity of the scientific assumptions behind trigger warnings and their use in the classroom. External stimuli causing distress is a feature common to many mental disorders, and trauma-based triggers of distress are an essential feature of PTSD. However, development of PTSD after a traumatic experience is relatively rare. Environmental triggers are often difficult to predict, but warnings may reduce distress among people with PTSD by allowing exposure to be controlled. To the extent that trigger warnings allow avoidance of hyperarousal when trying to learn, they should increase students’ classroom performance. However, avoidance of trauma reminders contributes to the persistence of PTSD symptoms. Although clinical research generally supports the notion of trigger warnings as an accommodation for individual students diagnosed with PTSD, the effectiveness of trigger warnings in the classroom is unknown. In addition, trigger warnings may be a legitimate accommodation for students with psychiatric disabilities, but this does not mean that they are relevant to nonclinical issues.
  • Article
    Recently, a heated debate has risen in Academia following numerous student initiatives petitions for the formal incorporation of rigger warnings in course syllabi. When contextualized within the intersecting politics of disability and feminist pedagogies, a number of fundamental contentions within this debate become apparent. First, grave misunderstandings remain regarding about practices of accommodation and the possibility of establishing the classroom as a “safe space.” Second, resistance within the academy to understand trauma as a pedagogical issue illustrate a failure to consider experiences of and responses to trauma as issues of disability (in)justice. Through an exploration of these issues, it becomes evident that the conflicting approaches to trauma in the classroom demand the more integrated, collaborative praxis of a “Feminist Disability Studies Pedagogy” (FDSP). When approached through this hybrid pedagogy, the conversation shifts from whether we should use trigger warnings, to why trauma itself is an imperative social justice issue within our classrooms.
  • Article
    Full-text available
    Crowdsourcing has had a dramatic impact on the speed and scale at which scientific research can be conducted. Clinical scientists have particularly benefited from readily available research study participants and streamlined recruiting and payment systems afforded by Amazon Mechanical Turk (MTurk), a popular labor market for crowdsourcing workers. MTurk has been used in this capacity for more than five years. The popularity and novelty of the platform have spurred numerous methodological investigations, making it the most studied nonprobability sample available to researchers. This article summarizes what is known about MTurk sample composition and data quality with an emphasis on findings relevant to clinical psychological research. It then addresses methodological issues with using MTurk-many of which are common to other nonprobability samples but unfamiliar to clinical science researchers-and suggests concrete steps to avoid these issues or minimize their impact. Expected final online publication date for the Annual Review of Clinical Psychology Volume 12 is March 28, 2016. Please see http://www.annualreviews.org/catalog/pubdates.aspx for revised estimates.
  • Article
    We examine the trade-offs associated with using Amazon.com's Mechanical Turk (MTurk) interface for subject recruitment. We first describe MTurk and its promise as a vehicle for performing low-cost and easy-to-field experiments. We then assess the internal and external validity of experiments performed using MTurk, employing a framework that can be used to evaluate other subject pools. We first investigate the characteristics of samples drawn from the MTurk population. We show that respondents recruited in this manner are often more representative of the U.S. population than in-person convenience samples-the modal sample in published experimental political science-but less representative than subjects in Internet-based panels or national probability samples. Finally, we replicate important published experimental work using MTurk samples. © The Author 2012. Published by Oxford University Press on behalf of the Society for Political Methodology. All rights reserved.
  • Article
    To ascertain the prevalence of posttraumatic stress disorder (PTSD) and risk factors associated with it, we studied a random sample of 1007 young adults from a large health maintenance organization in the Detroit, Mich, area. The lifetime prevalence of exposure to traumatic events was 39.1%. The rate of PTSD in those who were exposed was 23.6%, yielding a lifetime prevalence in the sample of 9.2%. Persons with PTSD were at increased risk for other psychiatric disorders; PTSD had stronger associations with anxiety and affective disorders than with substance abuse or dependence. Risk factors for exposure to traumatic events included low education, male sex, early conduct problems, extraversion, and family history of psychiatric disorder or substance problems. Risk factors for PTSD following exposure included early separation from parents, neuroticism, preexisting anxiety or depression, and family history of anxiety. Life-style differences associated with differential exposure to situations that have a high risk for traumatic events and personal predispositions to the PTSD effects of traumatic events might be responsible for a substantial part of PTSD in this population.
  • Article
    Full-text available
    Self-concept consistency and short-term stability were investigated in the United States, Australia, Mexico, Venezuela, Philippines, Malaysia, China, and Japan. Evidence for substantial cross-role consistency and reliable within-individual variability in trait self-perceptions were found in each culture. Participants in all cultures exhibited short-term stability in their self-reported traits within roles and moderately stable if–then patterns of trait self-perceptions. Cultural differences, which primarily involved Japan, were partially accounted for by cultural differences in dialecticism, but not self-construals or cultural tightness. In all cultures, satisfaction of needs in various roles partially accounted for within-individual variability in self-reported traits. The results provide support for integrating trait and cultural psychology perspectives, as well as structure and process approaches, in the study of self-concepts across cultures.