Work Description

Date: 15 May, 2020

Dataset Title: Data and Syntax for Trigger warnings as an interpersonal emotion-regulation tool

Dataset Creators: Gainsburg, Izzy; Earl, Allison

Dataset Contact: Izzy Gainsburg, izzyg@umich.edu

Funding: This research was supported in part by the Department of Psychology and Rackham Graduate School at the University of Michigan.

Disclaimer:
As its title suggests, our dataset contains some content (notably a fictional essay portraying domestic violence in Study 3) that is emotionally loaded and that some users may find difficult to read.

Key Points:
- Study 1 demonstrated that anticipated anxiety for warned-of content predicts intentions to avoid information. Furthermore, beliefs about trigger warnings as protective (versus coddling) best predicted anticipated anxiety for warned-of content and subsequent intentions to avoid.

-In Study 2, participants had higher anticipated negative affect for videos with trigger warnings, compared to those without, and this mediated increased avoidance for warned-of videos.

-In Study 3, trigger warnings preceding essays increased anticipated negative affect and attentional-regulation strategies, but reduced experiences of negative affect.

Research Overview:
This research explores trigger warnings as an interpersonal emotion-regulation strategy, introducing a temporal dimension to interpersonal emotion-regulation by regulating others' future, anticipated emotions.

Across studies, believing that trigger warnings are protective (versus coddling) increased their effect on anticipated negative affect, but weakened their effect on experienced negative affect.

Methodology (see associated manuscript for full methods):
Study 1
120 individuals from across the United States were recruited through Amazon's Mechanical Turk (MTurk); an additional 28 additional participants signed up but did not finish the survey. Of these 148 individuals, 4 were excluded on account of English being their second language and 64 participants dropped out of the survey after informed consent but prior to answering any questions, resulting in a final sample of 80 participants.

Participants were told that the survey examined “experiences, beliefs, and feelings about trigger warnings.” See attached survey instrument for full questions asked of participants.

Study 2
240 individuals from across the United States were recruited through MTurk. An additional 46 participants provided usable data despite not finishing the survey and 10 participants were excluded on the basis of not selecting a video to watch. Sample size was determined a priori and analyses were conducted only after the data collection ceased.

Participants were told the study was about “people's thoughts and feelings around videos that are available on the internet.” Participants were shown two fictional video titles and asked to choose one to watch. The titles were chosen from a pilot test of eight original fictional titles among 30 participants to determine two titles similar across 14 dimensions known to influence approach and avoidance. See attached survey instrument for full questions asked of participants. Participants never watched the videos.

Study 3
720 individuals were recruited through MTurk. An additional 290 additional people signed up despite not finishing the survey. Of these 1010 participants, 23 non-native English-speakers and 8 people that did not consent were excluded, leaving 979 participants. All analyses were conducted only after the data collection ceased.

Participants read an essay involving domestic violence developed by the authors and a research assistant. Participants read one of two versions of the essay, which were identical excepting three moments that varied in severity. Two versions of the essay were used to test whether effects of the trigger warnings would generalize to content of different intensity that could credibly have trigger warnings. We did not use content so mild such that trigger warnings would not provide any emotional benefit (i.e., there are no negative emotions to be reduced), or content that is so severe that it would violate ethical standards for the present research context.

Study 3 had a control condition and two different trigger warning conditions. One trigger warning condition merely warned that content might be distressing (“Trigger Warning: This article contains distressing content,” i.e., Trigger Warning Only) and the other alerted participants to the specific subject matter that might be distressing (“Trigger Warning: This article contains content around domestic violence,” i.e., Trigger Warning with Content). The Trigger Warning Only condition, like the control condition, made no mention of specific subject matter, thus making it a more internally valid manipulation. On the other hand, the Trigger Warning with Content condition was higher in external validity, as real-life trigger warnings often occur in the context of recipients being aware of the potentially distressing subject matter.

Furthermore, there is reason to think the two warning conditions could have different effects on anticipated negative affect, attention-regulation, and experienced negative affect. Study 3 was designed to test competing hypotheses on which type of warning might have a stronger effect on these factors.

Instrument and/or Software specifications: Data were collected using MTurk and processed using SPSS. Modeling was conducted using the Lavaan package for R statistical software (Rosseel, Y. Lavaan: An R package for structural equation modeling
Journal of Statistical Software, 48 (2) (2012), pp. 1-36).

Files contained here:
See Table_of_documents.xlsx file

Related publication(s):
Gainsburg, I., & Earl, A. (2018). Trigger warnings as an interpersonal emotion-regulation tool: Avoidance, attention, and affect depend on beliefs. Journal of Experimental Social Psychology, 79, 252–263. https://doi.org/10.1016/j.jesp.2018.08.006

Use and Access:
This data set is made available under a Creative Commons Attribution-NonCommercial 4.0 International license (CC BY-NC 4.0).

To Cite Data:
Gainsburg, I., Earl, A. Data and Syntax for Gainsburg & Earl (2018) in JESP [Data set]. University of Michigan - Deep Blue. https://doi.org/10.7302/xsjc-1k55