Affect labeling and political polarization

How to cite

Hemsley, Pedro; Pavão, Lynda. 2024. "Affect labeling and political polarization." AEA RCT Registry. AEARCTR-0013305. https://doi.org/10.1257/rct.13305-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Pedro Hemsley* , Lynda Pavão

* Primary investigator

Registry ID
AEARCTR-0013305
DOI
https://doi.org/10.1257/rct.13305-1.0
First registered
Last updated
Status
Completed
Trial dates
2021-09-23 – 2021-09-24
Countries
United States of America ()
JEL codes
D91
Registry record
http://www.socialscienceregistry.org/trials/13305

Abstract

Affect labelling, the process of putting feelings into words, has been shown to have a calming effect on the brain. This study examines the impact of affect labelling on political polarization over Covid-mitigation measures, an emotionally-charged topic that led to large differences of opinion. We conducted an online experiment in which participants in a treatment group, randomly assigned, underwent a questionnaire to label their feelings before reporting their opinions. We found that affect labeling reduced polarization for young individuals, up to the age of 29, a demographic characterized by intense emotional responses due to increased activity in the amygdala, the region of the brain whose activity is reduced by affect labeling.

Intervention

We assigned randomly two treatments to participants in this online experiment. The first one was an affect labeling questionnaire: during the experiment, subjects answered this questionnaire as a way to express emotions freely (that is, perform affect labeling). The second one was the presentation of a short text with referenced information about Covid-19 mitigation measures. Then all subjects stated their opinions about these measures. Participation was entirely online and whole process took less than five minutes.

Experimental design

We conducted an online survey on September 23, 2021, with 600 participants from the United States, recruited from Amazon’s Mechanical Turk. The reward for participation was set at US$0.45. Respondents were not permitted to participate more than once and we only accepted workers who had previous experience and had at least a 90 percent approval rating. The questionnaire design and data collection from respondents were conducted through the Qualtrics platform.

The study began with participants answering socio-demographic questions via an online link. Next, they were informed that we could ask questions about the Covid-19 pandemic, an emotionally-charged topic that led to major differences of opinion. Participants were then randomly split into four groups: one control group and three treatment groups.

The first treatment group underwent an affect labeling questionnaire, designed for unrestricted emotional expression. This questionnaire combined items from various scales that assess emotional states (detailed in the appendix). The objective is to give participants a structured framework to identify and express their feelings, with simple implementation, taking around one minute to complete.

The second treatment group received referenced information about the effectiveness of social distancing and mask-wearing, including hyperlinked sources for reference. The third treatment group received both the referenced information and completed the affect labeling questionnaire, while the control group received neither.

Then all groups were directed to the concluding questions of the questionnaire and rated their agreement or disagreement on the efficacy of measures to control the spread of COVID-19 on a scale of -8 to 8. We choose this scale to have a natural center: zero.

Specifically, participants gave their opinions about the following statements:

1- Social distancing is a relevant measure to limit the spread of the coronavirus.
2- Wearing masks is a relevant measure to limit the spread of the coronavirus.

Design details

Randomization method
Randomization done online through the Qualtrics platform
Randomization unit
Individual participant in an online experiment
Clusters
600 individuals
Observations
600 individuals
Arms
146 participants in treatment 1 only (affect labeling) 150 participants in treatment 2 only (referenced information) 150 participants in treatments 1 and 2 150 participants in the control group
Primary outcomes
Political polarization, measured as the absolute value of the distance from a participant's opinion about Covid-mitigation policies to a neutral position (zero) in a scale from -8 to 8.

Cited by

No citation data available for this registration yet.

Keywords