Empathy for the Devil: Can AI-Illustrated Vignettes Manipulate Affective Polarization and Democratic Values?

How to cite

Peterlevitz, Tiago. 2023. "Empathy for the Devil: Can AI-Illustrated Vignettes Manipulate Affective Polarization and Democratic Values?." AEA RCT Registry. AEARCTR-0012387. https://doi.org/10.1257/rct.12387-1.1

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

Tiago Peterlevitz*

* Primary investigator

Registry ID
AEARCTR-0012387
DOI
https://doi.org/10.1257/rct.12387-1.1
First registered
Last updated
Status
Completed
Trial dates
2023-11-01 – 2023-11-04
Countries
Brazil ()
Registry record
http://www.socialscienceregistry.org/trials/12387

Abstract

This is the pre-analysis plan for a survey experiment investigating how AI-illustrated vignettes emphasizing tolerance and counter-stereotypes can foster empathy toward political adversaries.

Intervention

The intervention is designed to prompt respondents to consider the support of counter-stereotypical individuals or groups for the out-group candidate.

Experimental design

I will use a two-condition (control vs. treatment) between-subjects design. Participants will indicate their political preferences and be categorized into three groups: Bolsonaristas, Lulistas, or neutral. Within each group, participants will be randomly assigned to either the treatment or control groups. Participants in the neutral group assigned to the treatment condition will randomly receive either the Bolsonaristas’ treatment or the Lulistas’ treatment.

Design details

Randomization method
Randomization will be conducted by the online survey platform.
Randomization unit
Participant (individual).
Clusters
1
Observations
800-1000
Arms
400-500 participants in control and 400-500 participants in treatment.
Primary outcomes
Affective Polarization, Social Distance Index, Free and Fair Elections Index, Corruption and Individual Rights Index, Political Violence Index
Secondary outcomes
Social Distance Variables: Discuss Politics, Employee Hiring, and Daughter’s Marriage; Free and Fair Elections variables: Reduce Public Transportation, Contest Election Results, and Increase Traffic Stops; Corruption and Individual Rights Variables: Disclose Classified Information, Extend Pretrial Detention, and Illegal Wiretapping; Political Violence Variables: Threaten Political Leader, Harass Out-Group Supporter, Violence for Political Goals, and Violence After Electoral Defeat.

Cited by

No citation data available for this registration yet.

Keywords