Everyone’s Doing It: Affective Polarization is Inflated by Social Pressure

Download data and study materials from OSF

Principal investigator:

Elizabeth C. Connors

University of South Carolina

Email: connors4@mailbox.sc.edu

Homepage: http://www.elizabethchaseconnors.com/

Sample size: 3333

Field period: 02/07/2020-04/16/2020

American politics today is affectively polarized—partisans report disliking and distrusting out-partisans while liking and trusting in-partisans. In this study I examined if this climate could encourage partisans to report higher levels of affective polarization and tested this with a survey experiment. My findings demonstrated that partisan social pressure to report affective polarization could inflate survey responses, suggesting implications for the measurement of polarization, greater nuance in our understanding of polarization, and the potential for a “snowball effect” of political climates—where a climate of polarization could beget more polarization. Future work is needed for a broader understanding of how social context shapes expressions of partisanship.
I expected that altering the privacy respondents perceive they have will also change self-reports of affective polarization, where reports of affective polarization should be highest in the most public settings and lowest in the most private settings—and that this is also moderated by self-monitoring (or individual susceptibility to social pressure), where the change in privacy should be increasingly effective at altering reports of affective polarization as self-monitoring increases.
Experimental Manipulations
The audience. Participants were randomly assigned to one of four conditions: 1) a public condition, where they were told: “Just a reminder, the results based on your responses may be published”; 2) a private condition, where they were told: “Just a reminder, your responses are completely private”; 3) a friends condition, where they were told: "Imagine your friends will read your responses to the following questions"; or 4) a control condition, where they were told nothing.
Affective polarization (measured by feeling thermometers and trust) and issue polarization on climate change (which was largely exploratory).
Summary of Results

There were no main treatment effects nor a moderated treatment effect of the public treatment and self-monitoring on affective polarization. There was the predicted moderated treatment effect of the private treatment and self-monitoring on affective polarization as well as this same trend for the friends treatment (a treatment which was largely exploratory).

There was no treatment effect nor a moderated treatment effect on issue polarization as measured with the climate change question, although self-monitoring was negatively correlated with saying that climate change is not real (i.e., those who care about impressing others were more likely to say climate change is real).

Additional Information
This project is still in progress.