Political Identities, Information Demand and Processing (Wave 2)

How to cite

Chen, Yan; Kosfeld, Michael; Hett, Florian; Bauer, Kevin. 2021. "Political Identities, Information Demand and Processing (Wave 2)." AEA RCT Registry. AEARCTR-0007106. https://doi.org/10.1257/rct.7106-1.0

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

Yan Chen* , Michael Kosfeld , Florian Hett , Kevin Bauer

* Primary investigator

Registry ID
AEARCTR-0007106
DOI
https://doi.org/10.1257/rct.7106-1.0
First registered
Last updated
Status
In development
Trial dates
2021-01-30 – 2021-02-28
Countries
United States of America ()
Registry record
http://www.socialscienceregistry.org/trials/7106

Abstract

This study constitutes the second wave of our panel study on the role of social identity and ``groupiness'' in the demand for and processing of information using an online experiment with a representative sample of the US population. The first wave was deployed the week before the 2020 US presidential election, where 1,004 participants played a set of games and predicted both the outcomes of the 2020 presidential election and policy-sensitive statistics a year after the election conditional on its outcome. In Wave 2, we will explore the mechanisms behind participants' biased information demand and processing uncovered in Wave 1. Specifically, participants will again play the other-other allocation games as well as a variant of ``guessing which cup is used'', which measures the individual ingroup bias in information demand and processing in a neutral domain. We also administer a short survey to evaluate whether the election outcomes have led to any perceptions of status change among the Democrats and Republicans. In a treatment variation, we test whether informing participants that the quality of information is independent of party labels causes a reduction of ingroup bias in information demand and processing.

Intervention

We employ a between-subject design that varies whether or not a participant sees the proportion of correct guesses of Democrats and Republicans in Wave 1. After submitting their answers, participants in the control condition are not given the correct answers, whereas those in the treatment are given the correct answers.

Experimental design

We study the role of social identity and ``groupiness'' (the individual tendency to display in-group bias in allocation decisions) in the processing of and demand for information using an online experiment with a representative sample of the US population. The first wave of the study was deployed the week before the 2020 US presidential election (AEARCTR-0006670), where 1,004 participants played a set of games and predicted both the outcomes of the 2020 presidential election and policy-sensitive statistics a year after the election conditional on its outcome. The current study constitutes the second wave of our panel study on the role of social identity and ``groupiness'' in the processing of and demand for information. In Wave 2, we will explore the mechanisms behind participants' biased information demand and processing uncovered in Wave 1. Specifically, participants will play the other-other allocation games again, and will be given an opportunity to update their predictions on the same policy-sensitive statistics. Based on a simple version of a social learning experiment we check how subjects update and demand for information when given additional information with and without others' party labels. We also plan to administer a short survey to evaluate whether the election outcomes have led to any perception of status change among the Democrats and Republicans. In a treatment variation, we test whether informing participants that information processing accuracy is independent of party labels causes a reduction of ingroup bias in information demand and processing.

Design details

Randomization method
Randomization will be done by a computer.
Randomization unit
Individual participant
Clusters
1,004 individuals
Observations
1,004 individuals
Arms
500 participants in control, 504 in treatment condition.
Primary outcomes
1. (Partisanship) Answers to the Political Survey to classify individuals as either Democrat-leaning or Republican-leaning. 2. (Groupiness new) Difference in money allocated to members of different groups, both in the minimal and the political identities allocation tasks. 3. (Strength of social learning) Relative strength of weight put on guesses of others in comparison to weight put on own signals when guessing the right cup. 4. (Partisan bias in processing in a neutral domain) Difference in updating weight attached to guesses from political ingroups and outgroups. 5. (Partisan bias in demand in a neutral domain) Difference in the implied willingness-to-pay for guesses from political ingroups and outgroups.

Cited by

No citation data available for this registration yet.

Keywords