Political Identities, Information Demand and Processing

How to cite

Chen, Yan; Kosfeld, Michael; Hett, Florian; Bauer, Kevin. 2020. "Political Identities, Information Demand and Processing." AEA RCT Registry. AEARCTR-0006670. https://doi.org/10.1257/rct.6670-1.0

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-0006670
DOI
https://doi.org/10.1257/rct.6670-1.0
First registered
Last updated
Status
In development
Trial dates
2020-10-27 – 2020-11-03
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/6670

Abstract

We study the role of social identity and "groupiness'' in the processing of and demand for information using an online experiment with a representative sample of the US population, deployed the week before the upcoming 2020 US presidential election on November 3. Specifically, we ask whether (i) people update their beliefs in reaction to consuming news in a biased way, (ii) whether this bias increases if news source labels are observable, (iii) whether groupiness as measured by the other-other allocation game moderates both these effects, and (iv) whether exposure to news articles from a neutral source or the other side increases prediction accuracies. In our online experiment, approximately 1,000 participants have to predict both the outcome of the 2020 presidential election and policy-sensitive statistics a year after the election conditional on its outcome. To update their initial predictions, individuals receive or can select articles on the respective topics from different news sources. Our key experimental variation is whether the articles come with a label of the news sources or not, where news sources differ in their political leaning.

Intervention

In our online experiment, participants have to predict both the outcome of the 2020 presidential election and policy-sensitive statistics a year after the election conditional on its outcome. To update their initial predictions, individuals receive or can select articles on the respective topics from different news sources. Our key experimental variation is whether the articles come with a label of the news sources or not, where news sources differ in their overlap with the participants' political identity.

The main experiment will be implemented the week before the 2020 presidential election. We will contact the participants again in the last week of January 2021 to pay them for the accuracy of their predictions of the election outcome, as well as the last week of October 2021 to pay them for their predictions of policy-relevant statistics.

Experimental design

1. In stage 1, we elicit a groupiness measure using the minimal group paradigm, priors for two prediction tasks, and participants' political identities.

2. In stage P (information processing), we examine how labels of news sources influence the processing of information.

3. In stage D (information demand) we test how labels affect the demand for information, and end by measuring their updating tendencies for neutral information.

Design details

Randomization method
randomization done online by a computer
Randomization unit
individual participant
Clusters
1,000 participants
Observations
1,000 participants
Arms
500 participants in the control condition and 500 in the treatment condition
Primary outcomes
1. (Prior beliefs) Differences in the answers in the prediction tasks between the two different election outcomes before consuming additional news. 2. (Posterior beliefs) Differences in the answers in the prediction tasks between the two different election outcomes after consuming additional news. 3. (Change of beliefs) Change of the differences in the answers in the prediction tasks between the two different election outcomes before and after consuming additional news. 4. (WTP for partisan news) Money spent to affect the probabilities in the news demand task. 5. (Prediction accuracy) A dummy variable which equals one if a prediction is correct, and zero otherwise.

Cited by

No citation data available for this registration yet.

Keywords