Information Bundling and Polarizing Persuasion

How to cite

Bonomi, Giampaolo. 2024. "Information Bundling and Polarizing Persuasion." AEA RCT Registry. AEARCTR-0012729. https://doi.org/10.1257/rct.12729-2.1

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

Giampaolo Bonomi*

* Primary investigator

Registry ID
AEARCTR-0012729
DOI
https://doi.org/10.1257/rct.12729-2.1
First registered
Last updated
Status
Completed
Trial dates
2024-02-15 – 2024-09-01
Countries
United States of America ()
Registry record
http://www.socialscienceregistry.org/trials/12729

Abstract

We study how bundling together policy recommendations on different policy issues affects voters' policy views. Voters are randomized into (i) treatment messages, each consisting of policy recommendations on two policy issues, and (ii) control messages, where policy recommendations are sent separately. The issues bundled differ in ideological value, policy domain, and complexity. We investigate the presence of belief spillovers across policy domains, and the role played by trust and identity in explaining these spillovers.

Experimental design

The experiment is implemented online on Amazon Mechanical Turk via Cloud Research. Before randomization, we collect the socio-demographic characteristics of respondents, as well as their perceived importance and knowledge of the policy issues used as main outcomes. Participants are then randomized into (i) treatment messages, each consisting of written policy recommendation on two policy issues, and (ii) control messages, where the same policy recommendations are reported separately (respondents observe one recommendation at a time). The issues bundled differ in ideological value, policy domain, and complexity. Before and after reading the messages, respondents are asked to provide their policy preference for the two issues (in the main experiment, one preference is asked before treatment while the other is asked after treatment).

Design details

Randomization method
Randomization performed through Qualtrics' block randomization option.
Randomization unit
Individual randomization
Clusters
450-500 subjects for each treatment arm. Main experiment: 5,000 individuals for the main policy pair (abortion and trade) For other policy pairs, the main treatments will be implemented, up to an (estimated) total of 12,000 participants taking part in the research study.
Observations
450-500 subjects for each treatment arms
Arms
For each pair of proposals on issues X and Y, an equal number of respondents will be allocated to the following three conditions: (i) X and Y made by the same source (ii) X and Y sent separately (i.e., not sent together by the same source). (iii) no message shown. Subjects will be assigned evenly to conditions. We plan to assign 450-500 subjects to each condition.
Primary outcomes
Policy preferences on economic policy issues (trade policy, healthcare regulations, and redistributive policies) and social policy issues (abortion policy, affirmative action, LGBTQ+ rights, etc)

Cited by

No citation data available for this registration yet.

Keywords