What can we learn from survey experiments?

How to cite

BRISCESE, GUGLIELMO; Grignani, Maddalena. 2024. "What can we learn from survey experiments?." AEA RCT Registry. AEARCTR-0014161. https://doi.org/10.1257/rct.14161-1.0

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

GUGLIELMO BRISCESE* , Maddalena Grignani

* Primary investigator

Registry ID
AEARCTR-0014161
DOI
https://doi.org/10.1257/rct.14161-1.0
First registered
Last updated
Status
In development
Trial dates
2024-08-12 – 2024-09-18
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/14161

Abstract

There is a growing interest in understanding how subjective beliefs shape behaviors. Survey experiments, which randomly vary exposure to information treatments, aim to provide causal evidence of this link, but the methodology used by existing studies might suffer from limitations to its external validity. This study employs a survey experiment to address this gap by comparing the effects of 'forced' versus 'self-selected' information exposure on beliefs and behavioral intentions. This study is concerned with understanding whether giving participants the option to choose information consumption (rather than being forced to consuming it) can enhance external validity and reduce social desirability bias. The study aims to (i) provide evidence on the willingness to consume belief-altering information when not coerced, (ii) offer methodological insights to enhance the validity of survey experiments, and (iii) explore how information consumption influences trust in institutions and overall societal welfare.

Intervention

This study employs a survey experiment where respondents are randomly assigned to different information treatments.

Experimental design

Survey respondents are automatically randomly assigned to a control or one of two treatment groups.

Design details

Randomization method
Automatic randomization done by survey software.
Randomization unit
Individual respondent level.
Clusters
0
Observations
Approximately 1600 respondents.
Arms
Approximately 500 respondents per randomized group.
Primary outcomes
Information consumption, knowledge accuracy, and trust in institutions.

Cited by

No citation data available for this registration yet.

Keywords