The persistent effect of narratives: additional mechanism and long-term effects

How to cite

Liu, Manwei; Zhang, Sili. 2024. "The persistent effect of narratives: additional mechanism and long-term effects." AEA RCT Registry. AEARCTR-0013582. https://doi.org/10.1257/rct.13582-1.0

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

Manwei Liu* , Sili Zhang

* Primary investigator

Registry ID
AEARCTR-0013582
DOI
https://doi.org/10.1257/rct.13582-1.0
First registered
Last updated
Status
In development
Trial dates
2024-05-09 – 2024-05-24
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/13582

Abstract

This study is a follow-up of the project “The persistent effect of biased narratives: Evidence from an online experiment”, preregistered in the AEA RCT registry (AEARCTR-0005595). The current study adds five new treatments (between subjects), plus one auxiliary treatment, to the original study to explore additional mechanisms for why narratives have persistent effects and whether the persistent effects of narratives are long-lasting.

Intervention

First, in Phase I, we vary whether subjects are exposed to one of the two narratives (pro-GMO or anti-GMO) first or both narratives (in random order). In Phase II, we vary whether subjects are immediately provided with the exact narratives from the other side that they did not see in Phase I, or they are provided with more arguments to counteract the influence of the Phase I narratives, which is scheduled two weeks after Phase I.

Experimental design

In the two mechanism treatments, Pro.Nrt and Anti.Nrt, subjects are randomly exposed to one of the two narratives (pro-GMO or anti-GMO) in Phase I. In a third treatment, Both.Nrt as a benchmark for this new set of treatments, subjects are exposed to both narratives (in random order). In Phase II, subjects in Pro.Nrt and Anti.Nrt are provided with the exact narrative from the other side that they did not see in Phase I.

Two additional treatments, Pro.Long, Anti.Long, are identical to the Exogenous treatments in the original study except that the two phases are two weeks apart. They are added to test whether the persistent effects of narratives are long-lasting. In Phase I (the first session), subjects are randomly exposed to one of the two narratives (pro-GMO or anti-GMO). Phase II (the second session), where subjects are provided with more information to counteract the influence of the Phase I narratives, is scheduled two weeks after the first session.

We implement one auxiliary treatment, Both.Long, in which subjects' average donation functions as the target of estimation of Pro.Long and Anti.Long.

Design details

Randomization method
By a computer embedded in the program. We employ stratified randomization based on subjects' self-reported prior attitude toward the issue.
Randomization unit
Individual
Clusters
800 individuals.
Observations
800 individuals.
Arms
150 per treatment group in the five main treatment groups and 50 in the auxiliary benchmark treatment group.
Primary outcomes
(1) Self-reported attitude on whether the field practice takes technology too far or is appropriate usage of technology. (2) Allocation of a donation of 200 dollars between two organizations that favor two different sides.
Secondary outcomes
Evaluation of second-stage information, i.e. narratives or arguments.

Cited by

No citation data available for this registration yet.

Keywords