Does the effect of repeated exposure to fact-checks to combat misinformation vary over time? Evidence from a field experiment in the Global South

How to cite

Liu, Shelley; Croke, Kevin; Bowles, Jeremy; Larreguy, Horacio; Marshall, John. 2022. "Does the effect of repeated exposure to fact-checks to combat misinformation vary over time? Evidence from a field experiment in the Global South." AEA RCT Registry. AEARCTR-0009517. https://doi.org/10.1257/rct.9517-1.1

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

Shelley Liu* , Kevin Croke , Jeremy Bowles , Horacio Larreguy , John Marshall

* Primary investigator

Registry ID
AEARCTR-0009517
DOI
https://doi.org/10.1257/rct.9517-1.1
First registered
Last updated
Status
On going
Trial dates
2020-10-23 – 2023-09-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/9517

Abstract

Potentially harmful misinformation runs rampant on social media in many countries. We explore whether repeated exposure to fact-checks can inoculate citizens against misinformation by increasing their social media literacy. In the initial phase of this study, we partnered with Africa Check to deliver variants of a fact-checking podcast or text messages to around 4,500 recruited participants in South Africa twice a month for six months. This study follows up with the same set of participants to examine the longer term effects of this intervention, between 6 months and one year after the initial study had concluded. This follow-up study seeks to understand the longevity of fact-checking and inoculation efforts, with significant policy implications for combating misinformation. It also estimates the dynamic effects over three periods: after three months of exposure to treatment, after six months of exposure to treatment, and around nine months after treatment concluded.

Intervention

In the initial phase of this study, we partnered with Africa Check to deliver variants of a fact-checking podcast or text messages to around 4,500 recruited participants in South Africa twice a month for six months. This study follows up with the same set of participants to examine the longer term effects of this intervention, between 6 months and one year after the initial study had concluded.

Experimental design

Study participants are randomly assigned to either control or one treatment group. The treatments are distinguished along three dimensions: (1) mode of information delivery; (2) messaging encouraging information consumption; (3) whether participants are incentivized to take up the treatment.

Design details

Randomization method
We block randomize batches of participants into treatment conditions once every two weeks. We block on a set of variables including demographic characteristics, social media usage, attitudes towards different media sources, and knowledge regarding pieces of misinformation.
Randomization unit
Individual
Clusters
Information included in PAP.
Observations
Information included in PAP.
Arms
Information included in PAP.
Primary outcomes
We investigate the following groups of outcomes: 1. Increased knowledge of information covered by treatment delivery. 2. Knowledge and discernment of misinformation circulating online. 3. Trust in social media platforms 4. Consumption, verification, and sharing 5. Attitudes and behaviors towards COVID-19 and politics

Cited by

No citation data available for this registration yet.

Keywords