Evidence and Influence

How to cite

Trachtman, Hannah; Serra-Garcia, Marta. 2026. "Evidence and Influence." AEA RCT Registry. AEARCTR-0017882. https://doi.org/10.1257/rct.17882-4.0

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

Hannah Trachtman* , Marta Serra-Garcia

* Primary investigator

Registry ID
AEARCTR-0017882
DOI
https://doi.org/10.1257/rct.17882-4.0
First registered
Last updated
Status
In development
Trial dates
2026-02-13 – 2026-12-31
Countries
United States of America ()
Registry record
http://www.socialscienceregistry.org/trials/17882

Abstract

We investigate the attention and influence that scientific evidence receives in online settings. In an online experiment, participants in the role of "senders" view sets of articles based on research studies, and decide which articles to share and what content to provide alongside the article. We vary features of the articles and the communication process to shed light on the drivers of online attention to different types of evidence. In a separate experiment, participants in the role of "receivers" will choose which articles to click on to read more about.

Intervention

Participants in the role of sender are presented with a set of articles (press-releases), each of which summarizes the evidence provided in a research study.

Participants in the role of receiver are presented with similar set of articles, accompanied by captions written by senders.

Experimental design

Participants in the role of sender are presented with a set of articles (press-releases), each of which summarizes the evidence provided in a research study. Their task is to help disseminate research findings for a broad population and their advice is sought on (a) what content to feature and (b) how to describe this content.

The sender experiment will consist of three main treatments. First, in the Baseline treatment, senders will view four articles, each on a different topic. The type of study will be randomized across topics. The articles will also vary in the certainty of the findings that they present, according to the abstract. To study the mechanisms behind senders’ decisions, we will conduct two additional treatments. In one treatment, we will vary how some study types are framed (or described) in the articles. In a second treatment, we will reduce senders’ ”cost” of communicating about evidence by providing them with the caption for any article they choose to share.

Participants in the role of receiver are presented with similar set of articles, accompanied by captions written by senders.

Design details

Randomization method
Qualtrics randomization (Prolific), randomization done in office (freelancers)
Randomization unit
Individual
Clusters
300 freelancers (or as many as possible) 1600 professionals on Prolific (or as many as possible) 3000 participants on Prolific
Observations
300 freelancers (or as many as possible) 1600 professionals on Prolific (or as many as possible) 3000 participants on Prolific
Arms
Freelancers: 100 per treatment group Prolific professionals: 533 per treatment group Prolific participants: 1000 per treatment group
Primary outcomes
In the sender experiment, the primary outcome is which article was selected to be posted. In the receiver experiment, the primary outcome is which article was clicked to be read.
Secondary outcomes
In the sender experiment, secondary outcomes include features of the captions written by senders.

Cited by

No citation data available for this registration yet.

Keywords