Impact of media coverage of experts and politicians

Last registered on August 11, 2025

View Trial History

Pre-Trial

Trial Information

General Information

Title

Impact of media coverage of experts and politicians

RCT ID

AEARCTR-0016484

Initial registration date

July 31, 2025

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published

August 04, 2025, 6:01 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated

August 11, 2025, 10:35 AM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Country

United Kingdom of Great Britain and Northern Ireland

Region

Primary Investigator

Name

Clement Minaudier

Affiliation

City, University of London

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Greg Chih-Hsin Sheen

PI Affiliation

Academia Sinica

Contact Investigator

Additional Trial Information

Status

In development

Start date

2025-08-05

End date

2025-08-26

Keywords

Electoral, Governance

Additional Keywords

Media consumption, information acquisition

JEL code(s)

Secondary IDs

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

This project seeks to understand how individuals acquire information online after being exposed to media reports. When reporting on policy debates, the media often chooses to cite different sources of information: some of the information comes from elected politicians who may be presenting their party's position on the policy, while some of the information comes from experts who may have deeper knowledge about the topic. Our project aims to answer three interrelated questions: (1) does the use of different sources of information (politician vs. expert) in the media affect readers' attitudes towards these policies, (2) does the use of different sources of information affect readers' decision to acquire further information, and (3) does the use of different sources of information affect the type of information that readers seek to acquire? To answer these questions, we will randomly present one of two excerpts from an online news media article to participants. While both excerpts cover the same policy topic, one excerpt mostly contains quotes from experts, while the other mostly contains quotes from politicians. We will then measure how exposure to either article affects the respondent's policy attitude, desire to acquire more information, and source of further information.

External Link(s)

Registration Citation

Citation

Chih-Hsin Sheen, Greg and Clement Minaudier. 2025. "Impact of media coverage of experts and politicians." AEA RCT Registry. August 11. https://doi.org/10.1257/rct.16484-2.0

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

We will conduct a survey experiment in which we randomly present one of two excerpts from an online news media article to participants.
While both excerpts cover the same policy topic, one excerpt mostly contains quotes from experts, while the other mostly contains quotes from politicians.

Intervention (Hidden)

Intervention Start Date

2025-08-05

Intervention End Date

2025-08-26

Primary Outcomes

Primary Outcomes (end points)

Decision to obtain further online information and source of that information.

Primary Outcomes (explanation)

We will measure the decision to obtain further information based on the respondent's answer to the question "How likely would you be to seek further information on this topic?" and the source of information based on the respondent's actual click on a list of information sources.

Secondary Outcomes

Secondary Outcomes (end points)

Support for the policy and perception of the political nature of the policy.

Secondary Outcomes (explanation)

We will measure the respondent's support for the policy based on their response to the question "Do you agree that [INSERT POLICY DEBATE]" and using the donations they authorise us to make on their behalf to some selected charities. We will measure their perception of the nature of the policy based on their response to the question "Do you believe that policies about [INSERT POLICY TOPIC] should be determined by the UK parliament, a referendum, or by a committee of experts?".

Experimental Design

We use a between-subject design: we compare the responses to a subset of questions between participants assigned to the relevant treatment vignettes and those in the control group, or across participants assigned to different treatment vignettes, depending on the hypothesis.

Experimental Design Details

We use a survey experiment in which we randomly assign subjects to two different newspaper excerpts where we vary whether the quotes used in the excerpt are from politicians or from experts. We repeat this design across three topics: (1) A proposal to ban smartphones, (2) The UK government's defence spending, and (3) Policy approaches to address illegal Channel crossings.

To increase external validity, we use excerpts from actual articles posted on the BBC News website, a popular news media perceived as trustworthy and not biased towards any of the major political parties in the UK. The excerpts have been slightly edited by the researchers to emphasise the use of different sources across the two excerpts while keeping their messages comparable. We identify the BBC news source by showing the name and logo to increase the similarities between the experimental conditions and real-life media consumption. The politicians are elected members of parliament while experts are either civil servants, scientists, or charity leaders. The title of each source is stated in the excerpt to ensure that respondents can identify whether the source is a politician or an expert.

Respondents are allocated across two possible treatment arms: a newspaper excerpt with expert quotes or a newspaper excerpt with politician quotes. In addition, one group of respondents is assigned to a pure control arm who sees no information at all. We use a between-subject design: we compare the responses to a subset of questions between participants assigned to the relevant treatment vignettes and those in the control group, or across participants assigned to different treatment vignettes, depending on the hypothesis.

Randomization Method

Randomisation will be carried out by an algorithm on the Qualtrics platform.

Randomization Unit

Randomisation is carried out at the individual level.

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

2500

Sample size: planned number of observations

2500

Sample size (or number of clusters) by treatment arms

1000 individuals in the expert treatment, 1000 individuals in the political treatment, and 500 individuals in the control group.

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

The minimum detectable effect with 80% power and 95% confidence interval is 0.06 on a binary indicator of whether the respondent wants to acquire further information. This corresponds to a 10% increase relative to a control mean of 0.60 given a standard deviation of 0.48 (calibrated based on an exploratory survey).

Supporting Documents and Materials

IRB