Polarized Media and Inflation Expectations

Last registered on November 07, 2025

View Trial History

Pre-Trial

Trial Information

General Information

Title

Polarized Media and Inflation Expectations

RCT ID

AEARCTR-0017085

Initial registration date

November 03, 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

November 07, 2025, 7:35 AM EST

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

Locations

Country

Turkey

Region

Primary Investigator

Name

Orkun Saka

Affiliation

City St George's, University of London

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Selva Demiralp

PI Affiliation

Koc University

Contact Investigator

PI Name

Michael Weber

PI Affiliation

Purdue University

Contact Investigator

Additional Trial Information

Status

In development

Start date

2025-11-04

End date

2026-01-31

Keywords

Behavior, Finance & Microfinance, Other

Additional Keywords

Macro-finance, Inflation, Media, Politics

JEL code(s)

Secondary IDs

Prior work

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

Abstract

We aim to study how exposure to inflation-related prime-time news on TV may shape households' inflation expectations. For this purpose, we undertake a survey experiment in which households are first asked a basic set of demographic questions alongside their preferred TV channel for watching prime-time evening news, inflation perceptions and expectations. After that, each household is randomly assigned to a treatment arm that differs both in terms of the political affiliation of the TV channel (gov't-friendly, neutral or opposition) whose logo is visible on the treatment page and the type of inflation narrative (positive or negative) that the treatment page contains in written form. There are also two randomly-assigned control groups, one of which gets exposure to a placebo (i.e., oral-health-related) news and the other skips the treatment stage (hence, gets exposed to nothing). After treatment, we re-elicit households' inflation expectations by using a differently formatted question (compared to the earlier pre-treatment question). We also ask questions regarding whether the respondent has heard the news before, how much they trust TV news, how much support they have for government's economic policies, their voting preferences as well as forward-looking consumption/saving/expenditure behaviour.

We will explore the causal effect of each treatment arm on the post-treatment questions (taking either of the two control groups as the baseline). We will also explore the heterogeneity in causal effects by the type of TV channel that the respondent normally prefers to watch for prime-time news.

External Link(s)

Registration Citation

Citation

Demiralp, Selva, Orkun Saka and Michael Weber. 2025. "Polarized Media and Inflation Expectations." AEA RCT Registry. November 07. https://doi.org/10.1257/rct.17085-1.0

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

Intervention (Hidden)

Intervention Start Date

2025-11-04

Intervention End Date

2025-11-30

Primary Outcomes

Primary Outcomes (end points)

Households' inflation expectations, trust in TV channels, support for government's economic policies, voting preferences, forward-looking consumption/saving/expenditure behaviour.

Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)

Secondary Outcomes (explanation)

Experimental Design

Through an online survey company who has access to a large panel of existing respondents, we will invite a nationally-representative sample of individuals to our survey. The survey company will aim to make the sample representative by imposing quotas on key socio-economic characteristics of the collected sample and deny those from entering the survey if the relevant quotas are already met.

The survey begins by asking respondents a set of questions on their socio-demographic characteristics alongside channel preferences for prime-time TV news. Through a question assessing attention level, we then eliminate those who are not paying sufficient attention to the survey. After that, we move on to elicit prior beliefs about past inflation in the country (i.e., past 12 months) as well as expectations about its future trajectory (i.e., next 12 months).

The next stage includes the "TV channel x Narrative" intervention, with the aim of investigating how the intervention could shift the respondents' inflation expectations. For this purpose, we randomly divide the total sample into 8 treatment arms as in the following:

T1. Government-friendly Channel x Positive Inflation News Narrative
T2. Neutral Channel x Positive Inflation News Narrative
T3. Neutral Channel x Negative Inflation News Narrative
T4. Opposition Channel x Negative Inflation News Narrative
T5. Positive Inflation News Narrative
T6. Negative Inflation News Narrative
T7. Placebo News Control
T8. No News Control

Each treatment arm that contains a channel (i.e., T1, T2, T3 and T4) will include the corresponding TV channel's logo before the written narrative that the respondents will read. For these treatment arms, we will also randomly vary between two different TV channels both of which fall into the same category to make sure that our results are not driven by a specific channel. For the gov't-friendly and opposition interventions (T1 and T4), we use either only positive or only negative narratives respectively to make the intervention credible in the eyes of the respondents. T5 and T6 provide only narratives (without TV channel logos) which we will use to disentangle the causal effect of the TV channel separately from the type of narrative it uses. T7 and T8 will constitute two separate control groups, one with a placebo news item (related to an oral-health-related innovation) and another which will entirely skip the treatment stage. Including and comparing the results between these two control arms will help us understand whether our results might be driven simply by reading any news item on the screen.

In the post-treatment stage, we will re-elicit respondents' inflation expectations by using a differently formatted question (compared to the earlier pre-treatment question). Our main variable of interest will be the difference between prior and posterior inflation expectations that the respondents state respectively before and after the treatment. We will also analyze post-treatment questions on whether the respondent has heard the news before, how much they trust the TV news, how much support they have for government's economic policies, their current voting preferences as well as forward-looking consumption/saving/expenditure behaviour.

We will also explore the heterogeneity in these results especially with respect to ex ante TV channel preferences of the respondents, which will link this specific survey to a previous analysis we have conducted on this topic by using observational data.

Experimental Design Details

Randomization Method

Randomization is performed by a computer that assigns incoming respondents to different treatment arms.

Randomization Unit

Individual

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

N/A

Sample size: planned number of observations

Around 8,000

Sample size (or number of clusters) by treatment arms

Approximately:
- Gov't-friendly Channel x Positive Inflation News Narrative (1,000)
- Neutral Channel x Positive Inflation News Narrative (1,000)
- Neutral Channel x Negative Inflation News Narrative (1,000)
- Opposition Channel x Negative Inflation News Narrative (1,000)
- Positive Inflation News Narrative (1,000)
- Negative Inflation News Narrative (1,000)
- Placebo News Control (1,000)
- No News Control (1,000)

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

Supporting Documents and Materials

IRB