Upholding a Good Image: Selective Country Comparisons in the Media

How to cite

Berger, Lara Marie; Doğan, Gönül; Strang, Louis. 2024. "Upholding a Good Image: Selective Country Comparisons in the Media." AEA RCT Registry. AEARCTR-0014795. https://doi.org/10.1257/rct.14795-1.0

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

Lara Marie Berger* , Gönül Doğan , Louis Strang

* Primary investigator

Registry ID
AEARCTR-0014795
DOI
https://doi.org/10.1257/rct.14795-1.0
First registered
Last updated
Status
In development
Trial dates
2024-11-18 – 2024-12-15
Countries
Private
JEL codes
D83, L82, C91
Registry record
http://www.socialscienceregistry.org/trials/14795

Abstract

We run an experiment to investigate whether German journalists have a home country bias when covering a global issue, namely, environmental performance. In the experiment, they generate headlines that make comparisons between either Germany and 10 other countries or France and 10 other countries. We additionally vary the audience of the headline, either German readers or French readers. Participants choose the comparison countries from a list containing a performance ranking in three different environmental metrics. We will compare the countries chosen across the treatments to document the extent of the bias as well as to test whether the bias is demand driven.

Intervention

We randomly assign participants to three groups in a between-subjects design. Between groups, we vary the fixed country (either Germany or France) and the targeted readers of the headlines (German or French). We randomly assign the following groups:

• Germany_German-Readers: Participants have to compare countries to Germany and write headlines for German readers.
• Germany_French-Readers: Participants have to compare countries to Germany and write headlines for French readers.
• France_German-Readers: Participants have to compare countries to France and write headlines for German readers.

Experimental design

Participants are recruited among German journalists who are either members of the professional association Journalistenzentrum Deutschland e.V., alumni of Cologne Journalism School or alumni of a journalism degree at University of Dortmund. They are approached via an email which is scheduled to be send out in November 2024.

In the main decision, participants have to compare a fixed country to another country, which they have to choose from a list. The list comprises ten countries, five that perform better than the fixed country and five that perform worse in the respective dimension. Subsequently, they are asked to provide a headline for a newspaper that compares the two countries, the fixed one and the selected comparison country. Participants face three decisions, where they have to select a different country in each decision. All three comparisons are in the realm of environmental protection. We use the Environmental Performance Index 2024 (Block et al. 2024) to construct the list of comparison countries. The selected dimensions are anthropogenic PM2.5 exposure, CO exposure, and projected emissions in 2050.

In a subsequent elicitation, the headlines are presented to another set of participants, the ‘readers’, who are recruited via Prolific. Their sole task is to rate these headlines on how likely it is that they would read the whole article. The participants of the main experiment who wrote the best-rated headlines then receive a bonus.

We hypothesize the following results:

H1: When Germany is compared to other countries, that is, in treatment groups 1 and 2, average performance of the comparison countries relative to Germany is negative.

H2: Countries that do better than Germany are more likely to be chosen for a French audience (treatment 2) compared to a German audience (treatment 1).

H3: When France is compared to other countries for a German audience (treatment 3), better performing countries are more likely to be chosen compared to when Germany is compared to other countries for a German audience (treatment 1).

See the Pre-Analysis plan for a description of our planned statistical analyses.

Design details

Randomization method
computer
Randomization unit
individual
Clusters
NA
Observations
240. We will end data collection either after 240 observations have been reached or the 15th of December 2024 has passed.
Arms
3 randomly assigned groups with 80 participants each
Primary outcomes
subjects’ choice of comparison countries
Secondary outcomes
• Demographics: age, gender, residence, education • Political preferences • National pride measure • Type of occupation • Beliefs about Germany’s performance during the COVID-19 pandemic in 2020, 2021 and 2022 regarding COVID-related deaths per capita

Cited by

No citation data available for this registration yet.

Keywords