When AI Selects Political News: Algorithmic Content Curation and Party Plurality

The research project investigates how automated recommendation systems for political messages can influence voters' political decisions in election contexts.

 

News App Study

The news app study examines the news usage of app users in the political election campaign for the election of the 19th Lower Saxony state parliament on 9 October 2022.

• Project duration: 2022-2023
• Project management: Prof. Dr Juliane A. Lischka and Prof. Dr Katharina Kleinen-von Königslöw
• Research team: Researchers from Journalism and Communication Studies at Universität Hamburg (UHH), the Journalism Department at Johannes Gutenberg University Mainz (JGU) and the Dynamic and Distributed Information Systems Group (DDIS) at the University of Zurich (UZH), Switzerland.
• The study is funded by the Federal Ministry of Education and Research (BMBF) and the Free and Hanseatic City of Hamburg as part of the Excellence Strategy of the German federal and state governments.
• Contact: Interested parties can reach the research team at nachrichtenapp.wiso@uni-hamburg.de

 

Aims of the Study

The study investigates the role that the visibility of political parties in news in a smartphone app has for the voting intentions of users of the app. The aim of the study is to find criteria that characterise voting intentions and to investigate the role of automated news recommendation systems.

• Does the visibility of parties in the news change the voting preferences of app users?
• Which political news environments are associated with which political attitudes?
• What implications can be derived for automated news recommendation systems?

Scientific Background

Party visibility refers to the thematisation of parties and their candidates in media coverage (Eberl, Boomgaarden & Wagner, 2017). Visibility is an indicator of a party's political relevance and power (Miller & Krosnick, 2000). Typically, governing parties have a visibility bonus in reporting (Castro, 2021; Eberl et al., 2017; Hopmann, de Vreese & Albaek, 2011; Jandura, Udris & Eisenegger, 2019). Visibility can influence voting intentions and political attitudes (Broockman & Kalla, 2022; Hopmann, Vliegenthart, de Vreese & Albæk, 2010), especially those of undecided voters (Geers & Bos, 2017).

Automated recommendation systems for messages, for example in apps, can influence party visibility depending on design principles. This project is dedicated to the question of what influence party visibility environments have on political attitudes. In doing so, the project aims to contribute to determining the disruptive or equalising potential news recommenders can have for digital public spheres.

Study Procedure

Participants will be asked to use a news app for four weeks and answer a questionnaire at the end of the four weeks. The surveys will focus on political attitudes, party preferences, political knowledge, news preferences, media usage behaviour and assessment of the news app.

1. We recruit participants in an online household panel (ISO 20252 certified) through the market research institute Innofact.
2. If they are willing to participate, participants download the news app from the app store and are randomly assigned to one of three news environments.
3. Participants use the app for four weeks from 11 Sept to 9 Oct (election Sunday).
4. After election Sunday, we will invite all active app users to take part in a follow-up survey.
5. Afterwards, participants will receive a special payment from their panel provider.

Participation is voluntary and can be revoked at any time until the end of the survey without explanation.

 

News App

The news app and its news recommendation system were developed at the Department of Informatics at UZH and customised for the current study by the UHH research team. With the app, participants have access to daily news from six news media: four regional newspapers in Lower Saxony and two national daily newspapers. The app shows a news mix of politics, business, miscellaneous and the like. The articles are displayed one below the other in a news feed.

 

News Environments

Participants in the study are randomly assigned to one of three news environments in the app. These news environments are based on typologies of public spheres (Ferree, Gamson, Gerhards & Rucht, 2002; Helberger, 2019) and differ in terms of the visibility of the respective political parties.

• Group A sees articles on parties in proportion to party preferences in Lower Saxony according to previous election polls (‘Sunday Question’).
• Group B sees each party equally often in the coverage.
• Group C sees the three major parties disproportionately often in the articles.

Risks for Participants

Previous research shows that a higher visibility of political parties in news can lead to a higher preference, especially among undecided voters. There are risks for participants in that their party preferences might be altered according to party visibility in the coverage of their news environment.

In group B, for example, smaller parties are given greater visibility than corresponds to voting preferences (‘Sunday question’). In this group, party preferences could turn out to be more in favour of the smaller parties after participation. This is contrasted with the lower visibility of smaller parties in group C and compensated for. For the compensation, group C is larger than groups A and B by a factor of 1.4, so that no party is put at a possible advantage or disadvantage.

Data Protection

The data protection declarations for the use of the app and the study can be viewed here.

 

References

Broockman, D. & Kalla, J. (2022). The manifold effects of partisan media on viewers’ beliefs and attitudes: A field experiment with Fox News viewers. https://doi.org/10.31219/osf.io/jrw26

Castro, L. (2021). Measuring Partisan Media Bias Cross‐Nationally. Swiss Political Science Review, 27(2), 412–433. https://doi.org/10.1111/spsr.12459

Eberl, J.‑M., Boomgaarden, H. G. & Wagner, M. (2017). One Bias Fits All? Three Types of Media Bias and Their Effects on Party Preferences. Communication Research, 44(8), 1125–1148. https://doi.org/10.1177/0093650215614364

Ferree, M. M., Gamson, W. A., Gerhards, J. & Rucht, D. (2002). Four Models of the Public Sphere in Modern Democracies. Theory and Society, 31(3), 289–324.

Geers, S. & Bos, L. (2017). Priming Issues, Party Visibility, and Party Evaluations: The Impact on Vote Switching. Political Communication, 34(3), 344–366. https://doi.org/10.1080/10584609.2016.1201179

Helberger, N. (2019). On the Democratic Role of News Recommenders. Digital Journalism, 7(8), 993-1012. https://doi.org/10.1080/21670811.2019.1623700

Hopmann, D. N., de Vreese, C. H. & Albaek, E. (2011). Incumbency Bonus in Election News Coverage Explained: The Logics of Political Power and the Media Market. Journal of Communication, 61(2), 264–282. https://doi.org/10.1111/j.1460-2466.2011.01540.x

Hopmann, D. N., Vliegenthart, R., de Vreese, C. H. & Albæk, E. (2010). Effects of Election News Coverage: How Visibility and Tone Influence Party Choice. Political Communication, 27(4), 389–405. https://doi.org/10.1080/10584609.2010.516798

Jandura, O., Udris, L. & Eisenegger, M. (2019). Die Medienpräsenz politischer Akteure in Deutschland und der Schweiz. UFITA, 83(1), 170–195. https://doi.org/10.5771/2568-9185-2019-1-170

Miller, J. M. & Krosnick, J. A. (2000). News Media Impact on the Ingredients of Presidential Evaluations: Politically Knowledgeable Citizens Are Guided by a Trusted Source. American Journal of Political Science, 44(2), 301. https://doi.org/10.2307/2669312