Algorithms in Digital Media and Their Influence on Opinion Formation
Algorithms decide what you see when on the Internet and in which order by deciding on the priority and contents. Since it also affects opinion formation on socially relevant topics, a research team of the Hans-Bredow-Institut summarises the state of research in a report for the German Bundestag and makes further recommendations.
Algorithms have a significant influence on the selection, prioritisation and arrangement of media content online. You decide, for instance, what websites are in fact relevant for a search request and in what order the results should be shown. The user’s behaviour will be included in this process as well. Personalised algorithms decide on the selection of messages (and advertisement) that users will see. A direct influence on opinion formation cannot be deduced from it but there is a certain potential for influencing users.
The Office of Technology Assessment at the German Bundestag (TAB) commissioned the Hans-Bredow-Institut with a report on the topic of “Auswirkungen auf die individuelle Meinungsbildung bei Nutzenden [Effects on Individual Opinion Formation of Users]“ within the TA-project “Algorithmen in digitalen Medien und ihr Einfluss auf die Meinungsbildung [Algorithms in Digital Media and Their Influence on Opinion Formation]“. In May 2018, the research team will send their report to TAB who will summarise it with further expert opinions to a final report.
This report has three scientific objectives:
To systematise algorithmic effects on processes of information selection, to classify the context of other mechanisms of information behaviour and to determine the relevance for individual and public opinion formation based on the available state of research
To determine the meaning of algorithmic selection for information behaviour from different perspectives and with different methods based on own empirical studies. For this purpose, we will conduct secondary analyses of the Reuters Institute Digital News Survey along with qualitative studies on intermediaries and opinion formation. There will be also analyses based on the “Datenspende [Collection of Data]”-project (Algorithmwatch), which provides an insight into the algorithmic personalisation of search results on Google and news recommendations on Google News, and the project “Robin” (University of Amsterdam) in order to assess the personalisation of Facebook News Feeds.
To combine the findings from the systematic analysis of the state of research and our own empirical analyses and, on this basis, to identify regulatory problem areas and provide initial recommendations for action.
Research programme: RP1 - Transformation of Public Communication