YouTube is one of the largest social media platforms, both in Denmark and globally, offering an infrastructure for various actors to create, distribute, view, and discuss video-based content. As such, YouTube holds a democratizing potential, as it can give a voice to ordinary people and perspectives that might otherwise go unheard.
However, the same opportunities can be exploited by different actors to spread misleading content and conspiracy theories or influence citizens' political opinions and behavior. Additionally, YouTube's opaque algorithmic content curation can foster political polarization, support addictive viewing behavior, and contribute to mental health issues.
Nevertheless, we know little about what content citizens watch on YouTube, how they engage with it, how algorithms shape viewing patterns, and how different viewing behaviors are related to factors such as conspiracy mentality, trust in societal institutions, and mental well-being.
Shedding light on these questions requires interdisciplinary research that combines approaches from STS studies, social psychology, and political science (including studies on institutional trust, political opinions, and political behavior). This is crucial for understanding who has the potential to influence whom and how on YouTube—and, consequently, what implications the increasing use of YouTube has for democracy.
To answer these questions, this project draws on computational and survey-based methods, utilizing a unique combination of large-scale donations data and survey responses from a sample of 1,000 Danes, representative of the adult YouTube user population in Denmark. With this dataset, we will be able to analyze Danish YouTube users’ viewing behavior with a high level of detail, down to the individual videos they have watched, and relate their viewing habits to various socio-demographic and attitudinal variables.
Methodologically, the project will contribute to the scientific field by developing methods for collecting and analyzing large-scale donations data. This includes the development of automated methods based on deep learning technology to analyze large volumes of video data (a current obstacle preventing insight into what content citizens consume on video-based social media) and to analyze video visitation trails, i.e., what leads users from one video to the next.
Empirically, the project will map and classify the various types of actors whose content Danish YouTube users engage with, identify different YouTube viewing patterns, and provide insights into what content Danes turn to during times of crisis, focusing on the Covid-19 lockdowns in 2020-2021, the Russian invasion of Ukraine, and the Israel-Palestine conflict, as well as during election campaigns, focusing on the 2022 Danish general election and the 2024 European Parliament election.
Furthermore, the project will empirically examine how the identified viewing patterns are linked to various sociodemographic characteristics, conspiracy mentality, institutional trust, and mental well-being. Finally, the project will use an experimental design to test how exposure to political content from influencers can affect political opinions and behavior.
The knowledge gained from this project will contribute to interdisciplinary fields through journal articles and conference presentations. By shedding light on the hidden dynamics of a central social media platform, the project will strengthen user and societal resilience, with a particular focus on current and future EU and national big tech oversight, policies, and regulatory practices.
Dissemination:
