The Impact of Intermediaries on Audience Fragmentation

• Applicant: Professorin Dr. Birgit Stark
• Subject Area: Communication Sciences
• Term: from 2016 to 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 323196109

For most of the time, audience fragmentation was a dystopian scenario that scholars saw as a potential outcome of an ever increasing supply of media outlets. The Internet - along with its explosion of content, new journalistic styles and technical processing of information - reinvigorated these fears. Websites allow for a drastically increased degree of selective and hence preference-driven news use. Within the normative frame of democratic theories, such individualized information behavior is predominantly seen as a negative trend: The resulting fragmented audience, thus the common assumption, being only a precursor to a disintegrated society. On the internet, users' selectivity is only one of the factors influencing fragmentation: Technical and social filters and recommendations play an ever-increasing, yet largely unknown role. Intermediaries such as search engines, news aggregators and social networking sites serve as brokers between content and users while subtly guiding selectivity. By collecting, structuring, weighting and aggregating information, they control the visibility and ultimately the findability of topics. Although users predominantly perceive intermediaries as helpful navigation tools, their filter and sorting logics may incur unintended side-effects on the structure of attention and thus on the structure of audiences. These society-level influences and especially their impact upon fragmentation have scarcely been subjected to empirical scrutiny. The proposed project directly contributes to the field's understanding of these issues. It measures the intermediaries' individual and combined influence on audience fragmentation and the resulting impact on attention to topics. The empirical design is supported by a novel network theoretic model of news use which exposes intermediary influence on multiple levels. At the project's core lies an innovative combination of methods: A content analysis identifies the topics contained in Germany's most important news sites while a representative panel tracks individual usage of their articles. In conjunction, these data reveal the precise topic exposure for each individual user, resulting in the first comprehensive insight into both the exposure diversity and fragmentation of news use. This allows for a long-desired empirical test of the frequently discussed filter bubble, along with a broad assessment of the impact of algorithmic and social information processing on society.

• DFG Programme: Research Grants