DNA.FRAG is a collaborative project iof the University of Hamburg and GESIS – Leibniz-Institute for Social Sciences funded by the Geman Research Foundation DFG.

What is it about?

Digitalization has brought the fragmentation of news audiences to the forefront of public and scholarly debates. With the increasing diversity of digital media offerings, preference-driven news selection becomes more prevalent. This raises concerns that the integrative function of the public sphere is eroding: if shared media experiences decline, the foundation for social cohesion and democratic deliberation could be weakened.

However, whether and to what extent such fragmentation tendencies actually exist remains scientifically contested—not least because previous research has focused primarily on self-reports and used news sources (not content).

What are the objectives?

Through methodological innovations, DNA.FRAG enables substantive scientific advances to capture patterns, prevalence, as well as individual and societal drivers and effects of audience fragmentation more comprehensively and in greater detail. Specifically, the following previously unclear key questions are addressed:

Content-Level Fragmentation

To what extent are audiences fragmented not only in terms of the news sources they use, but also in terms of the (political) news content they actually consume?

Role of Digital Intermediaries

How do online intermediaries (such as social media, search engines, and news portals) shape news audience fragmentation? Do they contribute to convergence or further differentiation?

Dynamics of Fragmentation

How does audience fragmentation unfold as a dynamic process? To what extent does this process reinforce and manifest individualization, segmentation, and polarization tendencies in modern democratic societies?

What is the approach?

To answer these questions, DNA.FRAG combines secondary analyses of two unique datasets with an innovative primary data collection via GESIS Panel.dbd—a new research infrastructure that links survey data with digital behavioral data.

Methodologically, the project combines web tracking data (on news exposure), panel surveys (on interests and attitudes), and automated content analysis with innovations in original network science (e.g., two-mode networks) and computer science (e.g., zero- and few-shot learning with large language models).

Who is conducting the research?

DNA.FRAG is a collaborative project of the University of Hamburg and GESIS – Leibniz-Institute für Social Sciences. To capture audience fragmentation in its complexity, the project brings together expertise from media and communication studies, network science, and computer science.

The project is funded by the German Research Foundation (DFG), the duration is three years (2025-28).

The Principle Inverstigators are: Prof. Dr. Katharina Kleinen-von Königslöw (University of Hamburg)  and Dr. Frank Mangold (GESIS – Leibniz-Institute for Social Sciences)

The Reseach Associates are: Judith Gilsbach (GESIS – Leibniz-Institute für Social Sciences), Lars Reinelt (University of Hamburg)