No Filter Bubbles? Evidence from an Online Experiment on the News Diversity of Personalizing News Aggregators

Joschka Andreas Hüllmann*, Leonard Wilhelm Sensmeier

*Corresponding author for this work

Research output: Contribution to conferencePaperpeer-review


People increasingly use personalizing news aggregators for acquiring news online. Despite benefits such as more relevance, filter bubbles, among other risks, have been elicited. So far, there is little empirical evidence on how personalization affects the news diversity of news aggregators. Furthermore, it remains unclear how the news diversity of news aggregators compares to edited newspaper websites. This study investigates the effect of personalization on news diversity and reports the results of an online experiment. Based on browser instrumentation, news articles were queried using personalized and non-personalized profiles. Using a fixed effects model, the news diversity’s change due to personalization was estimated at 8-12%. However, the absolute news diversity of the personalizing news aggregators is comparable to edited newspaper websites with a difference of <5%. Our results contribute empirical insights to the debate on news personalization, finding that filter bubbles stemming from low news diversity are unlikely.
Original languageEnglish
Number of pages17
Publication statusPublished - 2022
EventAnnual Pacific Asia Conference on Information Systems, PACIS 2022: PACIS 2022 - Virtual Conference, Taipei/Sydney, Taiwan
Duration: 5 Jul 20229 Jul 2022


ConferenceAnnual Pacific Asia Conference on Information Systems, PACIS 2022

Cite this