With Their Own Voice: Inducing News Media Trust with a Structural Topic Model for Open-ended Survey Responses

• In Event: ICA Interactive Paper/Poster Session III
  In Interactive Poster Session: Journalism Studies Interactive Poster Session

Abstract

This paper attempts to clarify what a core concept in the field of journalism studies, news media trust, means for the audience. We asked an open-ended question about what comes to mind when people think about "trust in news media", fielded in a probability based Norwegian online panel (N=1898). Using a semi-automated quantitative text analysis technique called Structural Topic Modeling (STM), we induce that Norwegian citizens' write about five different topics: objective reporting, ideology, correct reporting, critical reporting, and balanced reporting. We identify a new dimension of trust, ideology, that has yet to be taken into consideration in the literature. The topics vary with Norwegian's age and education, but vary little with the respondents' political preferences and news habits. These results give further depth to our understanding of trust in news media, and can inform scholarly debates on how news media trust can be conceptualized and measured.

Authors

• Erik Knudsen, University of Bergen

• Magnus Hoem Iversen

• Silje Nygård, University of Bergen

• Stefan Dahlberg, U of Bergen

• Mikael Poul Johannesson, University of Bergen