Big data, content analyses, time, media, and communication

Aktivitet: Föredrag eller presentationInbjuden talare


Political polarization is defined as a process in which the political opinions of citizens and political elites become more extreme, and collective identities form in ways that promote exclusion and intolerance (Mason 2018). In the past decade, research on political polarization has exploded globally. Although findings yield differing conclusions, there is something akin to a consensus that the US political system has polarized in the last few decades. Outside of the US, empirical results have been scarce and mixed. Political polarization tends to be easier to measure in a two-party system, whereas multiparty systems might not be as conducive to clear cases of polarization.

Possible media effects connected to political polarization might not be limited to explanations revolving around selective exposure hypotheses. News are told as a narrative in order to make sense of “fragmentary observations” and present it in a coherent way, by drawing on the dominant myths of a society, through storytelling (Bird and Dardenne, 2009). If the media conveys an image of increased political polarization in society, there could potentially be effects on the perception of the levels of political polarization among media consumers and shape how political elites and voters act.

In this ongoing research project, we try to uncover the content and development in the way that the Swedish printed press uses the concept of political polarization over time. We do this through a combination of qualitative content analysis and natural language processing techniques. The paper you find enclosed presents the results of a pilot study of a smaller sample of texts.
Period2021 sep. 22
VidKarlstad University, Sverige

Fria nyckelord

  • sociala medier
  • big data
  • sentimentanalys