"my stance decides my language": Modeling of Framing and Political Stance in News Media

Abstract

Framing is a political strategy in which journalists and politicians highlight certain aspects of an issue or a problem to influence public opinion. Frameworks for detecting framing in news articles or social media posts are necessary in order to understand the spread of biased information in our society. Prior research efforts have shown that their framework for framing detection works well by predicting political affiliation afterward. In this paper, rather than predicting stance after detecting frames, we incorporate stance prediction into a framing detection model to jointly capture framing languages better. We take advantage of political stance data, which are more readily available than framing data that require manual annotation of professionals, and propose automatic framing detection models, which can detect previously unseen framing phrases. We compare two different methods of incorporation and show that leveraging stance prediction improves the separation of liberal and conservative biased frame language.

Tasks

Reproductions