NewsSlant

It is widely believed that political news is often slanted to suit news publishers' ideology. Politically slanted news can influence its readers to focus on different aspects of contentious social and political issues and hinder effective discussions.  To identify political slants in news and their influence on readers, we analyze Election-related news and user response to the news on Twitter. We collect Election-related news from six mainstream US news publishers covering the 2020 US Presidential Elections. We compute news publishers' political slants based on the favorability of news toward the two major parties' Presidential candidates. We use a Target-dependent Sentiment Classification (TSC) approach to determine the favorability of news toward the two candidates. We find that the Election-related news coverage shows signs of political slant in both, news headlines and on Twitter. The difference in news coverage of the two Presidential candidates between the left-leaning news publishers (left) and right-leaning news publishers (right) is statistically significant. The effect size is larger for the news on Twitter than for headlines. Further, news on Twitter is more sentimental than news headlines. We further identify moral foundations in user responses to news on Twitter based on Moral Foundation Theory (MFT). We fine-tune a transformer-based deep learning model to identify moral foundations in user responses. Moral foundations in user responses to left and right differ significantly. Further, the shift in moral foundation in user responses to left and right differs across social and political issues. User engagement is highest in user responses to the right, and lowest in user responses to balanced news.