Examining negativity and positivity bias in Pulitzer prize-winning news: a study across six subgenres

Abstract Linguistic positivity bias (LPB) has been widely recognized as a universal linguistic phenomenon in various domains. However, this contradicts the prevalence of negativity, often considered a fundamental aspect of news value, in news discourse (Liu and Zhu 2023). Furthermore, the variation of sentiment orientation across different news subgenres remains largely unexplored. To address this contradiction, we conducted sentiment analysis on a corpus consisting of over a thousand Pulitzer Prize-winning news articles spanning 26 years, covering six distinct news subgenres: Breaking, National, Local, Explanatory, Investigative, and International News. Our analysis revealed that LPB might not hold uniformly in news discourse, as negative sentiment prevails in our Pulitzer corpus. Additionally, we identified distinct patterns of sentiment across subgenres, also challenging the notion of LPB. A negativity bias is evident in several subgenres, while positivity is observed in others, implying nuanced differences of sentiment within news discourse. Furthermore, we found that although topic preference contributes to sentiment patterns, it is not the sole determinant.