Effects of narratives and information valence in digital headlines on user responses

The way headlines are framed and presented can influence how audiences respond, even when the substance of the headline remains the same. This study scraped big data of digital headlines and associated user response metrics from three leading video-and-news-sharing platforms in China, and then conducted a comprehensive quantitative content analysis. Through path analysis, significant main and interaction effects were discovered, highlighting the role of narrative (as opposed to numerical) evidence and the negative affect inherent in headlines on viewing. Furthermore, viewing mediated the effects of narratives and information valence on deep responses, such as liking, sharing, and commenting. The study also revealed that information valence moderated the mediating effects of viewing on the relationships between narrative and deep responses. These effects remained significant even after controlling for headline word length, source influence, and discrepancies between publication and scraping dates. Based on these findings, a new concept, the 'Negative Narrative Effect' (NNE), is introduced, and the implications for data journalism are discussed.