Going negative across platforms and time: Assessing the supply and demand of negative campaigning

Abstract

This study investigates dynamics of negative campaigning by Norwegian political parties on Facebook, Instagram and Twitter from 2013 to 2024. Analysing 89,791 posts from nine major parties, it uses a novel content analysis approach combining Large Language Models (LLMs) and human annotation to categorize posts as negative, positive or neutral. The findings show an overall rise in negative campaigning across platforms, with significant changes following leadership shifts in the right-wing populist Progress Party (PP). The study examines engagement levels, measured by likes, for different content types. Initially, negative content attracted more engagement on Facebook. This trend, however, decreased over time, while Instagram saw increased engagement with negative content during later years. Twitter users appeared to have rather consistently favoured negative content, reflecting its reputation for incivility. The article enhances understanding of political communication strategies in a high-choice media environment, emphasizing the need to tailor campaign strategies to platform-specific user dynamics.