Since the widespread uptake of social media in the 2010s, research on emotion and sentiment in online sharing has grown rapidly across various fields. Yet, this expansion lacks standardized methods and shared concepts, resulting in fragmented evidence and even contradictory findings on the direction of these effects. No systematic synthesis across fields currently exists. This review fills that gap by offering the first comprehensive overview of findings, scope, and methodologies at a pivotal moment before large language models reshape this research landscape.
The systematic review identified 136 studies from 1,181 search results retrieved via Web of Science and Scopus in accordance with the PRISMA guidelines. Given the heterogeneity of the studies, the synthesis is presented in narrative form.
Emotions and sentiment, in general, increase the popularity of information across platforms. The effect is found most consistently for negative discrete emotions, such as anger, anxiety, sadness and fear. Positivity bias is often identified in health communication, while negativity bias is prevalent in political communication. Going forward, the field should focus on sampling across platforms, greater linguistic diversity, and analysis of visual material.
The systematic review provides the first synthesis of a rapidly expanding field relevant to scholars across disciplines. It identifies potential sources of variance in the results and highlights methodological heterogeneity and identifies paths for future research.
