Impact of emotions in social media content diffusion

Shivangi Chawla,Monica Mehrotra

DOI: https://doi.org/10.31449/inf.v45i6.3575

2021-08-30

Informatica

Abstract:Emotions present in social media content shape its diffusion. This study seeks to comprehensively examine the impact created by emotions of a social media message on its diffusion. Centered on a non-domain specific Twitter dataset, the authors define several measurement constructs to quantify the tweet diffusion process namely, speed, size, half-life, diffusion potential, and engagement.  Since a message may express a single dominant emotion or multiple categories of emotions, the current study focuses to investigate the influence of emotions in the single label as well as multi-label setting. Through extensive statistical analyses (Multivariate Analysis of Variance and Regression), we find that the impact of emotions on diffusion constructs was statistically significant. The findings shed light on how emotions aid or hinder the spread of information through social media. Specifically, the tweets containing joy or contempt as primary emotion attained faster and stronger diffusion. In contrast, anger or fear as primary emotion in tweets contributed to slower and weaker diffusion. Also, the combination of one or more positive and negative emotions increased the diffusion outcome.

computer science, information systems,mathematics, applied