Reducing Uncertainty Caused by Social Noise on Social Media Platform

Abstract

Social noise and social entropy are two concepts that have been the subject of studies in the last two decades (Bruno, 2010; Matei et al.,2010; Alsaid et al., 2020; Alsaid & Pampapura, 2022; Pampapura et al., 2022). Studies that discussed misinformation, fake news, conspiracy theories, controversies, and disinformation also focused on certain aspects of social noise (Romer & Jamieson, 2020; Van Prooijen & Douglas, 2018). In this study, social noise refers to users intentionally or unintentionally participating in creating, disseminating, or spreading misinformation on social media. Recent studies have identified six characteristics or constructs of social noise. The six constructs identified by the recent studies include image curation, relationship management, cultural agency, conflict engagement, affiliation & politics, and norms & beliefs. In this study, we use the constructs to quantify and measure social noise using entropy and uncertainty reduction theory. The study focuses on two questions. How do we measure entropy as a degree of uncertainty caused by social noise? How does reducing uncertainty and social noise contribute to reducing the spread of misinformation? The primary analysis indicated that entropy as a measure of disorders of a tweet 1.5. This result is in line with the results produced by Son et al. (2019). Sentiment analysis results show that 66.6% of the tweets are subjective, while 33.4 % of the tweets are objective (facts). Topic modeling results also indicated the presence of keywords that is characteristics of social noise constructs in the dataset.