2017/03/10 by Pedro Calais Guerra, Roberto C. S. N. P. Souza, Guerra, Pedro Calais +5 · 1 citation
Physics and Astronomy · Social Sciences · #Complex Network Analysis Techniques #FOS: Computer and information sciences #Opinion Dynamics and Social Influence #Social Media and Politics #Social and Information Networks (cs.SI)
paper · pdf · doi:10.48550/arxiv.1703.03895
openalex publication_date 2017/03/10 · openalex created_date 2023/07/15 · openalex updated_date 2026/07/28
In this paper, we study the implications of the commonplace assumption that\nmost social media studies make with respect to the nature of message shares\n(such as retweets) as a predominantly positive interaction. By analyzing two\nlarge longitudinal Brazilian Twitter datasets containing 5 years of\nconversations on two polarizing topics - Politics and Sports - we empirically\ndemonstrate that groups holding antagonistic views can actually retweet each\nother more often than they retweet other groups. We show that assuming retweets\nas endorsement interactions can lead to misleading conclusions with respect to\nthe level of antagonism among social communities, and that this apparent\nparadox is explained in part by the use of retweets to quote the original\ncontent creator out of the message's original temporal context, for humor and\ncriticism purposes. As a consequence, messages diffused on online media can\nhave their polarity reversed over time, what poses challenges for social and\ncomputer scientists aiming to classify and track opinion groups on online\nmedia. On the other hand, we found that the time users take to retweet a\nmessage after it has been originally posted can be a useful signal to infer\nantagonism in social platforms, and that surges of out-of-context retweets\ncorrelate with sentiment drifts triggered by real-world events. We also discuss\nhow such evidences can be embedded in sentiment analysis models.\n