A Data-Driven Framework for Coding the Intent and Extent of Political Tweeting, Disinformation, and Extremism

Disinformation campaigns on online social networks (OSNs) in recent years have underscored democracy’s vulnerability to such operations and the importance of identifying dissecting their methods, intents, source. This paper is another milestone a line research political disinformation, propaganda, extremism OSNs. A total 40,000 original Tweets (not re-Tweets or Replies) related U.S. 2020 presidential election are collected. The intent, focus, affiliation these determined through multiple discussions revisions. There three affiliations: rightist, leftist, neutral. 171 different classes intent focus defined for Tweets. 25% were left out while defining intent. purpose assure that would be able cover unseen (Tweets not used determine define classes) no new required. provides classes, definition size, example from them. If any information included Tweet, its factuality verified valid news sources articles. opinion it whether extreme, analytics with regard results show disinformation extreme opinions more common among rightists than leftist Additionally, Coronavirus pandemic topic almost half Tweets, where 25.43% express unhappiness how Republicans handled this pandemic.