Web Connections: Shared Reality and Extremism in Online Spaces

Canfer Akbulut, Dr. Maya Rossignac-Milon, Dr. Sandra Matz, Professor E. Tory Higgins, Department of Psychology, Columbia University, 1190 Amsterdam Ave, New York, NY 10027, United States

In light of mounting concern over ideologically-motivated violence worldwide, much scholarship has been devoted to the study of extremism in tight-knit online communities. Research on shared reality in both in-person and online settings has revealed that when people perceive that they share the same thoughts and feelings as others, they become more certain of their interpretations of the world around them. That is, when someone believes that others share their views, thoughts, and feelings, they are more likely to perceive their own world-view as accurate, objective, and truthful. In this study, we examined whether shared reality in online communities can give rise to the expression of more extreme beliefs by community members. In extracting naturalistic online dialogues on Reddit, a popular discussion website, we operationalized shared reality between commenters by computing their latent semantic similarity (LSS), a natural language processing (NLP) technique previously linked to shared reality creation. We operationalized extremism through the use of sentiment analysis, a well-validated NLP method of quantifying the direction and intensity of valence in written speech. In 166,289 posts from 23,360 users in Reddit communities dedicated to the general discussion of politics and world news, LSS significantly and positively predicted extremism. We replicated this finding in a sample of more ideologically narrow Reddit communities. Our findings indicate that greater shared reality predicts the expression of more extreme sentiments, suggesting that engaging with like-minded groups in an online space may promote the radicalization of attitudes.

Additional Abstract Information

Presenter: Canfer Akbulut

Institution: Columbia University

Type: Poster

Subject: Psychology

Status: Approved

Time and Location

Session: Poster 10
Date/Time: Wed 1:30pm-2:30pm
Session Number: 6580