Group Polarization in Social Media: An Effective Network Approach to Communicative Reach and Disinformation Strategies.
Abstract
The objective of this project is to develop an improved understanding and ability to measure and predict the formation of beliefs, a"nd in particular polarized beliefs in social media. The goal is both to develop new theory and to develop the associated metrics and methods for assessing group beliefs in social media. We do this in the context of counter-messaging and disinformationcampaigns. We address: What makes groups vulnerable? How can polarization be exacerbated? What leads to individuals not moving toward group~s po"lar extreme? The approach we propose is grounded in affect control theory, social influence, network theory and the rhetorical the""ory of communicative reach. Affect control theory posits that to maintain affective cognitive balance individuals, when confronted w""ith conflicting evidence, will move to either alter their affective state (which could make them more extreme) or will distance them"selvesfrom the group. Social influence theory posits that what individuals believe is a function of the beliefs of those with whom they interact. Network theory posits that the structure of the group and the individual~s position in it determines who has the power to spread more information and beliefs. Communicative reach posits that individuals who communicate using messages that contain more high valued rhetorical concepts will have greater reach ~ and their message will be received by and influence more individuals. We combine these theoretical propositions into a unified model of communicative reach and group polarization. Each theory will be r"ecast in terms of high dimensional networks composed of relations between actors, groups, messages, concepts,and beliefs, and the p"rocesses recast as springs and pipes whereby one dimension influences another. Social media data from various topic groups will be coded into these same dimensions. Data coding and the unified model are operationalized and made reusable using a combination of lang"uage technology, network analytics, and machine learning optimized for social media and the linkages between social media and other"" media. The anticipated result is an approach that enables empirical assessment of group attitudes and their polarization, forecasti""ng the impact of various courses of action to affect that polarization process, and an empirically instantiable unified theory of be""lief formation, communication reach and the role of counter-messaging and disinformation.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 23, 2018
- Source ID
- N000141812106
Entities
People
- Kathleen Carley
Organizations
- Massachusetts Institute of Technology
- Office of Naval Research
- United States Navy