A paper titled "An Experimental Approach to Identify Perceived Opinion Formation Thresholds from Social Media," coauthored by Gyorgy Korniss & Boleslaw Szymanski was posted as ARL TR

A paper titled "An Experimental Approach to Identify Perceived Opinion Formation Thresholds from Social Media," by Derrik E. Asher, Justine Caylor, Alexis R. Neigel, Casey Doyle, Mark R Mittrick, John T Richardson, Gyorgy Korniss, and Boleslaw Szymanski was posted Defence Technical Information Center as ARL Technical Report ARL-TR-8763 on August 1, 2019. the paper focuses on passive social media consumption and measures when it can result in the formation of an individual’s opinion. If an individual trusts the veracity of the social media information, such newly formed opinion becomes a belief, and it has been argued that the purpose of belief is to guide action. Therefore, it is important to understand how social media information contributes toward the opinion formation of individuals because it may shape their future actions. The present study aims to identify thresholds estimated or perceived by an individual for the amount of social media data needed to form an individual’s opinion. The goals of the current study are to accomplish the following: 1) identify perceived opinion formation thresholds for three distinct social media data-types (i.e., Images, Videos, and Messages), 2) understand the influence of different contexts (i.e., Low controversy, Medium controversy, High controversy, and None) over opinion formation thresholds, and 3) determine how opinion formation thresholds change with the source of social media information (i.e., Unknown or unspecified, Like-minded or similar perspectives, and Different-minded or diverse perspectives). An experiment on Amazon Mechanical Turk (MTurk) was performed and the results of 945 participants were analyzed. The results yield several findings: 1) opinion formation thresholds represented as population averages are identified across the three distinct data-types, 2) context has marginal influence over opinion formation thresholds, and 3) influence from sources depends on the data-type. These results provide an empirically derived set of opinion formation thresholds that correspond to different dimensions of social media information. It is available at the above link