Expansion and Exploration of the Superdiffuser Model With Agent-Based Modeling

Abstract

The superdiffuser model predicts that the diffusion of a new behavior can be accelerated if superdiffusers (people who are connectors, persuaders, and mavens) are recruited to promote the behavior. We propose an expanded model where the importance of these traits varies by network structure and other network member characteristics. We assessed the plausibility of these proposed moderators using a simulation of an agent-based model in which each moderator was varied along with the superdiffuser traits of those initially seeded with the behavior. Analysis of the results of those simulations showed support for many, but not all, of the predictions as well as several surprising results that are suggestive for future empirical research.