Network Analysis of Social Representations for Community Detection

Pascal Moliner

Pascal Moliner Laboratoire EPSYLON, Université de Montpellier III, Montpellier, France

Keywords: social representation, group, network analysis, community, betweeness centrality

Abstract

This article proposes a new approach to the study of social representations which aims at studying them from the perspective of network analysis and graph theory. To do so, we consider networks integrating both the constituent elements of the representations (opinions, beliefs) and the individuals who adhere to these elements or are bearers of them. In this perspective, a representation network is presented as a set of nodes (opinions, beliefs, individuals) and links (the adherence of an individual to an opinion is considered as a link between this individual and the opinion to which he adheres). This procedure allows us to apply the algorithms developed in the field of network analysis for the detection of communities to the studied representation network. Three studies illustrate the proposed approach by showing that it makes it possible to identify the heterogeneity or homogeneity of a group interviewed during a social representation study.

Author Biography

Pascal Moliner, Laboratoire EPSYLON, Université de Montpellier III, Montpellier, France

Pascal MOLINER is a professor of social psychology at the University Paul Valéry-Montpellier III. His work concerns social representations, social cognition, and identity.