Using Mixed-Mode Networks to Disentangle Multiple Sources of Social Influence

Abstract

Social network analysis has been widely used to model various forms of network influence on individual behavior. In network studies into peer influence on adolescent risk-taking health behavior, either friendship influence (using one-mode networks) or affiliation-based peer influence (using two-mode networks) have been employed to examine potential sources of network influence on adolescent substance use behavior. However, these sources of peer influence may be potentially confounded with each other. This paper introduces a new network method of incorporating one network’s influence into another network’s influence patterns, enabling us to decompose the different sources of network influences though the use of a combination of the networks. Specifically, the method incorporates affiliation-based peer influence (operationalized by a two-mode network) with either a second two-mode network, or the peer influence (operationalized by a one-mode network). Empirical examples are included to demonstrate the utility of this new approach.KeywordsSocial network analysissocial influenceone-mode networktwo-mode affiliation networkadolescent substance usehealth behavior