Estimating peer political influence with large N observational data on ego-centered social networks

• Published in: Social networks 2022-07, Vol.70, p.198-207
• Main Authors: Mach, Bogdan W., Jackson, John E., Sadowski, Ireneusz
• Format: Article
• Language: English
• Subjects: Conversation; Data; Ego; Friendship; Influence; Network analysis; Observational data; Peer influences; Political networks; Political power; Social networks; Voting
• ISSN: 0378-8733; 1879-2111
• DOI: 10.1016/j.socnet.2022.01.002

The work in this paper analyzes a unique set of large-scale observational data that links data from a national sample of Polish citizens to data obtained from their regular conversation partners. The analysis finds and interprets substantial and significant peer to peer influences on public opinions and voting choices. The contribution of the paper to prior findings on peer to peer influences in networks consists in (a) basing our results on large-scale, nationally representative, observational data rather than on data from local samples, experiments or analyses of flows of information in constrained environments, (b) using a mix of common, but seldom used in this context, estimation methods, and (c) developing an idea of constructing simulated friendships pairs which are then analyzed to cast doubt on a common alternative interpretation of obtained results. •In this paper we analyze a unique set of large N observational data from a national sample of Polish citizens to data obtained from regular conversation partners.•The analysis finds and interprets substantial and significant peer to peer influences on public opinions and voting choices.•The analysis uses a mix of common Econometric and new estimation methods.