Who is Most Influential in the Core Discussion Network? Core Discussion Network Evolution and Political Attitude Change in the Netherlands
Thijmen Jeroense, Niels Spierings, Jochem TolsmaAn often-cited cause of political polarization is network segregation: as people are less likely to meet and mingle with different people they are not confronted with different views and as such are likely to have entrenched views on political issues. While intuitively plausible, it has rarely been tested whether this argument holds or theorized under what conditions it does or does not. In this study, we will analyze a core of this argument: the network - attitude linkage.
More concretely, we will focus on the influence of core discussion networks (CDN) on political attitude formation. Previous research indicates that the CDN is far from stable, and fluctuations in composition and size are frequent. However, whether these changes have an effect on the attitudes and behavior of individuals is understudied. We contribute to this gap by assessing whether CDN evolution affects political attitude change of individuals. Particularly, we will zoom in on the socio-economic composition and size of the network and on changes in a set of political attitudes, including the liberal-conservative axis and the economic axis.
Our main contribution is twofold. In the first part of this study, we will assess whether ego’s CDN composition and stability affect political attitudes of ego. In this part, we take a network-level perspective and we theorize and test how changes in the overall composition and size of the core discussion network influence political attitudes of individuals. As such the first research question is: How are changes in the CDN related to changes in political attitudes? In the second part of the study, we will delve deeper and theorize and test the influence of confidants in one’s CDN. We want to know when and why specific confidants are influential. In this part, we consider the role of characteristics of the dyad (political discussion frequency and similarity with ego) as well as the specific network context (size and composition). Our second research question is: When do confidants of one’s CDN have the most influence on one’s political attitudes? To answer these questions we use unique longitudinal egocentric network data from the LISS panel. This data encompasses 11 waves and covers the period between 2006 and 2019, which provides a unique opportunity to study the influence of confidants on political attitudes of ego. The political attitudes include emancipation, EU-integration, immigration, cultural exclusion, and income differences. Furthermore, the socio-economic composition, size, and stability of the network is known. Dynamic panel models are used to estimate the influence effects. By estimating it dynamically, we are able to rigorously control for reversed causality (i.e. selection effects): that attitude change leads to a composition change in the CDN.