Network of Categories: A Method to Aggregate Egocentric Network Survey Data into a Whole Network Structure

Abstract

We propose a new method for analyzing and visualizing information on a large collection of personal networks to uncover the socio-centric structure of relationships among aggregated actors that we clustered into categories. The network of categories identifies the links between groups of sampled ego actors sharing a given attribute (e.g., “being immigrant organizations”) and abstract attribute-based groups or types of alters (e.g., “any immigrant organization”). This method takes advantage of both whole networks and ego-centered networks analytical approaches since it gains insights into the context’s whole network structure while generalizing the results to a wider population. The method can be applied to data obtained through modified versions of instruments such as position or resource generators, which are easy to collect within standard large- N representative surveys. We present a concrete application of the method to the analysis of immigrant organizations' position in local governance networks in 40 French and German cities.