Micro and Macro Spatial Networks in Two Contexts: Collaboration and Interpersonal Risk Communication

Networks researchers have been learning the prominent role geographic/physical proximity plays in network dynamics. In this paper, we examine spatial distance in the context of two distinct social networks differing in size, geographic boundaries, and relation types. The first network comprises egonetworks of Dutch farmers and permeates across a rural region of roughly 1600 square kilometers in the southwest of the Netherlands. The interpersonal, information-gathering contact list of each respondent is right censored and is accompanied by several similarity relations as well as self-reported spatial distances between the respondent and their contacts. The second network is whole and contains collaboration activity among faculty in a single academic department whose offices are physically proximal within tens of meters (i.e., walking distance). While physical distance is generally negatively correlated to tie existence and their weights (i.e., co-authorship and interpersonal contact), its effect is differently mediated by homophilous relations depending on the context: 1) regional cultural heterogeneity affects the impact of distance in the network of farmers and 2) both direct similarity (in research agendas) and meta-similarity interact with distance in predicting co-authorships. We also discuss some complications in employing spatial distance in network analysis, e.g., the identification of a null-hypothesis for distances in drawing inferences from ego-networks.