Modelling Societal Levels: An Integral Structural Framework of Nested Sets

Social theorists usually describe 'vertical' (e.g., classes, strata) and 'horizontal' (e.g., networks) dimensions of society as interplaying: Lower-level structuring results in emergent effects at higher levels, which enables change, while higher levels constrain lower levels which secures stability. Meanwhile, empirical research mostly focusses on either of the dimensions or, at best, unrealistically reduces either of these. To enable theoretically plausible modelling of social structure, we offer a cross-dimensional framework of Nested Set Structures and corresponding extensions of established statistical modelling approaches of Generalised Relational Hyper-Event Models (GRHEMs) and Nested Exponential Random Graph Models (NERGMs) that nest lower-level relations in higher level entities, thus combining both vertical-hierarchical and horizontal-network relations in a single framework. We illustrate the power of the approach by using three different datasets (Wikipedia, Conflicts, and Hospitals) to examine the interplay between societal levels while juxtaposing our effects of nested lower-level structures on higher levels and the inverse effects to those of mere affiliations of lower-level entities with higher level entities (as per the state-of-the-art methods we are challenging). The results propose an integral theoretically plausible account of interplay between micro, meso, and macro levels of society, bridge major streams in quantitative social science (network- and sample-based modelling), and shed light on emergence and constraint in society. Further expansion of the proposed perspective goes beyond social structure - incorporating symbolic/cultural and cognitive structures as well as combining human and material structures.