It’s about time: revisiting reciprocity and triadicity in relational event analysis

Abstract Societies are intricate systems comprising interdependent social actors interconnected through diverse relationships. It has long been recognized that reciprocity and triadic closure are two fundamental components of this interdependence, and have as such been included in social network models. However, computational limitations and modelling complexity have meant that reciprodicity and triadic closure statistics included in such models have necessarily been simplified. Using novel computational and modelling approaches in relational event analysis, the aim of this paper is to explore a spectrum of endogenous network effect definitions, ranging from straightforward binary variables to complex, temporal functions accounting for the diminishing relevance of past events. Through simulation studies and real-world dataset analyses, it highlights the importance of comprehensively considering temporal dynamics and subtle assumptions in defining network effects. Neglecting these aspects can lead to significant pitfalls in the analysis. Fundamentally, the paper highlights the time-varying nature of reciprocity and triadic closure effects evident in empirical datasets. While exponential decay functions sometimes capture their temporal structure, more complex continuous functions of time often describe the intricate structure of both effects more precisely.