A model for the dynamics of face-to-face interactions in social groups

Abstract

Face-to-face interactions in social groups are a central aspect of human social lives. Although the composition of such groups has received ample attention in various fields—e.g., sociology, social psychology, management, and educational science—their micro-level dynamics are rarely analyzed empirically. In this article, we present a new statistical network model (DyNAM-i) that can represent the dynamics of conversation groups and interpersonal interaction in different social contexts. Taking an actor-oriented perspective, this model can be applied to test how individuals’ interaction patterns differ and how they choose and change their interaction groups. It moves beyond dyadic interaction mechanisms and translates central social network mechanisms—such as homophily, transitivity, and popularity—to the context of interactions in group settings. The utility and practical applicability of the new model are illustrated in two social network studies that investigate face-to-face interactions in a small party and an office setting.

Loneliness and social interaction dynamics: An ambulatory assessment study in a student population

“…The present study is set out to investigate if loneliness predicts social interactions measured with automatically sensed social interaction data and if, in turn, these structural aspects of social interactions predict changes in loneliness. We differentiate between joining and leaving social interactions, as these are two different social processes (Hoffman et al, 2019). To date, these two aspects of the structure of social interactions have not been examined in relation to loneliness.…”

“…With these variables we can control for fluctuations in interaction patterns that be due to situational factors (e.g., conversation phases might be more frequently interrupted due to where people place their sensing smartphone, interactions happen more frequently during certain times of the day). Such time and window-control variables are commonly used when modelling interaction dynamics (Hoffman et al, 2019;Stadtfeld & Block, 2017). It is important to note that we did not compute the "time in state" for the overnight observation, hence these the time gaps between the last measure in the evening and the first measure in the morning are not part of the analyses.…”

Separable and semiparametric network-based counting processes applied to the international combat aircraft trades

“…In this article, we introduce a tie-oriented model for the analysis of network-based event data. Tie-oriented models assume a bilateral intensity governing the occurrence of events within a dyad, as opposed to actor-oriented models suggested by Stadtfeld (2012) and extended in Hoffman et al (2020), Stadtfeld et al (2017). This approach partitions the intensity into an egocentric senderspecific intensity and a probability selecting the receiver conditional on the sender along the lines of the discrete choice model of McFadden (1973).…”

Creation, evolution, and dissolution of social groups

“…Instead, their membership changes over time with individuals constantly deciding whether to stay in their current groups and perhaps adjust some opinions, or leave to join other groups 7 . To date, however, the dynamics of human group formation, evolution, and dissolution in real social groups remain poorly understood 8,9 . Among humans, groups represent a level of association mediating between the individual and whole societies, producing a variety of personal benefits, including companionship and support 1 , while also allowing people to come together to pursue instrumental goals or work together toward common purposes 10 .…”

“…Indeed, new technologically mediated interactions such as texting, online chatting, and social media complement rather than replace face-to-face interaction in small groups, and a great majority of online ties ultimately relies on (or emerges from) offline face-to-face interactions 11 . Extended to social connections within face-to-face groups, computational and empirical literature have found that these groups are continually evolving rather than being static 5,8,9,12,13 . Such evolution has been aided by the tendency of the contemporary societies to relax social and institutional restrictions on group joining and leaving behavior compared to the past or to the societies that are more traditional 10,14 .…”