Exploring the interplay of individual traits and interaction dynamics in preschool social networks

TL;DR

The study investigates how individual traits influence preschool social interactions using RFID proximity data, parent surveys, and language assessments collected over 10 months. It combines dyadic and higher-order network analyses with machine learning to link traits such as age, sex, siblings, and language skills to interaction patterns, both inside and outside classrooms. Key findings show that these traits predict extremes in social activity, that age differences modulate contact duration, and that group embedding (hypercoreness) relates to siblings, with out-of-class interactions revealing stronger higher-order effects. The work highlights the value of integrating heterogeneous traits into mechanistic models of early social development and provides a foundation for more context-aware simulations of child social dynamics.

Abstract

Several studies have investigated human interaction using modern tracking techniques for face-to-face encounters across various settings and age groups. However, little attention has been given to understanding how individual characteristics relate to social behavior. This is particularly important in younger age groups due to its potential effects on early childhood development. In this study, conducted during the Complexity 72h Workshop, we analyze human social interactions in a French preschool, where children's face-to-face interactions were monitored using proximity sensors over an academic year. We use metadata from parent surveys and preschool linguistic tests, covering demographic information and home habits, to examine the interplay between individual characteristics and contact patterns. Using a mixture of approaches, from random forest classifiers to network-based metrics at both dyadic and higher-order (group) levels, we identify sex, age, language scores, and number of siblings as the variables displaying the most significant associations with interaction patterns. We explore these variables' relationships to interactions within and outside classrooms and across mixed and single-grade classes. At the group level, we investigate how group affinity affects group persistence. We also find that higher-order network centrality (hypercoreness) is higher among children with siblings, indicating different group embedding despite similar total contact duration. This study aligns with existing literature on early social development and highlights the importance of integrating individual traits into the study of human interactions. Focusing on 2-5-year-olds offers insights into emerging social preferences during critical phases of cognitive development. Future research could use these findings to enhance mechanistic models of complex social systems by incorporating individual traits.

Figures (7)

• Figure 1: Descriptive statistics of individual socio-demographic and linguistic features. (a) Distributions of childrens' age by their grade and class they were assigned to. Notice the presence of classes with mixed grades. (b) Percentage of answers given by the parents when asked is their child rather sociable or shy, disaggregated by sex. (c) Distribution of number of siblings that the children have. (d) Distributions of aggregated scores in development skill tests disaggregated by sex.
• Figure 2: Descriptive analysis of preferred home activities among individuals. (a) Number of positive and negative responses associated to the different out-of-school activities during day (leftmost bars) and before nightime (rightmost bars). (b) Correlation matrix of favorite activities.
• Figure 3: Visualizations of the temporal aggregated network during the out-of-class context. The size of the nodes corresponds to their degree. The layout is induced by the bipartite representation of interactions, where coloured nodes representing pupils are connected through "phantom nodes" representing group interactions (hyperedges). (a) Colors indicate the sex of pupils: blue for male and red for female. (b) Colors indicate class affiliation. The location of nodes remains the same between the two networks. The visualization has been produced using Gephi bastian2009gephi.
• Figure 4: Classifying pupils into classes of social activity based on individual metadata during in-class (a-c) and out-of-class (d-f) time. (a,d) Pupils are divided into four classes of social activity based on individual time spent in social interactions. The 4 quantiles of the distributions are marked by dashed lines. (b, e) Confusion matrices associated to the classification task performed by a random forest model, that is able to distinguish the most and least socially active pupils. The bar plots (c,f) report the relevance the model assigns to each individual feature; the top 5 features are consistent across the two contexts.
• Figure 5: From the weighted interaction network of children aggregated across weeks, the duration of interactions is plotted against age difference of children interacting (a) during class time (b) during free time. The data is compared with 50 realizations each of hard and soft reshuffling models. The hard reshuffling does not preserve degree distribution nor the weight distribution while the soft reshuffling preserves degree distribution while shuffling the weights (duration of interaction) among edges. (c) Duration vs age difference for out-of-class interactions for pairs of nodes where both have siblings, both have no siblings, only one has siblings