Human Life Histories as Dynamic Networks: Using Network Analysis to Conceptualize and Analyze Life History Data

Abstract

The examination of multiple life history indicators is essential to evolutionary sciences. However, the statistical analysis of life history parameters’ covariation is not apparently clear, due to the statistical limitations of “classic” procedures, like Factor Analysis, and conceptual problems in interpreting covariation between life history indicators as latent factors. Here, we propose that Network Analysis represents a promising framework for the exploration of life history parameters. First, we briefly describe the following basic metric of Network Analysis: nodes, edges, proximities, clustering, centrality indices, and small-world estimations. Next, we show the implementation of Network Analysis using the empirical set of life history variables as an example (N = 460). We showed that Network Analysis provides the following: (1) optimal level of information—higher than factor analysis and lower than correlation analysis; (2) findings that are in accordance with the existing life history data; (3) the estimation of age at first birth as the central node in the network; (4) dynamic view of life history events which can represent a solid basis for causal life history models. In sum, Network Analysis shows high potential both for conceptualizing life history pathways as dynamic networks and for statistical analysis of the covariation between the life history indicators.