Using Bayesian network analysis in social sciences: A case study of domestic water and energy use

La Matta Romero F.; Lewis T.R.; Staddon C. · Journal of the Bulgarian Geographical Society (2025) · DOI: 10.3897/jbgs.e168308

Resumen

Understanding the factors that shape household water and energy use is essential for designing targeted conservation interventions that promote both sustainability and well-being. While studies in this area often rely on traditional “frequentist” statistical methods, which can struggle to capture the complex interdependencies among de-mographic, behavioural, psychological, and material influences. This paper introduces Bayesian network (BN) analysis as a novel and adaptable method with useful applications in water and energy studies and a wide variety of other social sciences. The paper offers a primer on how to conduct BN analysis, including underlying logic and range of choice of software platforms, before presenting a brief worked example based on the authors’ current research into household water and energy consumption in a UK city. The paper shows how Bayesian networks can generate valuable insights from relatively small and complex datasets, capture non-linear relationships, and support scenario-based reasoning, making them well-suited for exploratory studies, “what if?” scenario-testing and policy effectiveness review. The findings contribute to a more nu-anced understanding of domestic water and energy consumption and offer a practi-cal framework that can inform the design of targeted, evidence-based interventions to encourage sustainable water and energy use in households. We argue that there is much to be gained by proliferation of this analytical approach throughout the social sciences. © Fiorella La Matta Romero et al.

Instituciones: Universidad Peruana Cayetano Heredia