Influence of Religion, Culture, and Education on Perception of Climate Change and its Implications: Applying Causal Inference in Statistics

抄録

This study examines the causal relationships between factors related to religion, culture, and education that influence perceptions about climate change in the Republic of the Marshall Islands (RMI) and Kiribati. Building on previous research that identified these three domains as important, this study utilizes Bayesian networks to uncover deeper connections between specific variables. Questionnaire data were collected from university students in both countries and analyzed using the R package “bnlearn” to construct Bayesian networks. Key variables from each domain were selected based on prior structural equation modeling studies. The resulting networks revealed complex interconnections between religion, culture, and education in shaping climate change perceptions. While some similarities exist between the two countries, notable differences emerged in the relationships between variables. In the RMI, certain climate perception variables appeared more isolated, while in Kiribati they showed broader connections to religious and cultural factors. These findings suggest that the underlying structure of influences on climate change perceptions may differ between the two island nations, despite surface similarities. The Bayesian network approach provides new insights into the causal pathways between domains that were not apparent in previous analyses. This deeper understanding of how religion, culture, and education interact to shape climate perspectives can inform more targeted and effective climate change communication and education efforts in the Pacific island communities.