Community Disaster Resilience through Nutritional Status and Health Interventions in Indonesia: PLS-SEM Wild Bootstrap Approach

This study applies the Partial Least Squares Structural Equation Modeling (PLS-SEM) with wild bootstrap inference to analyze the effects of Specific and Sensitive interventions on the nutritional status of Indonesian communities. The wild bootstrap procedure was implemented directly on the factor scores obtained from the PLS model to enhance the robustness of statistical inference under heteroskedasticity and non-normal data conditions. Using 2,000 bootstrap replications, the empirical distributions of the path coefficients revealed that both Specific and Sensitive interventions have consistently negative and statistically significant effects on nutritional status, whereas the intercept was not significant. These findings indicate that increased intensity of health and nutrition programs leads to improved nutritional outcomes, thereby strengthening community resilience to disasters. Methodologically, integrating wild bootstrap into PLS-SEM provides more reliable confidence intervals and p-values compared to conventional bootstrapping, ensuring more accurate inference in heterogeneous and complex datasets such as the 2024 Indonesian Nutrition Status Survey (SSGI). This approach highlights the importance of robust statistical methods in supporting evidence-based policy for disaster resilience and public health development.