Abstract
Stunting, characterized by chronic malnutrition due to undernutrition, is a significant public health problem in Indonesia, with a prevalence rate of 21.5% in 2023-a slight decrease from 21.6% the previous year. The large impact of stunting on child health has made it a priority in global and national agendas. This study which aims to identify the causes of stunting to inform targeted policy interventions. Given the multicollinearity among the influencing factors, traditional regression models often fail. Principal Component Regression (PCR) is used to overcome multicollinearity and improve the predictive power of the model. Using secondary data from the 2023 Indonesian Health Survey (IHS), this study identified two principal components from seven initial predictors assumed to influence stunting: Exclusive breastfeeding, low birth weight, undernourished maternal food intake, undernourished child food intake, complete basic immunization, proper sanitation, and hereditary diseases. Component 1 with an eigenvalue of 3.271, explained 53.1% of the variance, while Component 2 with an eigenvalue of 1.425, explained 73.5% of the variance. The PCR method effectively addresses multicollinearity, as evidenced by Variance Inflation Factor (VIF) < 5. This study highlights the importance of using advanced statistical methods for robust policy development in the fight against stunting.