<![CDATA[Count data are commonly modeled using Poisson regression. However, this model has limitations when the equidispersion assumption is violated, particularly under conditions of overdispersion. This study aims to compare the performance of the COM-Poisson and Poisson–Tweedie regression models for handling overdispersion and to identify factors influencing the number of measles cases among children under five in Indonesia in 2024. The method used in this research is a quantitative approach using secondary data on laboratory-confirmed measles cases among children under five across 38 provinces in Indonesia. The data were analyzed using COM-Poisson and Poisson–Tweedie regression models. The explanatory variables include measles immunization coverage, the number of people living in poverty, the percentage of vitamin A supplementation, the percentage of exclusive breastfeeding, and the percentage of undernourished children under five. The results of this research indicate that the Poisson regression model exhibits overdispersion and is therefore not suitable for the data. Both the COM-Poisson and Poisson–Tweedie regression models can accommodate overdispersion. Based on model selection criteria, namely the Akaike Information Criterion (AIC), the Poisson–Tweedie regression model demonstrates the best performance for analyzing the number of measles cases among children under five in Indonesia in 2024. Based on the selected model, measles immunization coverage, the number of people living in poverty, and the percentage of vitamin A supplementation have a statistically significant effect on the number of measles cases among children under five. In contrast, the percentage of exclusive breastfeeding and the percentage of undernourished children under five do not show a statistically significant effect.]]>
