Negative Binomial Regression in Overcoming Overdispersion Poverty Data in Kalimantan

Abstrak

Poverty is one of the problems that Indonesia still faces. Kalimantan Island has large natural resources, but experiences inequality in the distribution of wealth in the region. In this study, data on the number of poor people is used as the dependent variable. The independent variables include the percentage of households that have access to non-PLN electricity (X₁) , access to proper drinking water (X₂) , proper sanitation (X₃) , non-own toilet facilities (X₄) , HDI (X₅) , Open Unemployment Rate (X₆) , average wages of informal workers and main employment (X₇) , population density per km² (X₈) , monthly per capita food and non-food expenditure (X₉) , the percentage of the population experiencing health issues but not seeking treatment due to lack of financial means (X₁₀) , and percentage of the population aged 15 years and above who do not have a diploma (X₁₁)  in 2023. A Poisson regression analysis is employed. The model accounts for the significance of every independent variable. The model was found to have overdispersion, which was resolved through negative binomial regression. The findings of the study revealed that the average wage of informal workers and primary employment, population density per km2, monthly per capita food and non-food expenditure, The percentage of the population experiencing health issues but not seeking treatment due to lack of financial means, and the percentage of the population aged 15 years and older who do not have a diploma all have a significant impact on the magnitude of the number of people living in poverty on the island of Kalimantan.