Count Data Modeling Under Over Dispersion Issue: A Comparative Study

Abstract

Statistical modelling of count data has been of extreme interest to researchers. However, in practice, the count data is often identified with overdispersion or underdispersion. The Conway-Maxwell-Poisson regression model (CMPR) has been proven powerful in modelling count data with a wide range of dispersion. In this study, the performance of CMPR is tested under different value of dispersions. Our Monte Carlo simulation results suggest that the CMPR can bring significant improvement relative to Poisson regression model, in terms of AIC, BIC, and Deviance.