A Flexible Mixed Model for Clustered Count Data

Clustered count data are commonly modeled using Poisson regression with random effects to account for the correlation induced by clustering. The mixed model allows overdispersion via nature of within-cluster correlation, however, departures from equi-dispersion may also exist due underlying process mechanism. We study cross-sectional COM-Poisson model—a generalized in light dispersion—together analysis clustered data. demonstrate flexibility intercept model, including choice effect distribution, simulated and real examples. find that models provide comparable fit well-known associated special cases discrete data, result improved intermediate levels over- or underdispersion Accordingly, proposed useful capturing dispersion not consistent used statistical models, serve as a practical diagnostic tool.