Accounting for misreporting in spatial zero-Inflated Poisson models

A bayesian modeling approach based on the zero-inflated Poisson (ZIP) framework is proposed to accommodate for the possible presence of misreporting and spatial dependencies in count data with excess of zeros. Structured spatial effects are introduced in the model specification to capture correlation and heterogeneity of counts while controlling for possible sources of measurement error. To evaluate the model performance, a simulation study is carried out under configurations that allow for alternative structures of the random effects (i.e. structured and unstructured spatial effects). Bayesian Markov Chain Monte Carlo (MCMC) inference is implemented using the NIMBLE package in R.