D. Leann Long, PhD, assistant professor in the WVU Department of Biostatistics, has published a new statistical model for correlated count data containing more zeroes than expected, commonly known in biostatistics as ‘zero-inflated’ count data. Dr. Long’s work is now available online in Journal of the Royal Statistical Society: Series C (Applied Statistics). For many fields of research, the new marginalized zero-inflated Poisson model with random effects provides more meaningful interpretations than traditional methods for count data with excess zeroes.
Traditional zero-inflated models used in public health studies usually produce parameters without direct interpretation on the entire population. Long and colleagues have extended their previous marginalized zero-inflated Poisson model to incorporate correlated observations, providing statistical models that more directly answer scientific hypotheses. The article “A marginalized zero-inflated Poisson regression model with random effects” is available here (http://onlinelibrary.wiley.com/doi/10.1111/rssc.12104/abstract).