Negative binomial regression is used for modeling count variables, usually for over-dispersed count outcome variables (e.g. accounting for the extra-Poisson variation). BESH Stats fits negative binomial regression with the traditional NB2[1] parametrization