Presents a letter to the editor commenting on W. Thomson and colleagues' study about the application of zero-inflated Poisson and zero-inflated negative binomial regression methods for modeling both cross-sectional and longitudinal dental caries data obtained from the same cohort study.