Fig. 2From: Bayesian negative binomial logit hurdle and zero-inflated model for characterizing smoking intensityComparison of the densities of each model fits a–c number of cigarette smokingBack to article page