We first evaluated the influence of smoke-free bar and restaurant legislation on transitions in AUD status across two waves of NESARC data using multiple logistic regression, after adjusting for a number of sociodemographic covariates. There was no effect of days since ban on AUD transitions among individuals in states with smoke-free legislation, and time since ban was not included in regression analyses. Next, we conducted a series of stratified logistic regression analyses to examine the influence of smoke-free policy change on AUD transitions separately by smoking status, sex, and age. Third, we investigated the influence of smoke-free legislation on changes in smoking status and evaluated whether any effects of smoke-free legislation on AUD transitions would remain after adjusting for changes in smoking status. This was done by adding three dummy variables (smoking remission, smoking onset and no smoking) as predictors to our logistic regression analyses. Persistent smoking (i.e., smoking at both Waves) was used as the reference group. All analyses were estimated in both the entire sample and the subset of public drinkers. Supplementary analyses with the full sample that
