Analyses were conducted using Stata Version 11 (StataCorp, 2009). Multilevel mixed-effects regression analyses were used to account for the nesting of individuals within states. Including state as a fixed effect helps to account for potential confounding by unmeasured state-level variables that may be correlated with our key variables of interest. While NESARC sampling weights have been constructed to provide nationally representative sample estimates, they were not necessary to utilize in the current study because the analyses were concerned with relative effect sizes, and not making population-level estimates of the number of people affected (Korn et al., 1991).
