We first examined all continuous variables for normality, and applied appropriate statistical transformations so that the skew for all variables in our path model was less than |1|. A logarithm (base 10) transformation was applied to variables that were positively skewed. The logarithm of zero is undefined. In instances where a skewed variable had zero as a possible score, we added “1” as a constant prior to taking the logarithm [e.g., log10(alcohol use + 1) in order to retain these individuals in our analyses.
