Data were assessed for normality both by visual inspection and by Kolmogorov–Smirnov testing. Non-normally distributed data were log transformed, and if still non-normal, categorized as non-continuous variables for further analysis. Males and females were compared with respect to demographics, diagnoses, MA-use variables, severity of MA dependence, and severity of MA withdrawal symptoms by using chi-square test or independent t-test when appropriate. Statistically significant variables in the initial analyses were then controlled in the adjusted model for sex difference by using logistic regression analysis. Demographic, diagnostic, and MA use variables were compared between individuals with and without MA dependence in both the total and sex specific groups by chi-square test or independent t-test. Effects of sex on MA dependence symptoms were tested by logistic regression analysis, controlling for variables identified as distinguishing men and women in the above analyses. Factors that achieved statistical significance (p<0.05) or were nearly significantly (p<0.10) associated with MA dependence in each sex were subsequently analyzed using logistic regression to identify factors that predicted the trait in males and females.
