First, we conducted exploratory analyses through basic
contingency tables with Chi-square tests and the decision
tree (Chaid) with a level of significance of 5%, followed
by logistic regression for complex samples with
all variables used in the bivariate analysis [36]. Analyses
were conducted on data weighted to correct for unequal
probabilities of selection into the sample. The outcome
variable of interest was binge drinking (BD) in the past
month. The independent variables included behaviors
related to leisure, religion and family as well as information
on drug use. Results are presented via weighted
proportions (wgt%), Odds Ratios (ORs), adjusted Odds
Ratios (aORs) and p-values.
The complex samples module of the SPSS Version 15
software was used to perform the analyses.