2.3. Statistical procedures
Analyses were carried out using SPSS 20.0 (IBM). Using frequency of
marijuana use as the outcome, main effects and time-varying effects of
symptoms of depression, frequency of cigarette use, gender, and their
interactions were examined. Mixed-effects regression models using
Generalized Linear Models in SPSS included both fixed effects of
variables and randomeffects to account for the repeated measurements
of participants over time. Randomeffects included intercepts (capturing individual differences in the frequency ofmarijuana use at baseline) and
time trends (capturing individual differences in the rate of change in
frequency of marijuana use across the years of the study), with an
unstructured covariance structure.