Ever users were more likely to have tried such marijuana, cocaine, hallucinogens, and prescription medications than those who had never tried a CCLC, whereas past 30-day CCLC use was not related to the likelihood that an individual had ever tried other substances of abuse (see Supplementary Table 1). Mental Health Ever users and past 30-day users reported higher levels of anxiety and antisocial behavior than nonusers. Past 30-day users also self-reported higher levels of depression than those who had not used CCLC in the past 30 days (see Supplementary Table 1). Predicting Ever and Past 30-Day CCLC Use We then conducted a multivariate binary logistic regression following a manual backward stepwise approach in order to determine the most parsimonious main effects model.
We started by placing all effects that were significant bivariate predictors in the model for initial consideration. Stepwise elimination of nonsignificant effects occurred by sequentially removing variables in order of largest p value. After each effect was removed, the model was refit with the remaining items. This process continued until a final model consisting only of variables that explained significant partial variability in CCLC use (p < .05) was ascertained. Only the final models with significant predictors are reported (Table 1). The significant predictors of ever CCLC use while controlling for the others in the model included being male, greater nicotine dependence (FTQ), use of hookah in the last month, ever use marijuana, ever use of prescription medications, and use of alcohol in the last 90 days.
From the array of variables included in the second multivariate model, being male, use of hookah in the last 30 days, and higher antisocial tendencies all were associated with an increased likelihood of being a past 30-day CCLC user. Table 1. Model Fit at Final Stage of Backward Elimination for Predicting Ever and Past 30-Day Cigars, Cigarillos, and Little Cigars (CCLC) Use Controlling for Frequency of Past Month Cigarette Use Finally, multivariate linear regressions were conducted to determine whether 30-day CCLC use was related to our correlates of interest after adjusting for frequency of past 30-day cigarette use. We separately regressed past 30-day CCLC use (independent variable) on each tobacco use, substance use, and mental health variable (separate dependent variables), controlling for frequency of past 30-day cigarette use.
Additionally, we also controlled for gender and GPA given that bivariate analyses showed that 30-day CCLC users and nonusers were different along these background demographic dimensions. Results AV-951 reveal that having smoked CCLC in the past month and frequency of past month cigarette use both accounted for unique variance in predicting concurrent use of bidis, kreteks, and hookah, frequency of past 90-day alcohol and marijuana use, and alcohol problems (p��s < .
Related posts: