Table 2

Smoking by young teenagers. Ordered logistic regression model estimation results

CategoriesFrequency (%)
Effective sample size: 1697. The odds ratios are calculated to compare “smokes daily” with “does not smoke”. Dependent variable, 0 for no smoke, 1 for smoke less than once a week, 2 for smoke on weekends, 3 for smokes daily. Variables contents: sex (0 male; 1 female), interest in school (ordinal for interest in school; 1 very little; 2 little; 3 quite; 4 high), alcohol consumption occasionally (1 yes; 0 no), alcohol consumption frequently (1 yes; 0 no), Smokers at home (percentage of smokers at home), best friend smokes (1 yes; 0 no). Odds ratios 95% confidence intervals in parentheses. *Odds ratios for categories extremes, smokes daily versus no smoke.
Dependent variable
“Do you smoke?”No smoke1456 (85.8)
Yes, less than once a week83 (4.9)
Yes, on weekends48 (2.8)
Yes, smokes daily110 (6.5)
Explanatory variables Unadjusted odds ratios Adjusted odds ratios (95% CI)
Sex (female/male*)1.751.85 (1.33 to 2.59)
Interest in school (high/very little*)0.170.23 (0.19 to 0.27)
Alcohol consumption (occasionally/never*)21.745.88 (3.95 to 8.75)
Alcohol consumption (more than once a week/never*)58.309.81 (6.12 to 15.75)
Number of smokers at home (one/none*)1.772.03 (1.22 to 3.36)
Best friend smokes (yes/no*)24.746.96 (4.93 to 9.84)
Log likelihood−696.0774098
χ2 for covariatesp=0.0001