Table 2.

Odds ratios (95% confidence intervals) for association of restrictions with stages of smoking uptake. Odds ratios are not making comparisons with baseline level (non-susceptible non-smokers) but with previous level of smoking stage

	Susceptible non-smoker threshold (95% C)I	Early experimenter threshold (95% CI)	Advanced experimenter threshold (95% CI)	Established smoker threshold (95% CI)
Public place restrictions	0.96 (0.86 to 1.06)	0.93 (0.84 to 1.02)	0.92*† (0.83 to 1.00)	0.90*† (0.81 to 0.98)
Total home ban	0.64* (0.52 to 0.76)	0.69*† (0.59 to 0.79)	0.71*† (0.60 to 0.82)	0.78*† (0.67 to 0.90)
Some home restrictions	0.83* (0.74 to 0.92)	0.83* (0.74 to 0.92)	0.83* (0.74 to 0.92)	0.83* (0.74 to 0.92)
School ban	0.92 (0.77 to 1.08)	0.98 (0.85 to 1.10)	1.07 (0.93 to 1.21)	1.22*† (1.07 to 1.37)
Enforced school ban	0.89* (0.85 to 0.99)	0.89* (0.85 to 0.99)	0.89* (0.85 to 0.99)	0.89* (0.85 to 0.99)

2logL=35 559.3 (df=57), intracluster correlation=0.042, cluster variance =0.143, P<0.0001. 

Odds ratios are adjusted for school grade, sex, race, adult smokers in home, and sibling smokers; n=14 977. 

^*Odds ratio significantly different from 1.0 (P<0.05). 

^†Odds ratio significantly different from odds ratio at first threshold (P<0.05).