Table 6.

Logistic regression analyses for associations with reporting being a "lights" smoker*

Variables	Adjusted Odds Ratio (95% CI)
	Model 1 (demo-graphics)	Model 2 (+ socio-demo-graphics)	Model 3 (+ smoking behaviour)	Model 4 (+ smoking beliefs)
Demographic
Age (years)#	1.14 (0.99 – 1.31)	1.13 (0.98 – 1.31)	1.14 (0.98 – 1.32)	1.22 (1.04 – 1.43)
Gender (women vs men)	1.78 (1.20 – 2.65)	1.77 (1.19 – 2.63)	1.57 (1.03 – 2.39)	1.46 (0.96 – 2.24)
Māori vs European	0.53 (0.35 – 0.80)	0.56 (0.37 – 0.84)	0.57 (0.37 – 0.87)	0.55 (0.35 – 0.84)
Pacific vs European	0.14 (0.05 – 0.40)	0.15 (0.05 – 0.44)	0.13 (0.04 – 0.41)	0.11 (0.04 – 0.35)
Asian vs European	2.09 (0.78 – 5.61)	2.15 (0.79 – 5.84)	0.99 (0.42 – 2.35)	0.92 (0.38 – 2.20)
Socio-demographic
Deprivation quintiles (increasing deprivation)	-	0.92 (0.79 – 1.06)	0.94 (0.81 – 1.10)	0.93 (0.79 – 1.09)
Smoking behaviour
RYO smoker (only) vs Others	-	-	0.58 (0.38 – 0.91)	0.57 (0.37 – 0.90)
Heaviness of smoking index (alternate version)** #	-	-	0.86 (0.78 – 0.96)	0.87 (0.78 – 0.96)
Smoking beliefs
Awareness of smoking harm (7-item scale)**	-	-	-	1.82 (1.17 – 2.83)
Holding at least one of the 3 beliefs that "lights" confer health benefits**	-	-	-	1.95 (1.29 – 2.95)

* The aORs in models 2, 3 & 4 are adjusted for the demographic and key socio-demographic variables (ie, deprivation), models 3 & 4 for smoking behaviour variables and model 4 for smoking beliefs. The included variables from the univariate analyses that became not significant in the models (at p < 0.05) were subsequently omitted from the final respective models, except for those considered critical to our conceptual framework (see Methods).

** See Tables 3 and 4 for further details on these indices and α scores (for multi-item indices) and Table 5 for the three questions used for the variable in the last row of the above table.

# Continuous variable and for age the actual age (years) was divided by 10 so as to interpret the OR estimates for an increment of 10 years.