Table 4.

Logistic Regression of Association Between Demographic and Smoking-Related Variables and Frequency of Crushing the Capsule

Base: factory-made cigarette smokers (18 and over), who smoke capsule cigarettes, have smoked in the last month, and have ever crushed the capsule (344a, unweighted)	1 = Crush capsule always/most of the time (258)
	0 = Crush capsule rarely/sometimes (86)
	Adj ORb	95% CI lower	95% CI upper	p
Gender
Male	ref
Female	1.973	1.143	3.408	.015
Age group				.001
18–24	ref
25–34 vs. 18–24	0.754	0.322	1.770	.517
35–44 vs. 18–34	0.609	0.299	1.238	.171
45–54 vs. 18–44	1.021	0.453	2.299	.961
55+ vs. 18–54	0.204	0.095	0.440	<.001
Social grade
ABC1	ref
C2DE	0.816	0.450	1.481	.504
Ethnicity				.042
White British	ref
Other White background	0.423	0.184	0.974	.043
Other	0.421	0.163	1.088	.074
Heaviness of Smoking Index (HSI, 0 = light, 6 = heavy)				.006
0	ref
1 vs. 0	0.572	0.242	1.354	.204
2 vs. (0–1)	0.543	0.261	1.131	.103
3 vs. (0–2)	0.615	0.308	1.225	.167
4–6 vs. (0–3)	0.310	0.127	0.759	.010
Whether planning to quit in next 6 mo
No	ref
Yes	1.105	0.639	1.960	.722
	Test of model coefficients: χ2 = 55.739, df = 13, p < .001. Nagelkerke R2 = 0.222.
	Hosmer & Lemeshow χ2 = 5.283, df = 8, p = .727.

95% CI = 95% confidence interval; Adj OR = adjusted odds ratio; ref = reference category.

^a15 missing cases because of missing data across social grade, ethnicity, and Heaviness of Smoking Index where cell sizes were too small to include the “not stated responses” as separate categories.

^badjusted for all other variables in the model.