Table 5.

Association between the Indian scale ^a and tobacco use, by grade level; Delhi, India (n=3512).

	8^th graders (n=2163)			10^th graders (n=1349)
	β	(SE)	p-value ^b	β	(SE)	p-value ^b
*Ever use*
Any kind of tobacco	−0.024	0.012	0.053	−0.066	0.014	<0.001
Chewing tobacco	−0.020	0.014	0.152	−0.068	0.016	<0.001
Smoking bidis	−0.046	0.020	0.020	−0.074	0.020	<0.001
Smoking cigarettes	−0.042	0.016	0.009	−0.064	0.016	<0.001
*Past year use*
Any kind of tobacco	−0.029	0.017	0.086	−0.090	0.018	<0.001
Chewing tobacco	−0.035	0.020	0.082	−0.096	0.019	<0.001
Smoking bidis	−0.014	0.028	0.621	−0.136	0.028	<0.001
Smoking cigarettes	−0.050	0.023	0.033	−0.100	0.021	<0.001
*Past month use*
Any kind of tobacco	−0.041	0.020	0.045	−0.112	0.020	<0.001
Chewing tobacco	−0.036	0.026	0.169	−0.118	0.022	<0.001
Smoking bidis	−0.034	0.032	0.287	−0.128	0.026	<0.001
Smoking cigarettes	−0.050	0.025	0.049	−0.113	0.023	<0.001
*Intentions to use*
Chew in college	−0.070	0.015	<0.001	−0.063	0.017	<0.001
Chew when adult	−0.072	0.014	<0.001	−0.065	0.017	<0.001
Smoke in college	−0.072	0.015	<0.001	−0.073	0.018	<0.001
Smoke when adult	−0.059	0.013	<0.001	−0.072	0.017	<0.001

Models adjust for the effect of one dimension (e.g., Western) of “westernization” on tobacco use, when considering the effect of the other (e.g., Indian).

p-value represents test of the association using a random intercept mixed effects regression model adjusted for gender, school type, and age. The models for past year use of bidis and past month use of bidis were adjusted for school type and age only, so that these models could converge appropriately, as well.