. 2018 Sep 28;1(5):e182765. doi: 10.1001/jamanetworkopen.2018.2765

Table 2. Differences in Cannabis Use and Sociodemographic Correlates of Cannabis Use by Administration Method.

Regressor	Outcome
	Ever Use^a		Past 30-d Use^a		Days Used in Past 30 d^b
	OR (95% CI)	P Value	OR (95% CI)	P Value	β (95% CI)	P Value
Main Effect of Administration Method^c
Combustible vs vaporized	4.15 (3.67 to 4.70)	<.001^d	3.21 (2.83 to 3.63)	<.001^d	1.75 (0.59 to 2.90)	.003^d
Edible vs vaporized	2.38 (2.04 to 2.78)	<.001^d	1.69 (1.28 to 2.22)	<.001^d	−0.91 (−2.33 to 0.52)	.21
Combustible vs edible	1.74 (1.58 to 1.92)	<.001^d	1.90 (1.56 to 2.31)	<.001^d	2.65 (1.40 to 3.91)	.001^d
Sociodemographic Correlates of Use by Administration Method
Sex × administration method interaction estimate^c^,^e	NA	.027^d^,^f	NA	.003^d^,^f	NA	.70^f
Sex estimates stratified by administration method^f
Female vs male (outcome: combustible)	0.99 (0.85 to 1.16)	.96	0.93 (0.76 to 1.14)	.50	−3.68 (−5.53 to −1.83)	<.001^d
Female vs male (outcome: edible)	1.08 (0.91 to 1.28)	.40	0.87 (0.67 to 1.12)	.28	−2.39 (−4.52 to −0.25)	.03
Female vs male (outcome: vaporized)	0.79 (0.63 to 0.99)	.04	0.62 (0.45 to 0.85)	.004^d	−2.72 (−5.69 to 0.23)	.07
SES × administration method interaction estimate^c^,^e	NA	<.001^d^,^f	NA	<.001 ^d^,^f	NA	.50^f
SES estimates stratified by administration method^f
Low vs high SES (outcome: combustible)	1.71 (1.41 to 2.07)	<.001^d	1.29 (1.01 to 1.67)	.04	−0.97 (−3.12 to 1.17)	.37
Low vs high SES (outcome: edible)	1.49 (1.20 to 1.84)	<.001^d	1.19 (0.86 to 1.64)	.31	0.86 (−1.71 to 3.43)	.51
Low vs high SES (outcome: vaporized)	1.04 (0.79 to 1.38)	.77	0.91 (0.62 to 1.33)	.61	−0.76 (−4.10 to 2.58)	.65
Race/ethnicity × administration method interaction estimate^d^,^e^,^g	NA	.49^f	NA	.62^f	NA	.44^f

Abbreviations: NA, not applicable; OR, odds ratio; SES, socioeconomic status.

^{^a}

Binary logistic regression models in overall sample with data available for all regressors (n = 2701).

^{^b}

Linear regression among past 30-day users (n = 393).

^{^c}

Estimates from generalized linear mixed models of association of sociodemographic characteristics, administration method, and their interaction as simultaneous regressors, adjusted for school random effects and respondents’ age. Main effects of sociodemographic variables are not presented.

^{^d}

Indicates statistically significant with Benjamini-Hochberg correction using a P value threshold of .028 as the criterion of statistical significance.

^{^e}

Interaction terms were added to models one at a time; main effect estimates exclude interaction terms.

^{^f}

Estimates from univariable generalized linear mixed models with school random effects.

^{^g}

Stratum-specific associations of race/ethnicity with cannabis use for each administration method are reported in eTable 4 in the Supplement.