. 2022 Jan 19;83(1):55–63. doi: 10.15288/jsad.2022.83.55

Table 3.

Hierarchical linear models predicting change in any cannabis use and number of days using cannabis from pre-closure to post-closure-1 and from post-closure-1 to post-closure-2 (N = 223)

Variable	Change in any cannabis use				Change in cannabis use days
Variable	B	OR	[95% CI]	P	B	OR	[95% CI]	P
Residential change from pre-closure to post-closure-1
Stayed dependent	−1.22	0.30	[0.09,1.00]	.049	0.30	1.34	[0.61, 2.94]	.459
Stayed independent	−0.44	0.64	[0.33, 1.27]	.200	0.36	1.44	[0.81, 2.57]	.218
Moved to dependent	−0.87	0.42	[0.23, 0.78]	.006	−0.41	0.66	[0.37, 1.16]	.148
Residential change from post-closure-1 to post-closure-2
Stayed dependent	0.25	1.28	[0.73, 2.26]	.385	0.12	1.13	[0.60, 2.15]	.706
Stayed independent	0.12	1.13	[0.67, 1.88]	.650	0.05	1.05	[0.63, 1.75]	.844
Moved to dependent	0.15	1.17	[0.66, 2.06]	.593	−0.26	0.77	[0.41, 1.45]	.416
Moved to independent	0.19	1.21	[0.68, 2.15]	.514	−0.09	0.92	[0.49, 1.73]	.792
Covariates^a
Sex (ref. = male)	−0.19	0.83	[0.51, 1.37]	.457	−0.19	0.83	[0.36, 1.93]	.664
Age	−0.02	0.98	[0.74, 1.29]	.882	−0.20	0.82	[0.49, 1.37]	.442
School A (ref. = not in school)	−0.75	0.47	[0.13, 1.78]	.268	−1.09	0.34	[0.04, 2.92]	.322
School B (ref. = not in school)	0.23	1.26	[0.36, 4.43]	.713	1.37	3.93	[0.54, 28.91]	.177
School C (ref. = not in school)	−0.13	0.88	[0.31, 2.52]	.806	−0.75	0.47	[0.10, 2.19]	.337
Other school (ref. = not in school)	−0.58	0.56	[0.16, 1.94]	.357	−0.54	0.58	[0.08, 4.51]	.601
Cannabis legalization (ref. = no)	−0.18	0.83	[0.37, 1.90]	.665	−0.09	0.91	[0.19, 4.37]	.907

Notes: Bolded effects represent p < .05. Results come from separate models in which the referent group was switched to obtain group-specific effects. OR = odds ratio; CI = confidence interval; ref. = referent group. Although not tabled for parsimony, models also included each Level 2 variable that was entered as a predictor of a Level 1 slope effect (i.e., dummy codes for change in living situation) as an additional predictor of the model intercept, to appropriately test the cross-level interactions.

^{^a}

Effects of covariates are presented from the model in which Stayed Dependent was the referent group, but remained nonsignificant across all models.