Table 2.

Average discrete change of cannabis use indicators over survey waves, and estimated causal effect of COVID-19 on cannabis use by difference-in-difference, stratified by sex, among eligible students attending the 43 linked-longitudinal COMPASS schools across the three study waves (2018, 2019, 2020)

		Cannabis Use
		Monthly Usea		Weekly Usea		Daily Usea
		Difference Mean (95% CI)		Difference Mean (95% CI)		Difference Mean (95% CI)
		Female	Male	Female	Male	Female	Male
Average Discrete Changeb	2019–2018 (pre-COVID-19 period)	5.0 (0.4, 0.6)	7.9 (6.4,9.4)	2.9 (1.2, 4.5)	5.2 (3.7, 6.8)	0.5 (0.1, 1.0)	1.2 (0.5, 1.8)
	2020–2019 (early COVID-19 period)	0.2 (−0.3, 0.3)	1.1 (− 0.9, 3.0)	0.9 (−1.7, 3.5)	1.1 (− 1.9, 4.1)	1.3 (0.5, 2.0)	− 0.4 (−1.4, 0.5)
		Difference-in-Difference (95% CI)		Difference-in-Difference (95% CI)		Difference-in-Difference (95% CI)
		Female	Male	Female	Male	Female	Male
Estimated Causal Effectb	(2020-2019) – (2019-2018)	−4.8 (−8.2, −1.4) p = 0.006	−6.8 (− 9.7, −3.9) p < 0.001	−2.0 (−5.9, 2.0) p = 0.325	−4.1 (− 8.2, −0.1) p = 0.045	0.8 (−0.2, 1.7) p = 0.138	−1.6 (− 2.9, − 0.3) p = 0.016

Notes: 95% CI (confidence interval)

^a Fixed effect model with a lagged variable as the outcome, controlling for time-invariant confounders but constraining sex and age effects on the outcome to be fixed across time, and sample selection correction with the predictors of age, weekly spending money, school connectedness, and province.

^b based on adjusted estimates.