Table 4.

Model 2: dynamic regression model predicting positive to negative ratio with ARIMA^a error term (1, 0, 0).

Measures			Positive to negative ratio	P value
Predictors of positive to negative ratio, estimate of effect (95% CI)
	Intercept		37.90 (34.60 to 41.20)	<.001
	Statutory holidays (1 for holidays, 0 for nonholidays)		–6.22 (–10.30 to –2.12)	.002
	Business closure (log transformed)		–1.14 (–2.26 to –0.01)	.046
	Regional differences in lockdown
		Regions are in the same stage of lockdown	Reference group	N/Ab
		Regions are in different stages of lockdown	5.75 (2.16 to 9.33)	.001
		Regions are not under lockdown	–10.50 (–18.70 to –2.29)	.01
	New COVID-19 case counts in Canada, excluding Ontario (in hundreds of cases)		0.17 (–0.14 to 0.48)	.29
	New COVID-19 case counts (in hundreds of cases)		–0.98 (–1.81 to –0.16)	.02
	Business closed × new cases (increase of 1 log unit in business closure + 100 new cases)		0.37 (0.04 to 0.70)	.02
Goodness of fit
	With covariates, AICc		1612.43	N/A
	Without covariates, AIC		1723.89	N/A

^aARIMA: autoregressive integrated moving average.

^bN/A: not applicable.

^cAIC: Akaike information criterion.