Table 2. Spatial lag models estimated via maximum likelihood to predict log-transformed COVID-19 case distribution between Iranian provinces (model including all variables and model obtained through a selection scheme).

Variable	Coefficient		SE		z-value		p-value
	All variables*	Selection scheme**	All variables	Selection scheme	All variables	Selection scheme	All variables	Selection scheme
Spatial parameter (ρ)	0.723	0.737	0.107	0.104	6.734	7.069	<0.001	<0.001
Model constant	2.510	2.853	2.550	2.425	0.984	1.176	0.325	0.239
Urban population (%)	0.026	0.026	0.011	0.010	2.345	2.653	0.019	0.008
Population aged ≥60	0.383	0.331	0.075	0.062	5.089	5.324	<0.001	<0.001
Population density	-0.0002		0.0007		-0.258		0.797
Literacy	-0.110	-0.103	0.040	0.040	-2.771	-2.591	0.006	0.010
Average temperature (°C)	0.113	0.114	0.030	0.028	3.748	4.105	<0.001	<0.001
Precipitation levels (mm)	-0.0003		0.0003		-0.98		0.327
Physician distribution	0.0008	0.0008	0.0001	0.0001	5.740	5.746	<0.001	<0.001
GDP	-0.051	-0.057	0.038	0.034	-1.343	-1.699	0.179	0.089
Consumer Price Index	0.032		0.039		0.812		0.417
TEI	0.157	0.155	0.074	0.074	2.112	2.081	0.035	0.038

*Likelihood Ratio Test = 15.628, p< 0.001 (no spatial vs spatial model); R2 = 0.877; AIC = 57.165; σ² = 0.146

** Likelihood Ratio Test = 18.682, p< 0.001 (no spatial vs spatial model); R2 = 0.872; AIC = 52.704; σ² = 0.152