Table 3.

Comparison of spatial and non-spatial models for predicting total mortality during days without prolonged dust events

Variables	Multivariate linear: predicted mortality on non-dusty days (95% confidence intervals)	Spatial error (lag 1): predicted mortality on non-dusty days (95% confidence intervals)	Spatial error (lag 2): predicted mortality on non-dusty days (95% confidence intervals)	Spatial error (lag 3): predicted mortality on non-dusty days (95% confidence intervals)
SVF (in 10%)	−1.7 [−2.1, −1.4]*	−1.9 [−2.3, −1.5]*	−2.0 [−2.4, −1.6]*	−1.8 [−2.1, −1.4]*
% low education (in 10%)	1.5 [0.8, 2.2]*	1.5 [0.7, 2.2]*	1.4 [0.7, 2.2]*	1.4 [0.7, 2.1]*
% low income (in 10%)	0.8 [0.2, 1.4]*	0.6 [0.0, 1.2]	0.7 [0.1, 0.3]*	0.8 [0.2, 1.4]*
% elderly (in 10%)	−2.1 [−3.1, −1.1]*	−1.8 [−2.9, −0.7]*	−1.9 [−2.9, −0.9]*	−1.9 [−3.0, −0.9]*
Lambda	N/A	0.3 [0.2, 0.5]*	0.4 [0.2, 0.6]*	0.2 [−0.1, 0.5]
AIC	1612.48	1582.04	1585.25	1610.28

* Are the results with significant p values (<0.05)