Table 2. Importance weights and results of model averaging for predictor variables in spatial modeling of EEEV exposure risk in sentinel chickens in Walton County, Florida in 2009 and 2010.

	Model-averaged	Weighted	95% Confidence Interval		Importance
Variable	Estimate	Unconditional SE	Lower	Upper	Weight
Intercept	0.0113	0.0019	0.0075	0.0150	1.00
NOCA	0.0066	0.0021	0.0025	0.0107	0.97
mel	0.0028	0.0020	−0.0012	0.0068	0.41
avian	0.0016	0.0024	−0.0032	0.0063	0.23
EUST	−0.0006	0.0022	−0.0049	0.0037	0.19
Intercept	0.0102	0.0019	0.0065	0.0140	1.00
NOCA	0.0040	0.0025	−0.0010	0.0090	0.50
mel	0.0025	0.0020	−0.0015	0.0064	0.38
avian	0.0025	0.0022	−0.0017	0.0068	0.36
EUST	−0.0011	0.0021	−0.0051	0.0029	0.21

Results below the dotted line are based on the development of models using the dataset that excluded an influential observation.

Variable names: mel = Culiseta melanura abundance, avian = avian species richness, EUST = European starling density, NOCA = northern cardinal density.