Table 3.

Multiple Hazardous Waste Model Performance Using Matched Houses

				Estimated WTP: Motorola					Estimated WTP: Landfill
Model	Sample	Specification	Instruments^{^a}	Mean	Median	5%	95%	% Error^{^b}	Mean	Median	5%	95%	% Error^{^b}
True WTP	Full			9.8	3.7	-8.9	43.0	_	41.3	41.8	16.3	66.6	_
Difference	Full	Linear		7.1	0.0	0.0	28.5	-27.5	7.6	0.0	0.0	30.4	-81.7
IV-Difference	Full	Linear	Mot, Clean	10.7	0.0	0.0	42.9	9.1	10.7	0.0	0.0	42.9	-74.1
Difference	Full	Semilog		6.9	0.0	0.0	27.1	-30.0	7.6	0.0	0.0	30.2	-81.5
IV-Difference	Full	Semilog	Mot, Clean	10.7	0.0	0.0	42.4	9.6	10.7	0.0	0.0	42.4	-74.0
True WTP	Motorola			8.3	6.5	0.0	23.0	0.0
Difference	Motorola	Linear		6.0	3.8	0.4	17.7	-27.9
IV-Difference	Motorola	Linear	Mot, Clean	9.0	5.7	1.6	26.7	8.5
Difference	Motorola	Semilog		5.9	3.8	0.4	18.6	-28.1
IV-Difference	Motorola	Semilog	Mot, Clean	9.3	5.9	0.7	29.2	12.4
True WTP	Landfill								31.6	30.3	11.8	51.3	—
Difference	Landfill	Linear							23.9	19.3	15.3	46.8	-24.3
IV-Difference	Landfill	Linear	Mot, Clean						33.8	27.3	21.6	66.1	7.0
Difference	Landfill	Semilog							24.0	19.5	15.2	46.1	-24.2
IV-Difference	Landfill	Semilog	Mot, Clean						33.6	27.4	21.4	64.8	6.5

Note: WTP, willingness to pay.
↵a "Mot" indicates identification of original houses located near Motorola waste site. "Clean" indicates a house located near the cleaned landfill.
↵b Error computed based on mean WTP.