2002 World Congress of Environmental and Resource Economists

Previous tests of the validity of benefit transfer have tested the null hypothesis of equality between mean WTP estimates. When rejected, they implicitly assume that the alternative hypothesis of inequality is true. In many cases the null hypothesis of equality is rejected when transfer errors are small, while studies producing benefit estimates with high variance, often due to small sample sizes, pass the benefit transfer tests in spite of large transfer errors. We argue that non-rejection of a null hypothesis of equality cannot be used as evidence of its truth, and suggest equivalence testing as more appropriate when testing the validity of benefit transfer. Equivalence testing is better than traditional testing because it combines the concepts of statistical significance and policy significance into one test, by defining an acceptable transfer error prior to conducting the validity test. The level of acceptable transfer error will depend on the intended policy use.



	Paper Title:	Validity Tests Of Benefit Transfer: Are We Performing The Wrong Tests?
	Presenting Author:	Stale Navrud (Agricultural University of Norway)
	Coauthor 1:	Dadi Kristofersson
	Coauthor 2:
	Coauthor 3:
	Abstract:	Previous tests of the validity of benefit transfer have tested the null hypothesis of equality between mean WTP estimates. When rejected, they implicitly assume that the alternative hypothesis of inequality is true. In many cases the null hypothesis of equality is rejected when transfer errors are small, while studies producing benefit estimates with high variance, often due to small sample sizes, pass the benefit transfer tests in spite of large transfer errors. We argue that non-rejection of a null hypothesis of equality cannot be used as evidence of its truth, and suggest equivalence testing as more appropriate when testing the validity of benefit transfer. Equivalence testing is better than traditional testing because it combines the concepts of statistical significance and policy significance into one test, by defining an acceptable transfer error prior to conducting the validity test. The level of acceptable transfer error will depend on the intended policy use.
	Link to paper:	Not available

	Session / Day / Time	6H / Tuesday / 10:15 - 11:45 am

	Please use your browser's "back" button to get to the previous page.