The Optimal Sample Size for Contingent Valuation Surveys: Applications to Project Analysis
By Arthur H. Darling, William J. Vaughan (04/00, ENV-136, En) See also Environment and Natural Resources
NOTE: This paper can only be downloaded with Internet Explorer version 4.0 or better. If you want a hard copy, please send an e-mail to
ligiaes@iadb.orgAttachments contain spreadsheet solution files in QuattroPro versions 7 and 8 (WB3) and 9 (QPW). To download right click.
Whenever a cost-benefit analysis of a prospective investment has to be undertaken using project benefits that are estimated from household survey data, the size of the survey sample must be specified. The most obvious case is the valuation of environmental amenity improvements that must be obtained through contingent valuation (CV) surveys of willingness to pay. One of the first questions that has to be answered in the survey design process is "How many subjects should be interviewed?" The answer can have significant implications for the cost of project preparation; in Latin America and the Caribbean costs per interview can range from US$20 to US$100.
In the CV literature, the usual way to determine the sample size needed for household interviews is to use some variant of a standard statistical tolerance interval formula. It indicates how many interviews are needed to bring the sample mean of willingness to pay within some pre-specified degree of accord with the population mean, ignoring the benefits of additional sample information and the costs of collecting it. This paper proposes a different approach to CV survey sample size determination that, unlike the classical tolerance interval formula, guarantees that the sample size will be optimal. The paper illustrates both approaches with an example from a real project analysis case, and explains why the classical tolerance interval method is unreliable.
Last updated: 05/08/07
