Does Specification Error Explain the Discrepancy Between Open-Ended and Dichotomous Choice Contingent Valuation Responses? A Comment on Monte Carlo Benchmarks for Discrete Valuation Methods by Ju-Chin Huang and V. Kerry Smith
Poe, Gregory L.; Vossler, Christian A.
In this working paper we demonstrate that some of the statistical tests used by Huang and Smith in a recent Land Economics article (74(2 1998): 186-202) were erroneous, and raise concerns about their corresponding conclusions. Specifically, using data from one of the studies that they showcase, we demonstrate that Huang and Smith’s analysis suggesting statistical equality between hypothetical dichotomous choice responses and actual contributions is incorrect. We further show that their purported equality between dichotomous choice and open-ended response formats is unfounded. Based on these analyses we conclude that when real humans make real or stated decisions, the observed procedural variance across elicitation methods and the degree of hypothetical bias are more fundamental than relying on alternative econometric specifications.
WP 2001-18 September 2001
Funding for this research was provided by USDA Regional Project W-133 and Cornell University.
Charles H. Dyson School of Applied Economics and Management, Cornell University