Auctions, experiments and contingent valuation

Laboratory experiments are a welcome development that promise to make a significant contribution to the study of economic behavior (Heiner, 1985; Plott, 1982; Smith, 1982, 1985). While relatively new to economics, controlled laboratory studies of human behavior have long constituted the core of experimental psychology where there is a wealth of methodological experience from which experimental economists might profitably draw. In particular, current attempts to apply laboratory auction methods to valuing nonmarket environmental goods would be improved by more rigorous experimental design and data analysis procedures. Close scrutiny of the C/S experiment on valuing a private good revealed that: 1. Despite the experimenters' intentions to the contrary, it is unlikely that strategic behavior was eliminated. 2. Reanalysis of statistical tests using more appropriate procedures produced significantly different results from those obtained in the original study. 3. The WTP and CD conditions were not equivalent, and hence it is impossible to know how to interpret a comparison of their results. Despite these shortcomings, the auction procedure employed by C/S includes some features that undoubtedly will prove important in helping to obtain valid measures of value for nonmarket environmental goods. In conjunction with quite different approaches, including the lottery experiments of Knetsch and Sinden (1984), the field trials of Bishop and Heberlein (Bishop et al., 1984), and the response mode studies of Brown (1984), carefully designed experimental auctions should make a significant contribution to improving our methods for the economic evaluation of nonmarket goods. Copyright Martinus Nijhoff Publishers 1987