Testing bid design effects in multiple-bounded, contingent-valuation questions.

摘要

This study investigates the centring of responses and bid design effects in multiple-bounded questions using data from moose hunting surveys conducted in Maine, USA, in 1996 and 1997. In 1996, independent samples of respondents answered either an open-ended or multiple-bounded question. The multiple-bounded question presented a skewed bid design to investigate centring and range effects. In 1997, independent samples answered either a single-bounded question or one of three versions of a multiple-bounded question. The 1997 multiple-bounded question, with three bid designs, provides an additional test of centring and an investigation of bid range effects. The results indicate that multiple-bounded questions are not susceptible to a centring effect.However, multiple-bounded questions with careful bid designs do appear to reduce the fat tails (or yea saying) that has been observed in single-bounded data. These results suggest that the design of bids for multiple-bounded questions must proceed with the same care that has been advocated for single-bounded questions.