ESTIMATION USING CONTINGENT VALUATION DATA FROM A DICHOTOMOUS CHOICE WITH FOLLOW-UP QUESTIONNAIRE

Dichotomous-choice (referendum) contingent-valuation questions are ine fficient in that a very large number of observations is required to id entify a distribution of resource values with any degree of accuracy. An alternative questioning strategy introduces a follow-up dichotomous -choice question. We generalize upon previous analyses of this type of data by relaxing the assumption that the identical unobserved resourc e value motivates both responses. While values implied by the first an d second responses are highly correlated and may be drawn from the sam e distribution, they are definitely not identical. Furthermore, assumi ng that they are can severely distort the estimated valuation distribu tion. (C) 1994 Academic Press, Inc.