Would You Choose Your Preferred Option? Comparing Choice and Recoded Ranking Experiments

Abstract

Previous research has shown that results from a choice experiment are statistically different from those obtained from a ranking experiment that is recoded and treated as a choice experiment using only the first rank. By avoiding some of the shortcomings of previous comparisons, we obtain the opposite results using data from the valuation of a cork oak reforestation program in the south of Spain. Structural models and welfare estimations are statistically indistinguishable irrespective of the use of parametric or bootstrapping tests. Further, we employ follow‐up questions and subsample analysis to test whether divergences appear when potential effects are isolated.