Revealed and stated preference analysis of ski resort attractiveness

The paper compares a revealed and stated preference approach for estimating the site attractiveness of Vermont ski resorts to weekend skiers. Unexpected bias in attractiveness estimates, seemingly caused by the geometry of origins and destinations, is discovered when the revealed preference approach is used. The attractiveness scales of resorts, which were estimated using separate metric (Thurstone) and nonmetric (nonmetric scaling) models of preference, are shown to be very similar despite the different assumptions of the two models. Eighty‐six percent of variance in the attractiveness estimates of resorts is explained by only two variables, mean perceived length of slopes and mean perceived level of crowding. A multidimensional scaling analysis of a matrix showing the mean cognitive similarity of each pair of 23 resorts reveals that the resultant two‐dimensional cognitive map of the resorts can be largely characterized by the same two attributes, slope length and crowding, that explain the preference‐based attractiveness estimates.