Testing the external validity of hierarchical conjoint analysis models of recreational destination choice

The primary purpose of this article was to test the external validity of the hierarchical information integration (HlI) approach to modeling recreational choices. The HII approach assumes that individuals evaluate and compare recreation options on cognitive constructs such as "locational convenience," "scenicness, " and so forth. Such constructs are in turn defined by elemental attributes that vary across options. Individuals integrate the attribute information to define constructs and then use the constructs to compare and evaluate options. Individuals are assumed to use personal utility functions to value the construct levels of options, integrating these values into an overall "utility" for each option. Overall preferences, or utilities, for recreation options are revealed by individual responses to surveys that describe hypothetical recreation options as bundles of construct levels. Utilities for the attributes of constructs are revealed by responses to separate construct surveys that describe hypothetical recreation options as bundles of attribute levels associated with particular constructs. A statistical choice model was estimated by combining responses to the separate surveys and applied to predict the expected choice shares for 40 actual recreation areas in a particular area of The Netherlands. We could not reject the hypothesis that the predicted probalities were proportional to be observed choice proportions, which suggests that the HII approach may be a useful and valid way to model such decisions.