Destination Choice Modelling: What's in a Name?

Abstract

Discrete choice modelling can be employed to analyse tourists' holiday destination choices. The data for a destination choice modelling analysis are often obtained from a survey in which prospective tourists indicate their choices in a series of hypothetical scenarios. In that context, an issue deserving investigation is the use of destination names in the stated choice task. Using the names of destinations to label the alternatives in a choice scenario would be expected to enhance the predictive validity of the choice model. On the other hand, the experimental design of the choice scenarios may yield destination descriptions that appear unrealistic to survey respondents. In this paper, the labelling issue is investigated using a choice modelling study of short-break destination choices by residents from Melbourne, Australia. The multinomial logit model is used to compare the results of labelled and unlabelled stated choice tasks. The comparative results are reported and discussed and implications for destination choice modelling applications are drawn.