This paper is concerned with the development of models for forecasting and predicting the choices of urban residents for urban recreational and cultural activities and, hence, the travel for such activities. In earlier work, reported elsewhere, some preliminary choice models, using the multinomial logit model, were reported with one-way segmentations on each of geographic location, perceived attractiveness, and stage in the family life cycle. Geographic segmentation was found to be statistically the most significant segmentation scheme which is an undesirable result. It was postulated subsequently, that geographic segmentation was a proxy for social and economic differences in the populations of the geographic units used. This paper reports on the results of two-way segmentations using location as one dimension and various sociodemographic variables as the second dimension. In all cases, it was still found that the effects of location were significant, although seemingly less so than in the one-way segmentation schemes. Lack of a sufficiently large sample prevented the investigation of more complex segmentation schemes than the two-way schemes reported here. It must be concluded that there is no evidence to reject the postulate that geographic location affects recreation-travel behavior.
ASJC Scopus subject areas
- Environmental Science(all)
- Earth and Planetary Sciences(all)