Many attempts have been made to identify the variables that condition the destination choices of individuals and groups. Diverse models of spatial choice now incorporate many descriptors of the socioeconomic characteristics of decision-makers, of their cognitive and evaluative processes, and of the objective features of destination alternatives. No single model is generally acceptable. This seems to reflect both the complexity of the destination choice decision and the difficulty of developing a single model to predict the choices of a heterogeneous population group. This paper explores an alternative approach. We assume that the individual's destination choices over time, for a given purpose, may appear to be a random process because of the conflicting or interacting effects of many variables. A simple Bernoulli model is developed to describe this process for a heterogeneous population group. A preliminary test of the model is carried out by using data on successive grocery store choices of a sample of 90 households in Uppsala, Sweden. The model fails to fit sections of the data describing the use of particular stores, the use of different classes of stores, and the behavior of different population groups. The population groups were differentiated by their degree of store familiarity, by their distance from both the stores they used and from all stores, and by their life-cycle stage. The consistent rejection of the model lends some support to new efforts to isolate variables conditioning destination choice, for example, through the application of learning models and perhaps even the multinomial logic models currently used to study mode choice.

Media Info

  • Media Type: Print
  • Features: References; Tables;
  • Pagination: pp 33-44
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 00097404
  • Record Type: Publication
  • ISBN: 0309023742
  • Files: TRIS, TRB
  • Created Date: Aug 13 1975 12:00AM