In this paper, three sampling techniques for calibrating disaggregate travel demand models are considered: random, stratified, and choice-based sampling. In a random sample, the probability of all members of the population being in the sample is equal; in a stratified sample, the population is divided into groups based on one or more characteristics and each group is sampled randomly but at different rates; and in a choice-based sample, the number in the sample selecting each alternative is predetermined, i.e., the sample is based on the outcome of a behavioral choice process. Existing disaggregate choice calibration methods yield consistent parameter estimates for random and stratified sampling techniques. Although maximum likelihood estimation for the third technique is extremely complex, an alternative, tractable estimator whose estimates are both consistent and asymptotically normal exists. This new estimation technique can be applied by using existing capabilities in ULOGIT or other multinomial logit estimation programs with only minor revisions. This implies that choice-based samples such as on-board surveys and roadside interviews can now be used for disaggregate model calibration. This should substantially reduce the cost of data collection in disaggregate model development. In addition, it opens an entire range of questions regarding the most appropriate sample design for future data collection efforts oriented toward the development of disaggregate choice models for urban travel demand forecasting. /Author/

Media Info

  • Media Type: Digital/other
  • Features: References;
  • Pagination: pp 24-28
  • Monograph Title: Perception and values in travel demand
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 00145326
  • Record Type: Publication
  • ISBN: 0309025613
  • Files: TRIS, TRB
  • Created Date: Feb 16 1981 12:00AM