ANYTHING YOU CAN DO, WE CAN DO BETTER: A PROVOCATIVE INTRODUCTION TO A NEW APPROACH TO STATED PREFERENCE DESIGN

This paper presents a new method for designing stated preference (SP) experiments. Currently, many SP designs used in practice maintain orthogonality between the attribute levels, because it is though this will produce lower standard errors associated with the parameter estimates. We however contend that the standard errors of the coefficient estimates derived from a logit model are not necessarily minimized when using an orthogonal design. We have obtained expressions for minimizing the variance of the estimated parameters of a logit model which indicate that choices should be offered to respondents which have probabilities of being chosen of 0.917 and 0.083. We have also derived a limit for the t statistic associated with each parameter, so that any design can be assessed against the theoretical optimum in terms of the t statistic it produces. Clearly, an optimal design depends on some foreknowledge of the parameters, but when we subjected the technique to empirical application, we found that our advanced design out performed a standard orthogonal design notwithstanding that the orthogonal design actually produced estimated parameters much closer to the design points than the advanced design.

• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/isbn/0080435904
• Corporate Authors:

  Elsevier

  The Boulevard, Langford Lane
  Kidlington, Oxford  United Kingdom  OX5 1GB
• Authors:
  • TONER, J P
  • CLARK, S D
  • Grant-Nuller, S M
  • FOWKES, A S
• Conference:
  • World Transport Research: Selected Proceedings of the 8th World Conference on Transport Research
  • Location: Antwerp, Belgium
  • Date: 1998-7-12 to 1998-7-17
• Publication Date: 1999

Language

• English

Media Info

• Features: Figures; References; Tables;
• Pagination: p. 107-120

Subject/Index Terms

• TRT Terms: Choice models; Coefficients; Design; Empirical methods; Logits; Orthogonality; Standard error; Statics
• Subject Areas: Design; Highways; Operations and Traffic Management; Planning and Forecasting; Public Transportation;

Filing Info

• Accession Number: 00784018
• Record Type: Publication
• ISBN: 0080435904
• Report/Paper Numbers: Volume 3
• Files: TRIS, ATRI
• Created Date: Feb 14 2000 12:00AM