A New Microeconomic Theory-Based Model for Ranking Data

This project will address the importance of flexible specifications for the utility kernel error terms for rank-ordered data models. This project will also explore why, just like in the rank-ordered logit (ROL) model, a mixed ROL that superimposes a distribution on the variable coefficients cannot be expected to resolve the problem of unstable coefficients across rank depths. Also, extending the mixed ROL in the ways that the ROL has been extended result in the corresponding models not being based on microeconomic theory. The project will instead adopt a finite-mixture approach to specify random coefficients on the variables as well as on the kernel error term, while using a multivariate normal distribution (including the kernel error term) within each mixture. As importantly, we propose the use of a robust composite marginal likelihood (CML) approach that guarantees estimator consistency under usual regularity conditions, while also entailing no more than the evaluation of bivariate cumulative normal distribution functions in the case of cross-sectional data, regardless of the number of random coefficients or number of alternatives. In the case of repeated ranking exercises, as is typical in stated preference surveys, the project proposes the MACML approach, which again entails the evaluation of no more than two-dimensional cumulative normal distribution functions. The project demonstrates an application of our formulation and estimation approach to study bicycle commute route choice. The use of non-motorized modes for commuting presents many benefits both to society (reduction in congestion and vehicle emissions) and to the individual (health benefits from an active lifestyle). Therefore, encouraging the use of bicycles through the provision of adequate infrastructure is of vital importance. In order to better plan for such infrastructure, it is necessary to first understand how bicyclists make route decisions, what their preferences regarding route attributes (such as pavement condition, presence of big uphills, or travel time) are, and what determines such preferences.


  • English


  • Status: Completed
  • Contract Numbers:


  • Sponsor Organizations:

    Office of the Assistant Secretary for Research and Technology

    University Transportation Centers Program
    Department of Transportation
    Washington, DC  United States  20590
  • Project Managers:

    Bhat, Chandra

  • Performing Organizations:

    Data-Supported Transportation Operations and Planning Center

    University of Texas at Austin
    Austin, TX  United States  78701
  • Principal Investigators:

    Bhat, Chandra

  • Start Date: 20170501
  • Expected Completion Date: 20180930
  • Actual Completion Date: 0
  • Source Data: 148

Subject/Index Terms

Filing Info

  • Accession Number: 01634955
  • Record Type: Research project
  • Source Agency: Data-Supported Transportation Operations and Planning Center
  • Contract Numbers: DTRT13-G-UTC58
  • Files: UTC, RIP
  • Created Date: May 18 2017 9:10PM