SIMULATION AS A TOOL FOR INVESTIGATING MODEL BIAS

Two case studies are presented of the way in which simulated data sets can assist in understanding bias in transport models. First, an aggregate time series data set is produced for public transport demand, based on a model in which age-related elasticities and the process of ageing are explicitly considered. When a conventional form of demand model is "fitted" to these data, satistical fit is excellent but the elasticities are underestimated. Secondly, a disaggregate cross section data set is produced for mode choice, based on a model where individuals make choices according to generalized cost but have a resistance to change. When values of time are estimated without considering this inertia, it is found that "best fit" values are biassed downwards if journey times have changed, and upwards if costs have changed. (Author)

• Corporate Authors:

  University of Oxford

  Transport Studies Unit, 11 Bevington Road
  Oxford,   United Kingdom  OX2 6NB
• Authors:
  • Goodwin, P B
• Publication Date: 1984

Media Info

• Pagination: 11 p.

Subject/Index Terms

• TRT Terms: Age; Demand; Disaggregate analysis; Mathematical models; Mechanical elasticity; Public transit; Simulation; Statistical bias; Transportation; Transportation planning; Travel time
• Uncontrolled Terms: Transportation models
• Old TRIS Terms: Disaggregate models; User characteristics
• Subject Areas: Planning and Forecasting; Transportation (General);

Filing Info

• Accession Number: 00386727
• Record Type: Publication
• Report/Paper Numbers: TSU/REF-233
• Files: TRIS
• Created Date: Jul 30 1984 12:00AM