A TWO-PERIOD ANALYSIS OF COMMUTER MODE CHOICE: THE PREDICTIVE CAPABILITY OF INDIVIDUAL CHOICE MODELS

The estimation of travel demand is considered to have assumed an increasingly important role in public-policy decision making, with today's ever-tightening government budgets and increasing awareness of the energy costs of alternative transport systems. It is therefore considered to be of paramount importance to estimate as accurately as possible future demands on transport networks if enlightened decisions concerning these networks are to be made. In this paper the important question of the accuracy of the predictions of transport demand models is considered, and the predictive capabilities of individual binary mode choice models using a two-period data set of commuters in the Sydney metropolitan area is investigated. The prediction test procedure is outlined, and shown applied with data collected in 1973 in a model estimated on 1971 data. The predictive capability of individual mode choice models is illustrated and discussed by reference to three studies: A 1970 study which used data gathered for the study of mode choice in the Chicago northwest corridor where the choice was between the car and a high quality suburban rail service; a 1973 study concerned with modal choice between car and train for medium-range intercity work in the Forth-Clyde corridor of the central lowlands of Scotland; and a 1974 study conducted in Sweden which estimated a model on survey data for commuter work trips in the Stockholm metropolitan area in 1968 and then tested it against a similar transportation survey in 1971 in the same area. The authors conclude that this two-period analysis has demonstrated the predictive capabilities of behavioural modelling procedures that adopt the individual traveller as the unit of analysis, and that it has also provided a benchmark for emphasizing the need to continue a data-collection effort every two to three years so that system changes over time can be investigated and an improved knowledge of traveller behaviour over time can be gained. /TRRL/

  • Corporate Authors:

    University of British Columbia, Vancouver

    Faculty of Commerce
    Vancouver, British Columbia  Canada 
  • Authors:
    • Hensher, D A
    • JOHNSON, L W
  • Publication Date: 1977

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 00177202
  • Record Type: Publication
  • Source Agency: Transport and Road Research Laboratory (TRRL)
  • Files: ITRD, TRIS
  • Created Date: Jul 19 1978 12:00AM