PREDICTIVE ACCURACY OF AGGREGATE AND DISAGGREGATE AUTO OWNERSHIP MODELS

A test is made of the predictive validity of aggregate and disaggregate auto ownership models. Three models--linear disaggregate, nonlinear disaggregate, and linear aggregate--are calibrated using 1964 home-interview data from Binghamton, N.Y.; forecasts of auto ownership are then made to 1970 and compared with Census. Results indicate that the nonlinear logit form is superior to the linear disaggregate model, and equivalent to the conventional aggregate linear model, despite a 75 percent reduction in observations used in model-building. /Author/

Media Info

  • Features: Figures; References; Tables;
  • Pagination: 31 p.

Subject/Index Terms

Filing Info

  • Accession Number: 00131076
  • Record Type: Publication
  • Report/Paper Numbers: PRR 95
  • Files: TRIS, STATEDOT
  • Created Date: Apr 21 1976 12:00AM