CONTACT US: Contact info
|
Citation |
Purvis CL. Transp. Res. Rec. 1994; 1443: 21. |
|
Copyright |
(Copyright © 1994, Transportation Research Board, National Research Council, National Academy of Sciences USA, Publisher SAGE Publishing) |
|
DOI |
unavailable |
PMID |
unavailable |
|
Abstract |
Disaggregate (household-level) automobile ownership choice models are typically estimated by using large-scale cross-sectional household travel surveys. Automobile ownership choice models typically stratify households into households owning zero, one, or two or more vehicles. This automobile ownership market segmentation is critical in the application of a regional set of disaggregate travel demand models for aggregate forecasting purposes. An alternative regional data set for estimating disaggregate automobile ownership choice models is the 1990 Census Public Use Microdata Sample (PUMS). PUMS consists of two disaggregate files of individual 1990 census records (household and population characteristics) of either 1% of an area's households or 5% of an area's households (the 1% and the 5% samples). Disaggregate workers in household and automobile ownership choice (logit) models were estimated on the basis of PUMS data files for the nine-county San Francisco Bay Area and the one-county San Diego region. These models were also compared with disaggregate models on the basis of the 1990 Metropolitan Transportation Commission household travel survey. The strengths and weaknesses of both approaches--PUMS versus household travel surveys--are discussed. The primary weakness of PUMS is the lack of data on neighborhood characteristics, such as land use density or accessibility measures, at a fine enough geographic level (i.e., regional travel analysis zone) for model estimation purposes. The transferability of the model estimation methodology to other metropolitan regions is discussed. Language: en |