Analysis of the First Two Waves of the Puget Sound Transportation Panel: PanelAttrition, Residential Location, Mode Switching, and Panel-Based Forecasting. Task D: Latent Demand Estimation Using Panel Observations

摘要

The need for an estimate of latent travel demand (i.e., the difference between areference benchmark--minimum or maximum--demand for transportation services and the actual use of transportation facilities) has clear potential for planning applications. Methods designed to quantify latent demand are needed with increasing urgency because of recent legislative trends and consequent research activities. Also, recent attempts to provide an adequate measure of latent demand are increasingly making use of more sophisticated information such as data from panel surveys. In this paper, a disaggregate measure of latent demand is defined and statistical models of the measure are provided. This measure permits estimation of potential travel behavior changes at the most elementary unit of analysis (household or person), and it is defined at the same unit as a random variable, allowing for stochastic variations from one unit of analysis to another. This new method is based on the recognition that boundaries in the movement of people exist and can be estimated using regression methods. Models at the household level are estimated using data from the Puget Sound Transportation Panel, and illustrations on the use of the newly defined measures are offered in this paper.