Despite the added complexities in model development, data analysis, survey design and administration, longitudinal methods and data constitute a powerful means for transportation analysis. A number of panel studies have accumulated over the past two decades or so (see Raimond and Hensher, 1997). The recent advent of faster and cheaper computational capabilities and methodological advances have been setting ground for the application of complex models to account for estimation problems that have existed in the past. The same advances are making prediction with longitudinal models, or a system of longitudinal models, practical by means of microsimulation. It can be reasonably expected that longitudinal methods can be developed and applied with the same ease as cross-sectional models, to make transportation analysis more coherent, accurate, and richer.
Abstract