An Empirical Study of Travel Behavior

Abstract

The empirical strategy described here involves matching travel diary data to land-use characteristics for two different southern California regions. The first data set is from Orange County and nearby parts of suburban Los Angeles, and the other data are from San Diego. Each data set poses somewhat different opportunities and challenges in measuring the factors of interest, so we describe these in turn. The Orange County/Los Angeles travel diary data set includes data for 769 southern California residents. These were obtained from a 1993 survey administered as part of the Panel Study of Southern California Commuters. Because that survey includes each respondent’s street address, we were able to match the travel diary data to land-use variables from the 1990 census and from the Southern California Association of Governments (for the years 1990 and 1994). The travel diary covered a two-day period, and respondents were pre-assigned days, so trip making on all days of the week is represented in the data. Individual respondents were first contacted through their employer, and then for follow-up waves the same persons were contacted at home. The sample is employer based, and consequently the respondents are not a random sample of southern California residents. About half of the respondents worked at the Irvine Business Complex, a large, diversified employment center near the Orange County Airport, and the other half worked elsewhere throughout the Los Angeles metropolitan area. The descriptive statistics shown in table 5.1 illustrate how the survey oversampled Whites, highly educated persons, and persons with high income. This suggests some caution is warranted in interpreting beyond these individuals. Yet restricting attention to an educated, upper-middle-income, largely suburban population still provides interesting information, because many of the new urban designs are intended for low-density, suburban environments with demographics similar to those in this sample. The dependent variable for the model is the number of nonwork automobile trips made by an individual during the two-day travel diary period.