This paper reports upon the first stage of a research project designed to investigate the effects of traveler characteristics on choice behavior. Data were collected on approximately 2,000 work trips to the cbd of a medium sized city (200,000 pop.). The data included household income, auto ownership, household size and structure, and type of dwelling unit for the respondent's household, together with the age, sex, marital status, occupation, workplace, and driver's license status of the respondent. Other data included levels of service experienced on the survey day in getting to work, and details of an alternative mode of travel (if used within the last year). Conventional disaggregate models were constructed for the total survey population and for various stratifications, based upon the socio-economic data. The primary characteristics investigated in this study were household income, stage in the family life cycle, age, and sex. The stage in the family life cycle was a composite variable derived from data on the respondent's age, marital status, and the number and ages of children living at home. Statistical and estimation comparisons were made between the models, in order to attempt to determine whether socio-economic characteristics affected stated choice behavior, and to determine which characteristics had the most significant effect. The comparisons were made using correlation ratios and f and t statistics. Some significant differences were found between strata for all the socio-economic variables tested. The most significant variable appeared to be sex, while age and income appeared to be the least significant of those tested. The results are somewhat inconclusive since "best" models were not sought for each stratum used, thus resulting in a failure to achieve statistically significant models in a number of cases. For the covering abstract of the symposium please see irrd abstract no. 224453.
Abstract