Tour-Based Path Analysis of Long-Distance Non-Commute Travel Behavior in California

Abstract

Data of long-distance tours by each household from an 8-week California Household Travel Survey travel log are analyzed in this paper. Each tour record contains summary data from a single-day diary, household sociodemographic information, and place of residence characteristics. Each tour contains a main trip, selected tours with a main trip that is not a commute trip, and added destination descriptors from Foursquare. Path analysis is used on three censored variables (tour miles by air, miles driving, and miles by public transportation) and two categorical variables (main trip tour purpose) and number of overnight stays. Annual household income is a major determinant of air travel and multiple overnight stays. Moreover, travel by air is more likely to be in large dense destinations with high attractiveness ratings. In contrast, car ownership is a major determinant of more mileage accrued driving a car to urban environments. In terms of main trip purpose, work and shopping are not associated with overnight stays and vacation/sightseeing is the purpose associated with longer overnight stays. Moreover, California destinations are more likely to be in tours completed within a day. It was also found that vacation, sightseeing, and leisure trips are more likely to be combined with other vacation and sightseeing trips. In contrast, business and combined business and leisure trips are less likely to be combined with vacation trips. There were also substantial and significant differences in long-distance tour behavior among residents of urban versus rural environments.