The optimal duration for a travel survey

Abstract

Empirical observations were carried out using a six-week travel survey, Mobidrive, in order to probe into the optimal duration problem of a single wave travel survey. The regressions on certain travel elements (i.e., number of daily trips, VkM per day, number of home based trip chains, daily travel time expenditure) suggest that a two-week survey duration might serve well to represent personal variability in various dimensions. However further analysis carried out by using nonparametric survival analysis points to a longer period as the elapsed time between activity participation does not seem to converge to true values, at least for rare activities such as recreation.