In travel surveys most respondents apply rounding of departure andarrival times to multiples of 5, 15 and 30minutes: in the annual Dutch travel survey about 85-95 percent of allreported times are 'round' ones. We estimaterounding models for departure and arrival times. The model allows oneto compute the probability that a reportedarrival time m (say m=9: 15 am) means that the actual arrival timeequals n (say m=9:21 am). Departure timesappear to be rounded much more frequently than arrival times. Aninterpretation for this result is offered bydistinguishing between scheduled and non-scheduled activities, and byaddressing the role of transitory activities.We argue that explicitly addressing rounding of arrival and departuretimes will have at least three positive effects.1. It leads to a considerably better treatment of variances ofreported travel times. 2. It enables one to avoid biasesin the computation of average transport times based on travelsurveys. 3. It overcomes the problem that the use oftravel survey data for the minute-per-minute records of thedevelopment of the number of persons in trafficdisplays erratic patterns.
travel time transitory activities rounding Bayesian approach scheduling