Abstract
Travel survey respondents often over or under-estimate their travel time depending on a number of factors. This research presents two methods of measuring average speed and thus travel time. The traditional method, stated travel times from travel surveys are compared and corroborated against measured implied travel time based on a GPS based floating vehicle study to identify the measurable operating speed. An example of this approach is shown for a survey of two wheelers (bicyclists and electric bike users) in two cities in China. In these cases, survey respondents report up to 32%higher travel times than what is measured on a floating vehicle study, depending on the case. The implication of this is that mode choice models calibrated on respondent stated travel time more than triples the effect of travel time compared with a model calibrated on average GPS speed (and inferred trip travel time).
