RESPONDENT TRIP FREQUENCY BIAS IN ON-BOARD SURVEYS

In this paper it is shown that on-board surveys are burdened with an inherent and serious sampling bias. The source and implications of this bias are presented, and a simple statistical correction procedure is developed. An example is used in which the bias leads to a 50 percent overestimate of average tripmaking by users and a 33 percent underestimate of the number of transit users in the population.