Strict and Deep Comparison of Revealed Transit Trip Structure between Computer-Assisted Telephone Interview Household Travel Survey and Smart Cards

Abstract

Large sample household travel surveys (HTSs) are an essential tool for the planning of urban transit systems. The progressive adoption by transit agencies of fare collection systems based on smart cards (SCs) has, for the first time, provided opportunities to compare the survey data with detailed, population-level data collected independently. These comparisons have produced some surprising results. Although the underreporting of non-home-based and off-peak trips was to be expected, the significant overestimation of transit use during peak periods was not anticipated. Using the Greater Montreal Area as a case study, this paper performs a strict and deep comparison of computer-assisted telephone interview (CATI) HTS data and SC data across several dimensions: transit agency usage, departure time from home, number of trips per traveler, and activity durations. The analysis reveals that the HTS constitutes a simplified portrayal of transit usage patterns. Non-home-based trips and trips made for activities of short duration are underrepresented in the survey data, leading to an underestimation of off-peak travel by transit. In addition, the systematic overestimation of peak period transit use appears to be because of the corrective weighting of the 20–29 demographic which is notoriously difficult to reach in a telephone-based household survey.