Who drops off web-based travel surveys? Investigating the implications of respondents dropping out of travel diaries during online travel surveys

Abstract Household travel surveys collect core datasets for modelling passenger travel demand. However, decline in survey completion rate is becoming a concern in recent years. One major cause is the transition from computer-assisted telephone interviews (CATI) to computer-assisted web interviews (CAWI) surveys, where respondents need to self-complete the surveys without any active help from an interviewer. Among all components, the travel diaries are the most challenging part of CAWI travel surveys and suffer significant dropouts of participation. Therefore, an investigation is necessary to understand the implications of such participation dropouts in CAWI-based household travel surveys on travel survey data quality. This study reports two travel diary designs developed to ease response burdens in travel surveys. Empirical investigations are conducted to understand survey participation drop-off behaviour while filling out travel diaries. In proxy household travel surveys, diary designs with stable repetitions outperform sophisticated diary designs. In a survey deployed with supposedly advanced diary designs, respondents with higher travel demand are more likely to drop out. Estimated using a generalizable analysis framework proposed by this study, the travel demand reflected in the final dataset collected using the sophisticated travel diary design might be underestimated by 10.2%. This study also proposes strategies and recommendations for future travel surveys.