Household travel surveys collect core datasets for modelling passenger travel demand. However, the decline in survey completion rate has become a concern in recent years. 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 influential factors contributing to participation dropouts while reporting their travel dairies and subsequent non-response bias on data quality. This study utilizes binary logit models to capture the relationship between respondents’ drop-off behaviours while filling out travel diaries, their socioeconomic characteristics and different travel diary designs. The study also develops a generalizable analysis framework to measure the impact of non-response bias on data quality. The analysis framework incorporates a trip generation model with microsimulation and bootstrapping techniques. The results show that a diary design with stable repetitions is preferred by respondents and results in less non-response bias in the final dataset. This study also proposes recommendations for future travel surveys.

家庭旅行调查收集用于模拟旅客旅行需求的核心数据集。然而，近年来调查完成率的下降已成为一个令人担忧的问题。在所有组成部分中，旅行日记是 CAWI 旅行调查中最具挑战性的部分，参与率大幅下降。因此，有必要进行调查，以了解导致参与退出的影响因素，同时报告他们的旅行日记以及随后对数据质量的不答复偏差。本研究利用二元 Logit 模型来捕捉受访者在填写旅行日记时的下车行为、他们的社会经济特征和不同旅行日记设计之间的关系。该研究还开发了一个通用的分析框架来衡量无响应偏差对数据质量的影响。该分析框架将行程生成模型与微观模拟和引导技术相结合。结果表明，受访者更喜欢具有稳定重复的日记设计，并且最终数据集中的无回应偏差较小。这项研究还为未来的旅行调查提出了建议。