Abstract:In this study, we investigate the heterogeneous effects of the COVID-19 pandemic on the travel behaviors of 6,012 individuals in the United States. With a latent-class cluster analysis, we identify five distinctive groups based on changes in travel behaviors between two timepoints: fall 2019 (before the pandemic) vs. fall 2020 (during the pandemic). Each group presents a unique set of changes (or non-changes) in physical commutes, (perceived) mode availability, and trip frequencies by ten modes for commutes and leisure trips. We also examine the socioeconomic, demographic, and attitudinal profiles of each class, which help understand the reasons behind these changes. The first three classes are those who commuted during the pandemic. Those in Class 1 depend highly on private vehicles with the least availability of other travel modes, but some of them even increased commute trips due to job needs during the pandemic. Class 2 are also auto-oriented, but they do so by choice given that other travel modes are available to them. In contrast, members of Class 3 are the most multimodal with the highest trip frequencies overall, with modest changes between two timepoints. Class 4 mainly includes teleworkers who reduce physical commutes the most. Class 5 contains empty nesters who commuted neither before nor during the pandemic with the least travel demand. In sum, this study offers unique insights on heterogeneous behavioral changes in response to the pandemic, which help identify effective policy options to mitigate the negative impacts of the pandemic while promoting sustainable and safe transportation.