Attitude-Based Segmentation of Residential Self-Selection and Travel Behavior Changes Affected by COVID-19

Abstract

This study evaluated the effects of COVID-19 on attitudes toward residential associated with travel behavior on decisions regarding future relocation. Chi-square automatic interaction detection was used to generate tree and classification segments to investigate the various segmentations of travelers and residents around mass transit stations. The decision tree revealed that the most influential variables were the number of transport card ownerships, walking distance to the nearest mass station, number of households, type of resident, property ownership, travel cost, and trip frequency. During the COVID-19 pandemic, people have concentrated on reducing travel time, reducing the number of transfers, and decreasing unnecessary trips. Consequently, people who live near mass transit stations less than 400 and 400–1000 m away prefer to live in residential and rural areas in the future. Structural Equation Modeling was used to confirm the relationship between attitudes in normal and pandemic situations. According to the findings, attitudes toward residential accessibility of travel modes were a significant determinant of attitudes toward residential location areas. This research demonstrates travelers’ and residents’ uncertain decision-making regarding relocation, allowing policymakers and transport authorities to better understand their behavior to improve transportation services.

Keywords

• COVID-19 affected
• residential self-selection
• mass transit
• decision tree
• structural equation modeling
• attitude-based