The COVID-19 pandemic triggered dynamic shifts in urban residential location preferences, yet individual-level evidence on how these preferences have evolved across different stages remains limited. Using a hypothetical, scenario-based discrete choice experiment with 7318 urban residents in China, this study estimates within-person changes in residential trade-offs across three decision frames: a pre-pandemic baseline (September–November 2019), the post-lockdown present, and an envisioned post-pandemic future without life-threatening risk. The results reveal four temporal trajectories (i.e. rising, falling, V-shaped, and inverted-V patterns), indicating that the pandemic altered residential location preferences in non-uniform ways. Sensitivity to monthly rent and proximity to local shopping rose sharply immediately after lockdowns but subsided thereafter, reflecting short-term economic and provisioning concerns. By contrast, preferences for shorter commutes, better public transport accessibility and higher school quality strengthened steadily across all three decision frames, signalling more durable reordering of daily-life priorities. Shifts in residential location preference were heterogeneous. Middle-income respondents and those with asymptomatic or mild infections adjusted location preferences across the broadest set of attributes, whereas the largest magnitudes of change occurred among households with older adults or school-age children and among highly satisfied residents. Family-centred households became increasingly sensitive to commuting burdens and facility distances, while higher-income residents, licenced drivers, service workers and satisfied residents became progressively less sensitive to distance and cost. These findings reveal emerging divergence between local-convenience and spatial-flexibility logics. Policy responses should therefore reinforce affordability, reliable commuting and educational access citywide while tailoring neighbourhood improvement and service provision to differing household needs.
Meng et al. (Fri,) studied this question.