Between pause and pulse: How travel time shapes opt-out preferences in Hong Kong’s urban street experiments

Street experimental interventions are increasingly used to test alternative street functions, yet their impacts on travel time and their trade-offs with aesthetic and social benefits remain poorly quantified, particularly in dense urban environments. Existing approaches rely heavily on qualitative, trial-and-error experimentation, offering limited empirical guidance for intervention design. This study introduces a quantitative framework combining generative artificial intelligence (GenAI) and stated preference modeling to assess public acceptance of street experiments in Hong Kong. GenAI-produced photorealistic visualizations were used in a survey of 150 participants (1,200 observations). A nested logit model was estimated, extending the multinomial logit (MNL) approach by relaxing the Independence of Irrelevant Alternatives (IIA) property inherited from MNL. This structure captures correlated preferences among intervention types while treating the opt-out alternative as a distinct manifestation of status quo bias. Preferences were examined across road types (alleys, pavements, minor, major), temporality (temporary vs. permanent, time of day), intervention forms (bike lanes, shared spaces, pocket parks, outdoor dining), and travel time impacts. Results indicate strong aversion to travel time disruptions, with even minor delays significantly reducing acceptance. Visually appealing, socially vibrant interventions featuring seating and greenery are preferred, while bike-related infrastructure encounters cultural and spatial resistance. Opt-out behavior is particularly pronounced among older adults, residents of high-density areas, individuals with lower educational attainment, and frequent public transport users, reflecting heightened time sensitivity, space constraints, and reliance on existing transit networks. Findings underscore the importance of explicitly modeling opt-out choices and provide a transferable, data-driven framework for designing incremental and socially accepted street experiments suited to high-density Asian cities, thereby advancing the evidence base for tactical urbanism.