A Hybrid Efficient Adaptive Discrete Choice Experiment for Investigating Residential Preferences with a Case Study on COVID-19 Pandemic in the Greater Toronto Area

This paper introduces the Efficient Adaptive Stated Preference (EASP) experiment design, an approach suitable for discrete choice situations where respondents may be unfamiliar with a subset of alternatives in choice experiments. The unfamiliarity will violate the rational choice theory's premise that respondents have complete knowledge of the entire choice setting. This study evaluates the EASP approach for residential location choice in the Greater Toronto Area (GTA). The two data collection cycles for this case study occurred in July 2020 and 2021, respectively. While the first cycle used Efficient design, in the second cycle, the EASP design is applied. Data of the EASP based survey compared to the data collected in the first cycle. Application of an artificial neural network demonstrates the performance enhancement of the EASP design. In addition, a statistical comparison of the two cycles of acquired data demonstrates changes in household attitudes toward the factors influencing their residential mobility decisions after a year of pandemic dominance