Findings in pedestrian route choice experiments: Tradeoff and hypothetical bias

Abstract

In this paper, we study the pedestrian route choice behaviors through a controlled experiment. Two routes with different lengths and waiting times are considered. Sometimes, extra rewards are given to test the different responses in diverse situations. The surveys before and after the experiments (S1 and S2) can aid in collecting data on pedestrians’ individual traits. In the binary logit model, the main factors for pedestrian route choice are identified and discussed, including the route, experimental and pedestrian attributes from the surveys and experiments. Next, we employ three methods to explore the hypothetical bias of pedestrian route choice behaviors: (1) Comparisons of logit model coefficients; (2) Predictions of revealed results by stated results; (3) Proposal of a new indicator (named PB) and related linear regressions. The results indicate that the hypothetical bias can be quantitatively measured and reduced, which could be beneficial for future studies.