Pedestrians’ Decision of Shopping Duration with the Influence of Walking Direction Choice

Abstract

To predict the duration of time that consumers spend in shopping environments, researchers often use hazard models to predict the probabilities of ending the shopping trip as a function of individual sociodemographics and static environmental factors. Models are established in such a way as if the duration has been determined prior to the shopping trip since the environmental factors are constant. The influence of environmental variations attributable to an individual’s shopping experience during the shopping process on the decision making has not been studied. This paper proposes a model that incorporates individuals’ dynamic shopping experiences into the decision of ending a shopping trip. A nested logit model is developed to explain pedestrians’ decisions of going home during a shopping trip, combined with the decision of choosing a walking direction. The model is estimated on two data sets of pedestrian shopping behavior in two Chinese cities. The results show that incorporating dynamic experiences can significantly improve the statistical accuracy of predicting the go-home decisions. However, the practical improvement is limited. Temporal factors turn out to be good predictors of the go-home probability; relatively robust spatial factors for the direction choice are identified by comparing the two cases.