GEOGRAPHICAL RESEARCH ›› 2010, Vol. 29 ›› Issue (2): 338-350.doi: 10.11821/yj2010020016

• Urban and Rural Studies • Previous Articles     Next Articles

A household-based model of shopping decision-making behavior in timing and destination choice: Formulation and application to spatial policy evaluation

ZHANG Wen-jia, CHAI Yan-wei   

  1. Department of Urban and Economic Geography, Peking University, Beijing 100871, China
  • Received:2009-04-20 Revised:2009-08-17 Online:2010-02-20 Published:2010-02-20


The researches of shopping behavior in urban geography and planning mainly focus on how to manage urban residential non-work travel, how to regulate and optimize commercial spatial structure, and how to improve residential everyday life quality. In this paper, we put forward the conceptual model of the shopping decision-making behavior on the perspective of intra-household interaction. Hereby, using the first-hand data from a activity-travel daily survey of Beijing in 2007, we derive and formulate a multinomial logit model (MNL) for shopping timing decision-making, and a conditional logit model (CLM) for shopping destination choice, to interpret the behavioral diversity among different groups and evaluate the spatial policies. The results indicate that: First, the influences of the factors about intra-household interaction and different decision-making levels on the shopping space-time decision-making process are more significant than that of traditional socio-demographic characters. Therefore, the characters of household-based decision should be paid more attention by the studies of activity-travel behavior. Then, the groups travelling in the morning peak period seem to be low-income or exurb-living, with non-motor mode or longer shopping duration. Also, the dwellers travelling in the evening peak period seem to be with family companions, suburb-living or choosing the tour mode of HWSH. Moreover, it does mean to attract more shopping activities with higher commercial density. At the weekend, residential shopping activities are mostly concentrated within 6 km far from the home, or appear at the high-density CBD. Furthermore, the impacts of personal characters on the destination choice give more attention to the home-to-shop distance, rather than the density of the shopping district. Finally, the results from estimation and simulation evaluation show that the policy of mixed land-use might indeed reduce the travel time, but increase the probability of multi-purpose shopping tour, which will more probably choose the peak-time travelling, then aggravate the traffic jam. Besides, suppose that the commercial establishment would be uniformly distributed, the more shopping activities will be concentrated surrounding the home. Although the cities have commercial hierarchical structure, it would be better to use the policy of decentralization, rather than aggregation, to improve the urban transportation and advance the residential everyday life quality.

Key words: shopping behavior, timing and destination decision-making, household-based, Conditional Logit Model (CLM), policy evaluation and forecast, Beijing City