A Modeling of KFC''''s Spatial Diffusion in Beijing City
KFC 在北京城区的空间扩散模型

Cultural diffusion is an important part of study in cultural geography and communication. The current models of cultural diffusion take time as a variable and some of them take location as a variable. Cultural geography pays attention to the spatial diffusion of cultural phenomena. Therefore, diffusion models describing spatial changes are required. Setting Kentucky Fried Chicken (KFC) as a research object, this paper builds models to analyze how foreign catering chain stores are spatially diffused within Beijing City. KFC's diffusion is a typical expending diffusion. This kind of diffusion has spatial preference. The social structure of Beijing made a good base for KFC's entering in the late 1980s. This modeling analysis is based on the data given by KFC's headquarter in Beijing and aims to set models with spatial variables. This study found KFC's diffusion does not follow the grade of commercial central places in Beijing City. Five diffusion rings in this paper are designed as spatial variables. The Ring Order Model and Radius Model can not pass their tests, while Alpha Model passes. Alpha Model depicts the relation between diffusion directions of KFC and their opening order. The three conclusions of this paper are: (1) In comparison to former models, this paper designs three new spatial variables for diffusion models of KFC. Alpha Model has passed its reliable test. It could be applied in forecasting the direction of a new KFC approximately. (2) These models, in addition to yielding new insights into the processes involved in spatial diffusion, are helpful for middle class catering chain stores in site selection. The Alpha Model is better for selecting than Ring-Order Model and Radius Model. (3) Economic factors have greater influence on KFC's diffusion. The Ring Order in this paper shows the consumption levels. One of the shortcomings in this paper is taking diffusion rings as static ones. Another one is ignoring the competition of other catering chain stores in the same commercial area. Lacking of census data, the models in this paper have less variables.