Abstract
This paper quantitatively analyzes parents' choice behavior in view of the possible introduction of a public school choice system in Japan. There exist a number of studies and discussions on the merits and demerits of the school choice system on a theoretical basis. However, since they are not supported by empirical data of actual school choice behavior, it is still unknown whether the school choice system is desirable, and if so, in what form the system should be introduced. This study aims to fill the lack of information on parental school preference, and to simulate the choice behavior. In this study we performed a questionnaire survey and obtained hypothetical choice data from 908 schoolchildren's parents in the Tokyo metropolitan area (primary school choice: 564 parents of first or second grade schoolchildren belonging to five primary schools; secondary school choice: 344 parents of fifth or sixth grade schoolchildren belonging to three primary schools) from February through July 1998. We conducted this survey in the Tokyo metropolitan area primarily because this area meets the requirements for school choice system, such as the density of population, the density of schools, and traffic conditions. The data obtained in this survey were analyzed using the rank-logit model, a disaggregate choice model that is suitable for analyzing individual choice behavior. The results show that 1) subjects were clearly classified into two groups in terms of choice behavior in both primary and secondary school choices, 2) these groups were different with respect to the rank-logit model representing school choice behavior, 3) there were no significant differences as to parent's educational background and occupation between the groups. On the other hand, we can find significant differences in individual taste or state such factors as the child's personality and dissatisfaction with real schools near their home. 4) It is possible that the school choice system, especially the secondary school choice system, causes over-concentration of children on a specific school or school ranking. These results enable us to simulate parent's school choice behavior under various conditions, and to examine what school characteristics and school disposition level will cause over-concentration on a specific school or school ranking. From another viewpoint, it is also necessary to analyze the survey data in areas besides Tokyo metropolitan area, and furthermore, to investigate why parents place a great deal of weight on proximity of school, especially on primary school choices.
