This study is a study that uses the PISA 2003 data to analyze the school effect by applying the multilevel structural equation modeling. The multilevel structural equation modeling reflects the characteristics of mutually different layer in school org...
This study is a study that uses the PISA 2003 data to analyze the school effect by applying the multilevel structural equation modeling. The multilevel structural equation modeling reflects the characteristics of mutually different layer in school organization, finds out the relationship of variables in structures, and through it, this is the model to describe the school effect reasonably that this study uses the PISA 2003 data to confirm the findings.
In general, in the event of data used in the education studies, it may be used together to the analysis with the variables measured in the group level like school and the variable measured in the student level. At this time, if the relational structure between the variables is disclosed by considering for single layer without consideration for measurement in mutually different layer, it would be an error to generate the problems in loss of information and aggregation bias or expand the variables to exaggerate the attributes of the raw data. Therefore, in order to minimize this type of error, it is needed to use the model that has the attributes of each layer reflecting to display the structural relationship.
In this study, through the multilevel structural equation modeling that considered the variables of measured injection/process variables from the student and school levels and the structure of the learning accomplishment, the gross effect from each variable is verified and the route is described and in particular, it confirms how the structural relationship of measured variables from different layer, such as student level and school level.
The variables of student level used in the study are assumed socio-economic background of student, learning related variable, self related belief on learning, and learning strategy, and in school level, it has added with the passion of teachers and educational resource of the school in addition to such variables. The analytical data are the data of Korea in PISA 2003(Programme for International Student Assessment 2003) with the use of data for 5,212 persons excluding the absented data on the applicable variables, and as the analysis programs, SPSS12.0 and Mplus4.1 have been used.
As a result of analyzing the above data, in the event of the data with the multileveled attribute, looking into the structural route following each layer through the multilevel structural equation modeling would be more appropriate. The study model suggested in this study reflects the structure to show the relational structure of the variables, and it may have some difference in the structure route following the layers that reflecting it would be more reasonable. In relation to the concrete route model, the self related belief on learning and learning strategy would show positive effect to the learning accomplishment. The assumed economic background displayed a great indirect effect when the attitude on learning is matched with the learning strategy. In the overall school level, the self related belief showed greatest effect for the learning accomplishment from the group level, and it makes direct and indirect impacts through various routes for the variables measured in the school level with efforts of teachers, school education resource or class related factors, attitude on learning, learning strategy and others.
The result of this study has a difference in the size and route of model that is displayed in the student level and school level as shown that the case of data with nesting structure would reflect it and analyze the structural relationship between the variables that impact on the school effect through the multilevel structural equation modeling.