고차 구조방정식모형에서 잠재변수의 측정방법에 관한 연구

Abstract

The estimation method of latent variables in high order factor model has strict limitations. In the PLS method, the accurate scores can be measured only when the number of manifest variables related to each latent variables are the same. In this study, the analysis was studied by comparing the numbers of manifest variable related to the latent variables when they are the same or not. Comparing the numeric analysis with the beta-value, through examples measuring the beta-value by increasing the number of manifest variables, and yielding the fit index by comparing the numbers of manifest variable related to the latent variables when they are the same or not. Therefore, the repeated use of manifest variables should not be used except for when the number of manifest variable related to the latent variables are the same.