EVALUASI MODEL FIT MULTILEVEL CONFIRMATORY FACTOR ANALYSIS (MCFA) DENGAN MENGGUNAKAN PENDEKATAN PARTIALLY SATURATED (PS)
Thesis ini mengkaji mengenai mengevaluasi model fit Multilevel Confirmatory
Factor Analysis (MCF A) dengan menggunakan pendekatan standar dan
pendekatan Partially Saturated (PS). Tujuan analisis adalah untuk mendeteksi
kurangnya fit pada level between group. Pendekatan standar untuk mengevaluasi
model fit memiliki keterbatasan dalam mendeteksi kurangnya fit pada level
between group. Beberapa pendekatan tingkat spesifik diusulkan untuk mengatasi
keterbatasan pada pendekatan standar, salah satunya adalah pendekatan Partially
Saturated (PS). Pendekatan partially saturated diperkenalkan oleh Box (2002)
kemudian pendekatan ini telah dikembangkan oleh Ryu dan West (2013) dengan
studi simulasi. Uji statistik yang digunakan yaitu uji chi square, Comparative Fit
iv
Index (CFI), dan Root Mean Squared Error Of Approximation (RMSEA).
Penaksiran parameter model dilakukan melalui metode Maksimum Likelihood.
Software yang digunakan dalarn penelitian ini adalah MPlus.
10. Abstract
This paper present a study of the evaluating model fit Multilevel Confirmatory
Factor Analysis (MCF A) using the standard approach and the approach of
Partially Saturated (PS). The goal of the analysis is to detect the lack of fit at the
between group. The standard approach to evaluate model fit has limitations in the
lack of fit at the between group level. Some specific level approach proposed to
overcomes the limitations of the standard approaches, one of which is the
approach Partially Saturated (PS). Partially Saturated approach introduced by
Hox (2002) then this approach has been developed by Ryu and West (2013) with
a simulation study. The statistical test used is chi square test, Comparative Fit
Index (CFl), and Root Mean Squared Error of Approximation (RMSEA). The
parameters of the model are estimated by a Maximum Likelihood. Software used
in this study is Mplus.
v
No copy data
No other version available