具相關誤差二階CFA組合信度之估計法

Abstract

本研究針對具相關誤差之二階驗證型因素分析模式提出一個有效的組合信度估計方 法，首先以逐步篩選法，利用卡方差異檢定，認定模式中具相關之誤差項，在模式之參 數估得後即可據以估計組合信度。本研究將採蒙地卡羅模擬檢視組合信度之統計性質， 並比較二階CFA組合信度與第一階各構念組合信度大小。所提組合信度估計法在二階驗 證性因素分析實證研究上具應用價值。

In this study, an effective approach is proposed to estimate the composite reliability in the model of second-order CFA with correlated errors. Correlated errors are first identified by using stepwise selection based on the chi-square difference test. An estimate of composite reliability can be obtained after model parameters are estimated. Monte-Carlo simulation will be used to evaluate the performance of the estimated composite reliability and compare between the second-order CFA composite reliability and the composite reliabilities for the first-order constructs. The proposed approach is useful in second-order CFA empirical studies.