Perbedaan Tujuan SEM Kovarians dan SEM PLS

Sumber Kutipan:
Jorg Henseler, Christian M. Ringle and Rudolf R. Sinkovics (2009),
The Use Of Partial Least Squares Path Modeling In International Marketing,
New Challenges to International Marketing Advances in International Marketing, Volume 20,
277–319

-------------------------------------------------------------------------------------

According to Jo¨reskog (1982, p. 270) ‘‘ML is theory-oriented, and emphasizes the transition from exploratory to confirmatory analysis. PLS is primarily intended for causal-predictive analysis in situations of high complexity but low theoretical information.’’ The philosophical distinction between these approaches is whether to use CBSEM for theory testing and development, or PLS path modeling for predictive applications. ..., in causal modeling situations where prior theory is strong and further testing and development is the goal, CBSEM is the most appropriate statistical methodology. Yet, due to the indeterminacy of factor score estimations, there is a loss of predictive accuracy. This occurrence, of course, is not of concern in theory testing where structural relationships (i.e. parameter estimation) between concepts is of primary concern. Moreover, hypothesis building and the assessment of CBSEM results through global goodness-of-fit
--------------------------------------------------------------------------------------

Menurut Joreskog (1982), Maksimul Likelihood atau taksiran model dalam SEM Kovarians berorientasi teori dan menkankan pada studi konfirmatori atau transisi studi eksplorasi ke konfirmasi. Studi PLS bertujuan pada analisis kausal prediksi dalam  situasi model yang kompleks akan tetapi memiliki informasi teori yang kurang/lemah/pengembangan teori. Filosofi mendasar perbedaan kedua analisis tersebuty adalah bahwa SEM Kovarians berguna untuk pengujian teori dan pengembangan sedangkan PLS untuk model prediksi. Ketika peneliti mempunyai rujukan teori yang mendasari pembentukan model yang kuat dan tujuannya adalah pengujian teori/ model pengembangan maka SEM Kovarians lebih baik digunakan.  Akan tetapi untuk menghindari permasalahan estimasi "indeterminacy of factor score" dan kehilangan akurasi predikasi maka lebih dianjurkan analisis SEM PLS.

Salam,
Sofyan Yamin
@2017, Bogor