A comparative study between GSCA-SEM and PLS-SEM

Zainudin, Awang and Mustafa, Mamat and Asyraf, Afthanorhan (2016) A comparative study between GSCA-SEM and PLS-SEM. MJ Journal on Statistics and Probability, 1 (1). pp. 63-72. ISSN 927-7032

Text FH02-FESP-17-09394.pdf Restricted to Registered users only Download (706kB)

Abstract

Structural Equation Modeling (SEM) now becomes one of the prominent methods that can ascertain researchers today to analyze the causal effect between latent constructs with multiple variables simultaneously. To add, SEM have two catego-ries namely Covariance and Component or Composite based SEM in which each approaches have their own advantages and disadvantages. Component model have two approaches namely Generalized Structured Component Analysis (GSCA) and Partial Least Square in which both of these applications itself using least square estimator and bootstrap technique in providing of parameter estimate and hypothesis testing. However, there are only a few studies using GSCA-SEM to ana-lyze the relationship between latent construct. Therefore, this study interest to compare the performance of PLS and GSCA in terms of the consistency factor loading, parameter estimation, standard error, and statistical power test using different of sample size. In this study, we perform the medical tourism data as a research subject. The study reveals that the perfor-mance GSCA-SEM is better than PLS-SEM in terms of the consistency, standard error and parameter estimate. However, both of this application is really powerful to minimize the type II error.

Actions (login required)

View Item