Comparison Between CB-SEM and PLS-SEM: Testing and Confirming the Maqasid Syariah Quality of Life Measurement Model

Mahadzirah Mohamad,Asyraf Afthanorhan,Zainudin Awang,Morliyati Mohammad

DOI: https://doi.org/10.32861/jssr.53.608.614

2019-03-10

The Journal of Social Sciences Research

Abstract:Structural Equation Modelling (SEM) is a powerful multivariate statistical analysis technique which combines both factor analysis and multiple regression analysis. It is capable of analysing the inter-relationships among latent constructs simultaneously in a model. These latent constructs are measured using certain number of items in a questionnaire. Covariance-based SEM (CB-SEM) or full SEM have become the choice for many researchers in a variety of disciplines because of their ability to evaluate complex relationships using parametric statistical approach. Researchers could also opt for Variance-based SEM (VB-SEM) or Partial Least Square-SEM (PLS-SEM) when their data failed the parametric assumptions such as multivariate normality distribution and minimum sample size. However, the approach of VB-SEM or PLS-SEM is a non-parametric instead of a parametric approach in CB-SEM. This article compared the performance of both SEM approaches using the same dataset to validate the Measurement Model for Maqasid Syariah Quality of Life (MSQoL). The findings of both analyses suggested that CB-SEM or full SEM is more appropriate to validate and confirm the MSQoL measurement model.