Application of Structural Equation Modeling – Partial Least Square and Mediation Analysis in Determining Factors of Human Development Index in Indonesia.

Abstract

Structural Equation Modeling (SEM) is a class of multivariate techniques that combine aspects of factor analysis and regression, enabling the researcher to simultaneously examine relationships among measured variables and latent variables as well as between latent variables. Meanwhile, partial least square is soft modeling approach to SEM without any assumptions about data distribution and minimal number of observation that often called as PLS SEM. Data that used in this research are the province’s data in Indonesia that has 34 provinces. The number of observation that equals to 34 observations is too small and could not fulfill the maximum likelihood estimation of SEM. In this research, the estimation method used was PLS, so the analysis called PLS SEM. This research used the PLS SEM to identify the factors that influence the human development index based on the province’s data in Indonesia. Human Development Index consists of three dimensions, which are Economy, Education, and Health that used as latent variables that influence the human development index direct and indirectly. The result of this research show that Economy and Health influence the Human Development Index directly significant. Education influences Human Development Index indirectly with partial mediation that mediated by Economy. The model in this research is evaluated by Q2 value that equals to 67% so the model obtained from this research was good enough In explaining the variance of Human Development Index.