Generalized Structured Component Analysis to Analyze Measurement Models: Utilization of Health Insurance

Abstract

This study aims to analyze measurement models of the Utilization of Health Insurance. A survey is conducted annually by Statistics Indonesia in measuring the Utilization of Health Insurance through National Socio-Economic Survey (SUSENAS). Code 81301-81309 on SUSENAS Core is blocking code XIII with a description of Health Insurance Utilization. Y1: Do you have health insurance? Y2: have you ever used health insurance? Y3: What are the reasons you have never used health insurance? Y4: in the last year, have you ever been refused a health check? Y5: what are the reasons you experience refusal to check your health?, Y6: in the last year, have you ever used health insurance for hospitalization?, Y7: What are the reasons you never used health insurance for hospitalization?, Y8: in the last year, have you ever been refused hospitalization?, and Y9: what are the reasons you experience refusal to be hospitalized? Generalized Structured Component Analysis (GSCA) is part of SEM-based on components that have global least-square optimization criteria, which can be consistently minimizing sum squares residuals to obtain model parameter estimates. GSCA is a powerful analysis method. This is because it is not based on many assumptions such as variables that do not have to have multivariate normal distributions (indicators with category, ordinal, interval to ratio can be used on the same model), the amount of data does not have to be large. The results showed that some indicator variables had no significant effect on the construct variables. The loading factor > 0.6 are latent construct indicators that provide good convergent validity.