Calculation of Composit Reliability and Maximal Reliability on confirmatory factor analysis.

Abstract

In the social research, the research variable is usually cannot be measured directly (usually called as latent variable). Therefore, the variable needs an indicator as its measuring instrument. One of multivriat techniques is structural equation modeling (SEM) that is used in social researchs. Structural equation modeling (SEM) consists of 2 stages, the first approach is using confirmatory factor analysis (CFA), the second is by doing the structural equation modeling (SEM). Confirmatory factor analysis (CFA) approach is aimed to confirm that indicators used as the measuring instrument  are good enough to measure direct unmeasurable variables ( latent variable) seen from reliability score of those indicators. According to Hair, et al (1998) that reliability score is good when it has bigger score than 0,7. The results show that both composit and maximal reliablities have bigger scores than 0,7. It means that the scores of composit and maximal reliabilities can be accepted, it shows that all used indicators in this research is reliable.