SEM have two part that is structural model and measurement model. Composite structural model and measurement model testing enable to test the measurement error as part indivisible of SEM. Based on that condition, SEM can be used and developed to know about a correlation measurement influence in exogen variable and endogen variable, and correlation measurement influence between exogen variable toward structural coefficient. The research use 7 secunder data. Correlation with modification indices information represents the decrease in chi-square and change value of CMINDF and RMSEA to be better whereas at structural coefficient and p-value have not significant change such as at data 1 get structural coefficient and p-value are 0,313 and 0,002, and after correlation error that value be 0,316 and 0,001. The conclusion is that the influence of the correlation measurement error in exogen variable, in endogen variable and between exogen variable not change structural the coefficient significantly. © 2014 Solimun.
