I have just run a SEM model and the CFI and other goodness-of-fit measures indicate that the model is a good fit. However, under on R-Square, the estimate of my 3 variables are listed as undefined and the residual variance these variables are high.
Is R-square estimate really important and should be dealt with or does it not matter as long as all other measures such as CFI, RMSEA, TLI are fine? If not then, whats the best way to deal with it?
This can happen when you have a reciprocal interaction in the model, with count or categorical variables, and if the model is misspecified. If you want us to look at it, send the output and your license number to email@example.com.