Path Analysis- SAS PROC CALIS: Fitted correlation warning message

ESM

New Member
#1
Hello All,
I need some help running a path analysis in proc calis. My ultimate goal is to assess which explanatory variables have direct and indirect effects on my response variable to gain a better understanding of which variables are more related.

After running the code, I received two error messages. The first message states that the fitted correlation for my response variable deviates from 1 by at least 0.01. The second message states that fitting a correlation matrix should lead to insignificant diagonal residuals and the maximum value of the diagonal residuals might be too high for a valid chi-square test statistic and standard errors. This makes sense as the chi-square estimates in the output seem to be "off". The log was also telling me that all effects in the model were direct effects so there were no displays for effects or standardized effects.

I realized that I had fixed all of the parameters in the model in the STD and COV statements. I fixed this problem by setting only my direct effects as fixed parameters in the STD and COV statements. However I'm still receiving the same error messages.

Long story short, I don't quite understand why I'm getting the same warning messages and I'm not real sure as to why all of my effects are considered direct effects after making the parameters of my direct effects fixed.

Any information that anybody might have would be greatly appreciated!!

Thanks for your time!
ESM