|
|
Almost perfect fit indexes? |
|
Message/Author |
|
|
Hi Dr. Muthen, I got close to perfect fit indexes in a path analysis model. I have a small sample size (52). I am wondering if fit can be calculated. I saw that you've told on other posts that when there are perfect fit is usually that Degrees of Freedom are 0 and you have a just-identified model. But this is not my case as I have 2 DF?. In brief, are these fit results right or I cannot asses fit and so I should not report them?. Many thanks in advance!. Chi-Square Test of Model Fit Value 0.072 Degrees of Freedom 2 P-Value 0.9646 RMSEA 0.000 CI 0.000 0.000 Probability RMSEA <= .05 0.969 CFI/TLI 1.00 /1.323 SRMR 0.005 |
|
|
You can report the fit but make clear that the fit statistics are not trustworthy at such a small N. At the very least, the power to reject the model is extremely low. |
|
Back to top |
|
|