Confirmatory Factor Analysis with its expectations of multi-variate normality and linear relationship amongst explanatory variables does not meet the criteria.
I tried Confirmatory analysis.
P-Value associated with Chi Square was 0.000, and SRMR is 0.106, RMSEA is 0.160 and CFI is 0.497.
All of this indicates a poor fit.
So, I went back to Exploratory Factor Analysis, and found that # factors 4 and 5 do not fit the data appropriately (p value = 0.00) while # factors=6 leads to 0 degrees of freedom and p-value close to 0.
I believe this is stemming from my non-normally distributed observed variables.
Hello,
Bear in my mind that CFA should be applied for already existing and documented in the past factor structures.
EFA on the other hand pertains to newly developed measures.
What scenario is in your case?
Regards, Witold
I am not sure what kind of Exploratory Factor Analysis you used that gave you probability values for your factors, and I don't understand what you mean that there were 0 degrees of freedom for one of the factors.
Based on those results and the very low fit values for your model, it seems like you are more likely to have problems with the specification of the model, rather than violations of the normality assumptions.
- Badges
- Science topic