I run a nine -factor CFA using the Diagonal Weighted Least Square (DWLS) estimator for a dataset with 1234 respondents from a developing context and this was my model fit indices:
X2 (866)=17177.58
p
Hello Milli,
There are several reasons. First, make sure all the assumptions are not violated. Check if there are missing values (eg. you can do imputation of missing values), multivariate normality, multicollinearity). Next, check the modification indices to identify which observed variables or errors you need to covary. I think this could improve the model fit. Thanks
You should run diagnostic tests too in order to find out that your data set is relaible or not. If any diagnostic test gives you an indication that the employed data set does not possess the characteristic which is important to possess before analyze then you should take this issue in considertaion. It will deviate the RMSEA results.
Hi Millicent
maybe you should consider running the analysis as a multi-group model. You may then be able to find out which part of the CFA model is different across the different groups.
Mark
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis