Hi everyone, it's my first post here and am new to SEM. :) Anyway, I would like some help and recommendations on how to improve the CFA model I'm working on. The CFA is based on n=310 with a bootstrap of 5000; and modification indices of 4. The results I have are:
Result (Default Model)
- Chi-square = 1090.271
- Degrees of freedom = 511
- Probability level =0.000
Model Fit Summary
- PCMIN/df = 2.134
- CFI = 0.904
- TLI rho2 = 0.895
- RMSEA = 0.061
Thank you!
Without knowing more details about your model and data, it is virtually impossible to give specific tips for model improvement. As a general tip, it can be useful to study standardized residual covariances to examine which variances or covariances may not be well-reproduced by the model (e.g., because there are residual associations between indicators).
Hey Dan, I agree with Christian that it's hard to offer advice with just the fit statistics to go on. But I can tell that this is AMOS output, and I also know that AMOS doesn't handle categorical data very well. There are some great methods to handle data that have a small number of categories (e.g. Likert type response formats with 2-4 options) but AMOS doesn't have the capability - so you will have to be treating all your variables as continuous, and this will likely lead to poorer fit.
Ignore this message of you are dealing with data with 5 or more categories!
Mark
Model Fit:
CMIN/DF = ≤ 5
P = 0.05
GFI/NFI = >0.90
CFI/TLI = >0.95
SRMR = ≤ 0.10
RMSEA =
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Climate Science
- Climate Change