Dear researchers, I want to update my knowledge regarding data analysis. I will be honored if you can help me to reduce in correcting SRMR in SEM_PLS. All the measures (AVE, CR, VIF, CA, HTMT) are good except SRMR = 0.15
Hello Jan,
The standardized root mean square residual is one way to quantify how well a model can reproduce the observed relationships among the measured variables in your data set. A large value (for many people, > 0.05) would suggest inadequate model-data fit.
How to fix this?
1. Get a better model. You may require different/additional paths among variables in order to capture what's going on with the data.
2. Get better measured variables. The technical quality of your scores may not be as high as you could otherwise find (e.g., poor reliability, questionable validity).
3. Gather additional data. It's possible that, if your sample size is modest, that it may not represent the target population well (all other things equal).
4. Modify any constraints you've imposed. If you've fixed multiple paths to have identical values, this may be responsible for the apparent poor performance of your model.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Correction