Dear all, I am conducting SEM with AMOS and I understood first I have to do CFAs to evaluate the model fit. I want to control for the country of origin in my path model (we have two countries). Do I have to include this control variable already in CFA or do I run CFA without it and then include the control in path analysis? Thanks for your help! Ilia
I would recommend running a multigroup CFA (with country as grouping variable). That way, you can check whether the same factor model applies in both countries and test for measurement equivalence across countries. You could do the same in your path analysis or SEM (run it as a multigroup analysis).
Dear
I think you should use both ways to check your model fit. Infact, it depends on the nature of hypothesis,you want to test.
Thank you, all! I appreciate you taking time to respond! Regards, Ilia
Dear Ilia,
The short answer:
You should run the CFA without the control variable, then include it in the path analysis after testing for measurement invariance.
The long answer:
Remember to assess the goodness of fit of the final structural model and interpret the path coefficients to understand the relationships among the latent variables and the control variable.
Best,
Wadie
Dear Wadie Abu Dahoud, are there perhaps any references you would suggest that are relevant to this methodology that you described? I am completing a similar analysis and found your instructions helpful, and am interested to find any resources to aid my understanding. Kind regards, Jasmine
Dear Jasmine Wu ,
You may find this helpful:
Brown, T. A. (2015). Confirmatory factor analysis for applied research
Wadie Abu Dahoud thank you, I will have a look at this, much appreciated!
- Badges
- Science topic
- Similar topics
- Statistics
- Survey