I have 4 factors. Every factor is predicted by around 10 items. I did not face a problem in running CFA. However, when I want to make path analysis in Amos, how can I deal with all these items? shall I take the mean of all items per one factor? any advice ?
@Osama Using the latent variables from the CFA model, you can run SEM model for testing your hypothesized relationships. However, you can also compute latent variables using mean of observed variables for each factor and then run path analysis but the results of SEM are more robust than path analysis.
Imran Anwar Thank you so much for the answer. This is what I was looking for. Because I have taken a course where SEM=CFA + path analysis. but your answer proves an important point for me because I watch on youtube that you can rearrange the arrows of CFA to look like path analysis and then to add disturbances to dependent variables and this is how to build SEM (for sure, based on the theory of the study).
Dear Osama,
Following Mr. Imran Anwar, Path analysis can only be done on the observed variables. If necessary, assuming the existence of latent variables, we can average the observed variables through the compute command and express them in the form of an observed variable.
Sincerely,
Zahra
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running