Running the estimations separately means you are studying two unrelated models.
Here is a good youTube post on mediation using AMOS: https://www.youtube.com/watch?v=0artfnxyF_A
That should help you in your situation, I think.
Simon
the amos do not need to use a separate estimate, because the resulting output will still show variable unobserve 5, 2 and 1 devendent variable mediators
if we want to separate, I suggest you use PLS
You should never separate the mediators. Otherwise you are testing the mediator in the new model, not in the original model.
How about the path, Initially i included the paths for both mediators and examined the direct/indirect effects for mediation roles of both, Was this the correct approach, I should remove the path for one mediator when testing the other?
No need to remove the path. Otherwise it becomes two separate models.
does not need to be done separately for AMOS will work together, if you are using SPSS path analysis may be performed separately
Hi Prof Zainudin. Thanks for the feedback. The practice of running a single model (and without have to delete path) is applied regardless the type of mediator (parallel versus serial multiple). What I have is serial multiple mediator model, where the two moderators are expected to be correlated.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling