I have multiple mediators model. I need a thorough guidance on whether I should use bootstrapping or Phantom model approach in AMOS to test the mediation of five mediators and test their differential effects. Please send relevant research papers in which the procedure of reporting the results with both techniques is outlined. Thanks
Which statistical technique to employ depends on the objective of your analysis. If you intend to analyze latent constructs, then SEM is the best since it is developed to cater for limitations in OLS Regression. If you want to analyze mediation effects in the model, then SEM is the most appropriate for the task.
Thank you so much for your guidance Sir, but there is one point on which I want to be more clear. SEM provides total effect, indirect effect in mediation through bootstrapping, but if the reviewers inquire about the specific indirect effects, then it is not available in AMOS. What should I do, Should I use both bootstrapping as well as Phantom model appRoach in AMOS
Razia, what you mean "specific indirect effects"?
Don't mixed up between SEM and AMOS. SEM is the statistical technique while AMOS is a software/tool to achieve it. If you employ AMOS to analyze your structural model, it will give you the usual SEM output for testing and you can also get the Boostrapped Output as well if you choose the Bootstrap option.
Thanks Sir Zain for your valuable input, Yeah u r right, can you please suggest how to calculate the specific indirect effects in AMOS.
Dear Razia,
Usually, if you really interested to assess the specific indirect effect, mean that, you have more than one mediation variable. If so, then, you can try learn the phantom variable. to do, you have to use Amos syntax
Thanks Sir Wan Mohamad, Yeah I have more than one mediator in my model, that's why I'm interested in specific indirect effect, can you please guide me any source/paper through which I may get more information
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques