Dear Researchgate users,
Here it goes a dilemma.
I have a mediation model - with 3 mediators.
Only one of them turned out to be significant. Now I would like to apply moderators - and here is the big question:
Should I just test using the significant effect found in mediation or all three? What is your opinion on that?
Thank you for sharing your opinions.
David,
I am not sure if you understood the question properly.
My models are based on theories, literature, etc. The question is whether I am supposed to test just sig effect using moderator (e.g. socio-economic status) or all mediated effects as one of them which initially was not significant may become sig for example for low SES group... In my opinion this has little to do with fishing...
Best wishes
Witold
Witold, the significance of the paths are not important. Just go on an add your moderator effects to the model. You may find out that the conditional effect of a predictor variable on the outcome or the mediator variable will just be the mean conditional effect which is not significant, but for larger values of the moderator variable (or smaller values) this relationship will be significant. Therefore, just test your moderation effects as you originally planned.
Kind regards,
Karin
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling