If the standard estimate in the first model is more than the second in mediation analysis, in that context can we consider that the predictor in the first model is stronger than the second one?
No not necessarily, a larger indirect effect (/standardized estimate) in Model 1 than Model 2 does not automatically mean that the predictor in M1 is "stronger" than in M2 in a general sense (as the indirect effect depends on both the path from the predictor to the mediator + the path from the mediator to the outcome, the standardized estimates are affected by the variances of the variables (if the predictor or mediator has more variability in one model, this can inflate the standardized estimate), list goes on)
If you want to compare indirect effects formally you have to use a multi-group analysis and test for differences in the indirect effect
- Badges
- Science topic