I am comparing two models using Path Analysis (pictured).
Model A explained 49% of the variance in my DV.
Model B (which is the same as Model A + an additional mediator) explained 48% of the variance of the DV.
So Model A appears to be the better model - However it makes more theoretical sense to include the Mediator (M3) from Model B. Moreover, indirect effects of the predictors through this mediator are significant.
Model Fit from both models is also good, but slightly better for Model A.
I'm hoping someone might be able to provide some advice around how to interpret these findings? How is it that the indirect effects of the predictor through the new mediator is significant, but the variance doesn't change?
Hello Lisa,
If one of your research goals is to generate the simplest possible model that maximizes explanatory power for your DV, then the first model wins that contest. Parsimony frequently trumps complexity.
What you didn't do, however, was:
(a) explain how much of a change in explaining variance in M1 and M2 is accomplished by the introduction of M3 into the model; or
(b) explain why M3 is not a potential mediator of DV.
Good luck with your work!
David Morse Thank you for very much for your help. In response to your questions;
a) adding M3 did increase the variance explained in both M1 & M2 (though not greatly - see attached image).
b) the pathway between M3 and the DV was not significant so was removed from the model
Just to add to the above. Both IV's also explained 53% of the variance in M3. So I am struggling to understand how adding this mediator does not explain additional variance in the DV
Hello Lisa,
Well, for your sample, the performance of M3 just doesn't bring a lot extra to the party as regards accounting for differences in DV (above and beyond what M1 and M2 can do in conjunction with IV2 and IV1. Note that you don't have a direct effect for IV1 on DV (so, there's evidence for mediation; albeit not necessarily from M3, except via M1 and M2.
There are many reasons why M3 might not be as potent as you had perhaps envisioned. (a) M1 and M2 account for nearly all of what M3 could have accomplished as a mediator; (b) your sample is peculiar in some way; (c) the measures used for one or more of the variables aren't as good technically (reliability, validty) as they might be; or (d) some combination.
Good luck with completing your write-up!
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling