Hi,
I don't know where to ask this question and I hope someone could give me some opinions.
I found that, in my study, fMRI signal of a region (as mediator M) mediated the relationship between a predictor (X) and an outcome (Y). In fact, X was positively correlated with Y, X negatively correlated with M, and M negatively correlated with Y. The model is:
M
(-) (-)
X (+) Y
(1) Here is my first question: If the mediation model is non-significant, I would say that both M and X predicted Y (I can maybe do a multiple linear regression analysis to differentiate the contribution of M and X). But if the model is significant, only X predicted Y. M did not predict Y but just acted as a mediator. Am I correct?
(2) Then is my second question. If the mediation model is:
M
(+) (+)
X (+) Y
You can see that X-Y correlation is still positive but the other two were negative. What is the difference in the significance between the two mediation model? Do they just lead to the same conclusion as previous model that M mediates the relationship between X and Y?
Thanks for any help.
Mark
|