Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. It only takes a minute to sign up.
Sign up to join this community
Does it make sense to test whether the effect of X on Y is mediated by M if X and M are not significantly correlated?
And what if M and Y are not significantly correlated either?
I know that according to Baron and Kenny's approach, it wouldn't make sense, but I also read that this approach is not state of the art any more and some of their preconditions (e.g. One needs to establish that X is related to Y in order to carry out mediation) are actually not true.
Thanks a lot!
What is required for mediation is that X has an effect on M and M has an effect on Y. If either effect is not significant than that does not bode well for your mediation effect.
What might happen is that one of the constituents effects is just insignificant and the other effect is very strong, in which case the mediation effect might still be significant. However, I would not consider that to be a very convincing result.
