I am conducting a mediation analysis with the mediation package in R. As a result I get values for an average direct effect (ADE) and an average causal mediation effect (ACME).

In the literature such mediations are often drawn as path models in which X represents the predictor, M the mediator and Y the dependent variable. X -> Y is then called c', X -> M becomes a and M -> Y is called b.

As I see it, the ADE represents the path c'. ACME is the product of paths a and b, that is ab.

How do I get values for a and b?

Your help is much appreciated!


When you run mediate(), you need to supply two models: the mediator model (M ~ T + X1 + X2) and the outcome model (Y ~ M + T + X1 + X2). The coefficient for T in the mediator model is a, and the coefficient for M in the outcome model is b. Note that these values only (directly) relate to the ACME and ADE if the outcome and mediator are continuous and there is no treatment-mediator interaction on the outcome.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.