Mediation analysis
A Causal inference with observational data method. But it can be used in combination with experiments. See also Instrumental variables estimation.
The basic idea is following (Baron1986moderator). When , running a regression without will show is predictive of . However, if is the mediator and there is no direct connection from to , adding M to the model will remove the effect of .
Imai2010general addresses the problem of the traditional approach and proposes an alternative. mediation is an R implementation.