The terms “moderators” and “mediators” have been used in both behavioral and medical research for at least the last half century. Early on, however, the terms were typically used idiosyncratically, meaning whatever seemed to be implied from the context of the statement in which they were used. The approach of Baron and Kenny (1986) provided strong conceptual definitions, but their mathematical approach to documenting which definitions applied, proved problematic. The MacArthur approach appearing around 2000 was an update to deal with those problems and is here presented, with illustrations and comments. Also discussed is how this approach resolves current conceptual problems such as “backward” and “reciprocal” causation. There are, however, as yet unresolved problems in its application, such as how to convey the clinical significance of moderation and mediation.