Designs for Studying Mediation

Once we find a cause-effect relationship, the next question is related to the mechanisms of change. Mediators play a prominent role in establishing these mechanisms because they are variables that intervene between the cause (i.e., predictor/independent variable) and its effect (i.e., outcome/criterion/dependent variable). A wide range of designs can focus on mediators. These include both nonexperimental designs (i.e., cross-sectional or longitudinal) and experimental designs. Experimental designs are more appropriate designs for studying mediation but no single study can meet all prescriptive requirements for considering a specific mediator as a causal mechanism. A strong recommendation is that any endeavor to establish mediation should be complemented by a conceptual analysis through the lens of a well-supported and/or well-specified theory. Keywords: experimental psychology; mental health; mediation; methodology; moderation; research; theory/mechanism of change