Testing Explicit Mediation Models by Set-Theoretic Causal Complexity

Although the linear mediation model developed by Baron and Kenny (1986) has made great contributions to the study of linear mediation mechanism, it has not substantially contributed to complex causality. Especially, when interactions or interdependent effects of multiple variables may occur in mediation mechanism, linear mediation model only considers individual variable’s separate mediation mechanism. Instead of a linear correlation, the explicit statement of necessity and sufficiency is needed for a falsifiable hypothesis in complex causality. In present article, we indicate the deficiencies of symmetry, simplification, and linearity relationship assumption in linear mediation model, and proposes a complex set-theoretic causality which can deal with asymmetry, complexity, and nonlinearity. Through the explicit discussion of various necessary or sufficient relationships of causal chains in complex causal mediation model, as well as the causal complexity characterized by conjunction and disjunction of multiple conditions, equivalent paths, asymmetric paths, the paper contributes to analyze the causal complexity of mediation mechanism. Instructions are provided on how to test complex causal mediation model.