|
Abstract:
|
Causal Mediation Analysis is concerned with distinguishing different causal pathways that may link a treatment and an outcome. In the simplest case we may wish to know whether a treatment acts directly on the outcome or via a mediator. We describe a conceptual framework for analyzing mediation (outlined in Robins & R 2010) based on decomposing the treatment into sub-components. We contrast this with a non-interventionist approach advocated by Pearl (2001) based on nested counterfactuals. The interventionist approach has several advantages: It does not require the existence of well-defined interventions on mediators; Identified effects are, in principle, empirically testable via interventions; The new theory preserves the dictum "no causation without manipulation"; It replaces the definitions of path specific effects as nested counterfactuals with definitions in terms of concrete experimental interventions; It facilitates communication with subject matter experts. However, when both effects are identified from data, the identifying formulae under both approaches are identical, even though the effects are different.
|
