Direct and Indirect Effects of Sequential Treatments

TL;DR

This paper reviews the concept of direct causal effects, reformulating it without counterfactuals, and explores conditions for identifiability in sequential treatments using graphical methods.

Contribution

It introduces a counterfactual-free formulation of direct effects and extends the concept to sequential treatments with new identifiability conditions.

Findings

Counterfactual-free formulation of direct effects

Graphical criteria for identifiability of sequential treatments

Comparison with existing criteria by Pearl and Robins

Abstract

In this paper we review the notion of direct causal effect as introduced by Pearl (2001). We show how it can be formulated without counterfactuals, using intervention indicators instead. This allows to consider the natural direct effect as a special case of sequential treatments discussed by Dawid and Didelez (2005) which immediately yields conditions for identifiability as well as a graphical way of checking identifiability. The results are contrasted with the criteria given by Pearl (2001) and Robins (2003).