Testing Identifiability of Causal Effects

TL;DR

This paper presents a systematic, polynomial-time method for testing the identifiability of causal effects in probabilistic models, providing explicit formulas when effects are identifiable.

Contribution

It introduces a polynomial-time procedure to determine causal effect identifiability and derives closed-form expressions for the probability of achieving specific goals.

Findings

Identifiability can be tested efficiently in polynomial time.

Closed-form formulas are available for identifiable causal effects.

The method applies to effects between a singleton variable and a set of variables.

Abstract

This paper concerns the probabilistic evaluation of the effects of actions in the presence of unmeasured variables. We show that the identification of causal effect between a singleton variable X and a set of variables Y can be accomplished systematically, in time polynomial in the number of variables in the graph. When the causal effect is identifiable, a closed-form expression can be obtained for the probability that the action will achieve a specified goal, or a set of goals.