Identifying causal effect

Given a causal graph and the set of observed variables, identification of causal effect is the process of determining whether the effect can be estimated using the available variables’ data. Formally, identification takes the target causal effect expression, e.g., \(E[Y|do(A)]\), and converts it to a form that can be estimated using observed data distribution, i.e., without the do-operator.

For an introduction to identification in causal inference, check out the book chapter.

• Backdoor criterion
• Frontdoor criterion
• Natural experiments and instrumental variables
• ID algorithm for discovering new identification strategies