Ignorability is an assumption under which treatment assignment carries no information about relevant potential outcomes after conditioning on specified covariates.
In a binary-treatment potential outcomes model, conditional exchangeability given pretreatment covariates is . Within each covariate stratum, treatment assignment then carries no information about the relevant potential outcomes. Together with consistency in causal inference and positivity, it permits outcome standardization to identify average causal contrasts. The assumption concerns unobserved counterfactual outcomes and cannot be established just by checking balance of the measured covariates. Including a large list of predictors is not by itself a guarantee that the assumption holds.
Articles by others on the same topic
Ignorability, often referred to in the context of causal inference and statistical modeling, refers to a condition under which the treatment assignment in an observational study can be considered as if it were randomized. This concept is crucial for identifying causal effects from observational data, as it allows researchers to make valid inferences about treatment effects without the biases typically associated with non-randomized studies.