Rubin Causal Model

The Rubin causal model, also known as the potential outcomes framework, formalizes causation as a comparison between counterfactual states of the world. For each unit, there exists a potential outcome under treatment and a potential outcome under control. The causal effect is the difference between these two potential outcomes.

The fundamental problem of causal inference is that only one potential outcome is ever observed — a unit cannot both receive and not receive the treatment. The framework solves this through randomized experiments, which create probabilistically equivalent groups, or through statistical methods that adjust for observed confounders under the assumption of no unmeasured confounding.

The Rubin model excels at estimating average treatment effects in settings with a single binary treatment and a well-defined outcome. It is less suited to systems with complex causal structures, multiple interacting treatments, or feedback loops — precisely the systems where do-calculus and structural models have advantages.