Potential Outcomes

The potential-outcomes framework imagines that every unit has two (or more) outcomes waiting in advance: the outcome if treated and the outcome if not treated. The causal effect for that unit is the difference between these two potential outcomes.

Because we only ever observe one of them — the one matching the treatment the unit actually received — the other is the missing counterfactual. This bookkeeping, often called the Rubin causal model, makes assumptions precise and is the language behind treatment effects, matching, and most modern causal methods.