A quasi-experimental method that exploits threshold-based treatment assignment to estimate causal effects for units near the threshold.