A method of estimation designed to find the causal effect of a discontinuous treatment, that is, a treatment applied when a certain variable, called the assignment (or forcing, or running) variable exceeds a known threshold. A classic example is the estimation of the effect of merit awards on subsequent academic outcomes when the award is given based on a test score. Assuming that individuals with a score just below the award threshold do not differ very much from those with a score just above the threshold, the difference in outcomes can be attributed to the effect of receiving the award.