Regression discontinuity is when assignment to a development program is made based on some sort of score (e.g., a poverty score or a test score), and we compare the people immediately above and below the score to each other, since they should be very similar except for (a) a few points on a test score and (b) some people received the program.