Differential reinforcement of successive approximations refers to the procedure of reinforcing only some responses and not others. Differential reinforcement is used in the "shaping" of complex behaviors. (see Shaping.)