Differential Reinforcement

Differential reinforcement (DR) is a procedure that is defined by the presence of two conditions: (a) positive reinforcement, delivered contingent on the emission of one behavior, and (b) withholding of positive reinforcement, contingent on the emission of another behavior. Delivery of these consequences may be event based (i.e., when a specified number of behaviors have been emitted) or time based (e.g., after passage of prespecified time and a behavior is emitted). If effective, an increase in the future occurrence of one behavior and a decrease in occurrence of the other behavior are observed.

If antecedent stimuli are manipulated systematically with DR, antecedents in a common stimulus class will occasion desired behavior, and antecedents that are not in that stimulus class will not occasion the desired behavior. In other words, if a desired behavior is emitted in the presence of a desired antecedent stimulus, contingent positive reinforcement is presented. If the same behavior is emitted in the absence of the desired antecedent stimulus (or in the presence of an undesired antecedent stimulus), positive reinforcement is withheld.

An antecedent stimulus that reliably occasions a particular behavior and signals positive reinforcement functions as a discriminative stimulus. Differential reinforcement is the process through which an antecedent stimulus becomes a discriminative stimulus. For example, to bring the letter b under the control of the associated letter sound /b/, a teacher would provide positive reinforcement only if the student's correct response—that is, saying /b/—occurs in the presence of the letter b. The teacher would not provide reinforcement for the same behavior if it were displayed following the presentation of different stimuli (e.g., p, d). In other words, the teacher would focus on presenting reinforcement whenever the correct response is emitted in the presence of a desired stimulus and would not provide positive reinforcement if the behavior was emitted in the absence of the desired stimulus.

Typically, as indicated, DR is used to increase behavior by linking a stimulus to the opportunity for reinforcement. However, DR can also be used to decrease the likelihood of occurrences of problem behaviors in four common variations: (a) differential reinforcement of incompatible behavior (DRI), (b) differential reinforcement of alternative behavior (DRA), (c) differential reinforcement of other behavior (DRO), and (d) differential reinforcement of low rates of behavior (DRL).

DRI and DRA are closely related. DRI involves the reinforcement of a behavior that is topographically incompatible with the problem behavior (i.e., the two behaviors cannot happen at the same time). For example, if a child roams the hallways during classtime, occurrences of staying in his or her seat would be differentially reinforced. Roaming and staying in seat are incompatible. In contrast, DRA involves positive reinforcement for occurrences of alternative behaviors that are functionally equivalent and more acceptable than the problem behavior but not topographically different. For example, two students who have frequent arguments might be placed on the same relay team. Occurrences of cooperative behavior would be positively reinforced, and occurrences of argumentative behavior would not be positively reinforced. Arguing and cooperative behaviors are functionally equivalent (i.e., result in peer attention), but occurrences of cooperative behavior are also associated with additional positive reinforcement.

...