Differential Reinforcement of Other Behavior

Description of the Strategy

A reinforcement schedule designed to differentially reinforce other behavior (DRO) is useful when the goal is to gradually eliminate a behavior. When a DRO schedule is in effect, a reinforcer is presented only if an unwanted response does not occur during a predetermined period of time. Reinforcing an increase in the time without the behavior produces a decrease in the rate of the behavior. Because reinforcement is not contingent on the emission of any particular response, any response other than the specified response may be reinforced (e.g., sitting still, sleeping, or doing “nothing”). The DRO schedule is also described as differential reinforcement of zero rates [Page 805]or as omission training because reinforcement is delivered for a zero rate (the omission) of the specified response. It is sometimes combined with punishment. For example, if no response occurs during the DRO interval, reinforcement is delivered; if a response does occur, the punishment procedure (i.e., response cost, time-out) would be immediately applied and a new interval would begin.

There are two variations of the DRO schedule. Under whole-interval DRO, the selected behavior must be absent for the entire interval for the reinforcer to be delivered. For example, if a response (whining behavior) occurs about once every 5 minutes during baseline, the initial interval would be set at 5 minutes and reinforcement would be delivered after each 5-minute period that did not include whining. Under whole-interval DRO, the occurrence of the unwanted response before the interval elapses either resets the interval or cancels the reinforcer programmed at the end of the interval. Thus, whining during the interval could either reset the period to 5 minutes or simply not deliver reinforcement at the end of the 5-minute period.

The other variation is momentary DRO. With this contingency, the problem behavior must be absent at the end of the interval for the reinforcer to be delivered. For example, reinforcement at the end of each 5-minute interval would occur if the whining behavior was absent when the interval ended. If whining continued for 4 minutes but was absent when the interval ended, reinforcement would be presented. Similarly, if whining had been absent for more than 4 minutes but occurred just as the interval ended, the reinforcer would be cancelled. The primary advantage of momentary DRO is that it is easy to monitor because evaluation need only take place at the end of the DRO interval. The disadvantage is that the target behavior may be adventitiously reinforced if it occurs at times other than the end of the DRO interval.

Research Basis

A protocol should be followed when implementing the DRO procedure. First, it is essential to identify the reinforcer that is maintaining the problem behavior. This necessitates a functional assessment prior to implementing the DRO procedure. Research indicates that extinction is an essential component of the DRO procedure. That is, the effectiveness of the DRO contingency is dependent upon eliminating the reinforcing consequence that maintains the problem behavior. Second, a functional assessment should empirically identify a consequence that functions as a reinforcer for the individual. Of course, one consequence that is certain to function as a reinforcer is the consequence (identified through a functional assessment) that is targeted for extinction. When self-injurious behavior (SIB) is maintained by contingent attention, procedures such as DRO (using attention) are effective in reducing SIB, particularly in the early stages of the intervention. Third, there must be an empirical determination of the time interval during which the behavior should not occur. The initial DRO value should be approximately equal to the mean value between occurrences of the target behavior as determined by a representative baseline period. Fourth, the DRO duration should be gradually increased to ensure frequent reinforcement until an acceptable rate of the target behavior is maintained. Fifth, the occurrence of the target behavior should be monitored in other settings to identify instances of behavior contrast. Finally, generalization of the absence of the target behavior to new situations should be systematically expanded into various settings. When this recommended protocol is followed, reduction of the target behaviors is usually quite rapid and tends to be maintained. If difficulties arise, look at whether the initial interval is too long or whether the interval without the behavior was increased too rapidly.

...