Differential reinforcement of other behavior (DRO): a yoked-control comparison.
DRO suppresses behavior faster than matched noncontingent food because the reinforcer lands only after the response stops, but keep an eye on other behaviors that may rise.
01Research in Context
What this study did
The team worked with lab rats that pressed a lever for food.
They compared two 10-second and 30-second DRO schedules against yoked noncontingent reinforcement.
Each rat in the DRO group had a partner that got the same food deliveries on a fixed time, no matter what.
What they found
DRO quickly cut lever presses to near zero.
The yoked food-only group also pressed less, but never as low as DRO.
Rats began to wait by the food cup, yet the act of waiting did not clearly help or hurt the result.
How this fits with other research
Horner (1971) ran a near-copy study the same year and showed DRO beats straight extinction, giving an early clue that reinforcing "not pressing" is powerful.
Bensemann et al. (2015) later repeated the setup with humans and saw the same drop in target behavior, but also found untargeted responses rose—a side effect the rat study missed.
Rogers-Warren et al. (1976) went one step further in monkeys: reinforcing a specific alternative response cut the target even faster than DRO, hinting that pure DRO may not be the quickest path.
Hangen et al. (2020) blended DRO with extinction and again recorded the unwanted jump in other behavior, locking in the pattern first seen in animals.
Why it matters
For you in the clinic, the old rat data still hold: DRO works because any free food that follows a pause can accidentally strengthen that pause.
Watch for new, untargeted behaviors that may pop up while the problem behavior falls—measure them from day one.
If speed is crucial, pair DRO with a clear alternative skill the client can do instead; later monkey and human studies show this beats DRO alone.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Add a simple tally for any new behavior that shows up while you run DRO, then plot both curves to see if you're accidentally strengthening something else.
02At a glance
03Original abstract
After training to press a lever on a variable-interval 30-sec schedule, one group of rats was shifted to a differential-reinforcement-of-other-behavior 10-sec schedule, while a second group was shifted to a noncontingent yoked-control schedule that provided the same frequency and distribution of reinforcement. Then, both groups were extensively retrained on the variable-interval schedule, after which the first group was shifted to a series of differential-reinforcement-of-other-behavior 30-sec sessions alternating daily with variable-interval 30-sec sessions, while the second group was treated like the first on variable-interval days and yoked with the first as before on differential-reinforcement-of-other-behavior days. In both phases, response-decrement was more rapid and more marked in the differential-reinforcement-of-other-behavior animals than in the controls. The difference was due, at least in large measure, to sustainment of response in the control animals by adventitious reinforcement. All the differential-reinforcement-of-other-behavior animals developed "other" behavior-the same distinctive pattern of waiting at the foodcup-but there was no direct evidence that it contributed in any way to the decrement in lever pressing.
Journal of the experimental analysis of behavior, 1971 · doi:10.1901/jeab.1971.15-237