Differential Reinforcement in ABA: DRA, DRI, DRO and DRL

Differential reinforcement in ABA means reinforcing one behavior while withholding or reducing reinforcement for another, so the reinforced behavior grows and the other one fades without punishment. There are four main types: DRA (alternative behavior), DRI (incompatible behavior), DRO (other behavior, meaning the absence of the problem behavior) and DRL (low rates). They differ in one thing: what has to happen for the learner to get the reinforcer.

What Is Differential Reinforcement in ABA?

Differential reinforcement has two parts running at once. One response keeps producing reinforcement. Another response stops producing it, or produces much less. The difference between the two is what changes behavior, which is where the name comes from.

The textbook version pairs reinforcement with extinction: the problem behavior gets nothing at all. In practice that isn't always possible. You can't always block the reinforcer for a behavior, and sometimes it isn't safe to try. Vollmer, Peters, Kronfli, Lloveras and Ibañez (2020) argued that DRA "should not require extinction, either procedurally or by definition." Their definition is giving greater reinforcement, along at least one dimension, for one behavior while minimizing reinforcement for another. So if asking for a break gets a longer, faster break than screaming does, that's still DRA, even though screaming isn't fully on extinction.

DRA also has a deep evidence base. Petscher, Rey and Bailey (2009) reviewed 30 years of research and coded 116 empirical studies that used DRA. After grouping them into categories that met the American Psychological Association's Division 12 Task Force criteria, they found DRA, with and without extinction, well established for treating destructive behavior in people with developmental disabilities and for food refusal.

The Types of Differential Reinforcement: DRA, DRI, DRO and DRL

DRA: differential reinforcement of alternative behavior

You pick a specific replacement behavior and reinforce it, ideally with the same reinforcer the problem behavior was getting. A child who grabs toys gets the toy for pointing or asking, and grabbing stops working. When the replacement is a communication response, such as asking for a break or for attention, the procedure is usually called functional communication training (FCT).

DRI: differential reinforcement of incompatible behavior

DRI is DRA with a stricter rule: the replacement has to be physically impossible to do at the same time as the problem behavior. Hands on the desk can't co-occur with hair pulling. Sitting can't co-occur with running around the room. The upside is that every second of the replacement is a second without the problem behavior. The downside is that incompatible behaviors often don't serve the same function, so the learner still has no way to get what the problem behavior was getting.

DRO: differential reinforcement of other behavior

DRO doesn't name a replacement at all. You deliver the reinforcer when the problem behavior hasn't happened, either across a whole interval or at the instant the interval ends. Despite the name, DRO really reinforces the absence of a behavior, and whatever the learner happened to be doing at reinforcement time gets strengthened too. If a learner was quietly rocking or flicking their fingers when the timer went off, that's what you just reinforced. DRO earns its place when there's no replacement you can teach yet.

DRL: differential reinforcement of low rates

DRL reinforces the problem behavior itself, as long as it stays at or below a limit. It's for behavior that's fine in moderation: reassurance questions, phone checks, talking in class. Deitz and Repp (1973) ran three classroom studies with it. In the first, a student's talking out dropped when the teacher gave 5 minutes of free time for a talk-out rate under 0.06 per minute. In the third, a whole high school business class reduced off-task talk under successively decreasing DRL limits. The teachers found it manageable.

DRH, differential reinforcement of high rates, is the mirror image of DRL. You reinforce a behavior when it reaches or passes a target rate, so it's used to build behavior up, not reduce it. It appears on the BCBA outline alongside the reduction procedures, which is why students tend to see it grouped with them.

Differential Reinforcement Comparison Table

All five types side by side, with the situation each one fits.

DRO in ABA: Whole-Interval vs. Momentary

DRO has a decision inside it that most explanations skip. In whole-interval DRO, the learner earns the reinforcer only if the behavior didn't happen at any point during the interval. In momentary DRO, the learner earns it if the behavior isn't happening at the instant the interval ends. Here's the same minute of behavior under each.

The target behavior showed up in 4 of the 6 intervals. Whole-interval DRO delivered 2 reinforcers, one for each clean interval. Momentary DRO delivered all 6, because the behavior never happened to be occurring at a check, and 4 of those 6 reinforcers came at the end of intervals that contained the behavior you're trying to reduce.

In practice, momentary DRO has looked weaker on its own. Repp, Barton and Brulle (1983) compared the two schedules for students' disruptive behavior. For two students, momentary DRO came first and didn't work, while the whole-interval DRO that followed did. For a third student, whole-interval DRO came first and worked, and the reduction held when the schedule switched to momentary. In a second experiment the momentary schedule reduced responding, but the whole-interval schedule was more effective.

Momentary DRO is much easier to run, though, because you only have to look up at the end of each interval. Lindberg, Iwata, Kahng and DeLeon (1999) compared fixed-interval, variable-interval and variable-momentary DRO with three participants whose self-injury was maintained by social positive reinforcement. All three schedules reduced self-injury, and for one participant variable-momentary DRO worked on its own. The authors suggested momentary schedules as an alternative to fixed-interval DRO partly because they don't require continuous monitoring. Taken with Repp's third student, that's an argument for trying a momentary schedule when continuous observation isn't realistic, and for watching the data closely when you do. The same trade-off between accuracy and effort shows up in data collection, covered in our guide to continuous vs. discontinuous measurement.

One more caution on DRO. In Vollmer and colleagues (1993), both DRO and noncontingent reinforcement reduced attention-maintained self-injury in three participants, probably because both delivered the functional reinforcer, attention. The authors reported evidence that noncontingent reinforcement reduced several of DRO's limitations. So if you're choosing DRO because you know the function but have no replacement to teach yet, noncontingent reinforcement belongs on the list of options too.

Differential Reinforcement and Extinction: What Happens in Practice

When differential reinforcement does include extinction, plan for the extinction burst. A behavior that used to pay off reliably can get worse before it gets better: more often, more intense, sometimes in new forms. If anyone on the team delivers the reinforcer during the burst, the learner has just learned that escalating works.

That's the strongest argument for a rich schedule on the replacement behavior at the start. If asking for a break produces a break every time, right away, and screaming produces nothing, the learner has an easy route to the same outcome and less reason to escalate. If the replacement takes more effort than the problem behavior and pays no better, expect the problem behavior to win.

How to Choose a Differential Reinforcement Procedure

Start with the function. Differential reinforcement works best when the replacement gets the same reinforcer the problem behavior was getting, and that's only possible once you know what that reinforcer is. Our guide to the four functions of behavior covers how that's identified. Then work through these three questions.

1. Should the behavior stop completely, or just happen less? If less, use DRL. If completely, keep going.
2. Can you teach a replacement that gets the same outcome? If yes, use DRA. If you also need the replacement to make the problem behavior physically impossible, pick one that's incompatible with it, which makes it DRI, and check that it still meets the learner's need.
3. Is there no replacement you can teach yet? Use DRO, start with whole-interval, and watch what else you're reinforcing.

The differential reinforcement glossary entry has the one-paragraph definition.

Differential Reinforcement on the RBT and BCBA Exams

Both exam outlines list these procedures by name, though not the same ones. The RBT outline names DRI and the BCBA outline doesn't. The BCBA outline names DRH and the RBT outline doesn't.

On the RBT exam, the RBT Test Content Outline (3rd ed.) puts differential reinforcement in Domain D, Behavior Reduction, which carries 14 of the 75 scored questions. Task D.3 is "Implement differential reinforcement procedures (e.g., DRO, DRA, DRI, DRL, FCT)." Task D.4, right after it, is "Implement extinction procedures." For an RBT, the questions are about running the procedure: which behavior earns the reinforcer, and what you do when the problem behavior happens.

On the BCBA exam, the BCBA Test Content Outline (6th ed.) has task G.2: "Design and evaluate differential reinforcement (e.g., DRA, DRO, DRL, DRH) procedures with and without extinction." Domain G, Behavior-Change Procedures, carries 25 of the 175 scored questions, the most of any domain. Note the phrase "with and without extinction," which matches the definitional debate above. The full breakdown is in our guide to the BCBA task list 6th edition.

Differential Reinforcement Practice Questions

These five are original questions written for this page. They aren't taken from any BACB exam. For a full-length run at RBT-level items, try our RBT practice exam.

Frequently Asked Questions

Sources

Last checked November 9, 2026.

1. Petscher, Rey and Bailey, A review of empirical support for differential reinforcement of alternative behavior (Research in Developmental Disabilities 30(3):409-425, 2009)
2. Vollmer, Peters, Kronfli, Lloveras and Ibañez, On the definition of differential reinforcement of alternative behavior (Journal of Applied Behavior Analysis 53(3):1299-1303, 2020)
3. Repp, Barton and Brulle, A comparison of two procedures for programming the differential reinforcement of other behaviors (Journal of Applied Behavior Analysis 16(4):435-445, 1983)
4. Lindberg, Iwata, Kahng and DeLeon, DRO contingencies: an analysis of variable-momentary schedules (Journal of Applied Behavior Analysis 32(2):123-136, 1999)
5. Vollmer, Iwata, Zarcone, Smith and Mazaleski, The role of attention in the treatment of attention-maintained self-injurious behavior: noncontingent reinforcement and differential reinforcement of other behavior (Journal of Applied Behavior Analysis 26(1):9-21, 1993)
6. Deitz and Repp, Decreasing classroom misbehavior through the use of DRL schedules of reinforcement (Journal of Applied Behavior Analysis 6(3):457-463, 1973)
7. BACB, RBT Test Content Outline (3rd ed.), tasks D.3 and D.4 and the Behavior Reduction domain weight
8. BACB, BCBA Test Content Outline (6th ed.), task G.2 and the Domain G weight