Differential reinforcement is a behavior change technique that works on a simple principle: you reinforce the behaviors you want to see more of while withholding reinforcement for the behaviors you want to reduce. Rather than relying on punishment to stop unwanted behavior, it reshapes behavior by making the desired alternative more rewarding. It’s one of the most widely used strategies in Applied Behavior Analysis (ABA) and shows up in classrooms, therapy settings, and homes.
How Differential Reinforcement Works
Every behavior is maintained by some kind of payoff. A child who throws a tantrum in the grocery store may be doing it because it gets a parent’s attention, because it results in getting a toy, or because it lets them escape a situation they don’t like. Differential reinforcement targets that payoff directly. Instead of reinforcing the problem behavior (even unintentionally), you redirect the reinforcement toward a more appropriate behavior that serves the same purpose for the person.
The technique has two active components working together. First, you deliver reinforcement when the person engages in a desired behavior. Second, you withhold reinforcement when the problem behavior occurs, a process called extinction. This combination is what makes differential reinforcement effective: the person learns that the new behavior “works” while the old behavior no longer does. Over time, the desired behavior increases and the problem behavior fades.
Choosing the right approach starts with understanding why the problem behavior happens in the first place. Behavior analysts typically conduct a functional behavior assessment to determine whether the behavior is driven by a desire for attention, access to a tangible item, escape from a demand, or sensory stimulation. That function then guides which type of differential reinforcement to use and what the replacement behavior should be.
Types of Differential Reinforcement
There are several variations, each designed for a different situation. The core logic is always the same: reinforce what you want, stop reinforcing what you don’t. The differences come down to what exactly you’re reinforcing.
Reinforcing an Alternative Behavior (DRA)
DRA involves teaching and reinforcing a specific alternative behavior that serves the same function as the problem behavior. The replacement behavior gives the person an appropriate way to get what they were already trying to get. For example, a child who hits to get access to an iPad might be taught to hand a therapist a picture card requesting iPad time instead. When the child uses the card, they get the iPad. When they hit, the request is ignored. The key is that the alternative behavior must meet the same need, otherwise the person has no reason to switch.
Reinforcing an Incompatible Behavior (DRI)
DRI is a more specific version of DRA. Here, the replacement behavior is chosen because it is physically impossible to do at the same time as the problem behavior. A child who pulls their hair can be reinforced for clasping their hands in their lap, because you can’t clasp your hands and pull your hair simultaneously. A student who runs out of the room can be reinforced for sitting in their chair. The physical incompatibility creates a natural barrier: the more the person engages in the replacement behavior, the less opportunity there is for the problem behavior to occur.
Other examples organized by what’s driving the behavior:
- Attention-seeking behavior: Reinforce polite requests for attention; do not respond to whining.
- Wanting a tangible item: Give a child more blocks when they play nicely instead of when they grab toys from others.
- Trying to escape a task: Give a break when the learner completes a portion of work instead of when they destroy materials.
Reinforcing the Absence of Problem Behavior (DRO)
DRO takes a different approach. Instead of reinforcing a specific replacement behavior, it reinforces the absence of the problem behavior over a set time interval. If a child goes 10 minutes without engaging in the target behavior, they earn reinforcement. The goal is to gradually stretch the time between occurrences of the problem behavior. DRO is particularly useful when the problem behavior doesn’t have a clear alternative, or when the priority is simply reducing how often it happens. A meta-analysis of DRO used for tic disorders found significant reductions in both tic severity scores and tics per minute after intervention.
Reinforcing Lower Rates of Behavior (DRL)
Sometimes a behavior isn’t a problem by itself but happens too often. DRL is designed for exactly this situation. It doesn’t aim to eliminate the behavior, just bring it down to a reasonable level. If a student named James greets his classmates so frequently during class that it becomes disruptive, a DRL plan might set a limit of five greetings per class period. If James stays within that limit, he earns a reward. He’s still allowed to socialize; he just learns to moderate the frequency.
Reinforcing Higher Rates of Behavior (DRH)
DRH is the mirror image of DRL. It’s used when a behavior is beneficial but isn’t happening often enough. Reinforcement is delivered when the behavior occurs at or above a set rate within a given time period. A student who rarely participates in class discussions, for instance, might earn reinforcement for contributing at least three times per session. The threshold can be adjusted upward as the behavior increases.
What to Expect During the Process
One of the most important things to know before starting any differential reinforcement plan is that behavior often gets worse before it gets better. When reinforcement for a problem behavior is suddenly withdrawn, the person will typically try harder, doing the behavior more frequently, more intensely, or for longer stretches. This temporary spike is called an extinction burst, and it’s a normal part of the process.
During an extinction burst, you might see new variations of the behavior emerge. A child who used to whine might start screaming. Someone who tapped a desk might begin pounding it. Emotional reactions like frustration or aggression can appear. This escalation is a sign that the person has noticed the old behavior no longer works, and they’re testing whether a louder or more persistent version might. It’s uncomfortable, but it typically means the intervention is having an effect.
The critical factor during this phase is consistency. If you give in during the burst, you teach the person that escalation is what works, which makes the problem behavior even harder to change later. Caregivers and teachers who anticipate the burst and plan for it are far more successful at riding it out. Team support helps too. Regular check-ins among everyone involved keep morale up and ensure everyone is responding the same way. If the behavior poses a safety risk during the burst, strategies like offering a break card or using safe prompting can help manage the situation without reinforcing the problem behavior.
Why the Function of Behavior Matters
Differential reinforcement doesn’t work as a one-size-fits-all tool. The replacement behavior you choose has to match the reason the problem behavior exists. If a child screams because they want to escape math homework, reinforcing them with a sticker for being quiet won’t address the underlying need. The child doesn’t want a sticker. They want a break. An effective plan would teach them to calmly request a short break and then honor that request when they use the appropriate communication.
This is why a functional assessment is considered essential before designing an intervention. Without understanding whether behavior is driven by attention, escape, access to something tangible, or sensory input, you’re guessing at what the replacement behavior should be. And a mismatch between the function and the replacement is one of the most common reasons differential reinforcement plans fail. Current best practice guidelines also emphasize incorporating the individual’s own preferences when selecting interventions. Treatments that feel acceptable and meaningful to the person are more likely to stick over time.
Putting a Plan Together
A practical differential reinforcement plan involves several connected decisions. First, you identify the problem behavior in specific, observable terms. “Being disruptive” is too vague. “Calling out without raising a hand” is something you can count and track. Next, you determine the function of that behavior through observation or a formal assessment.
Then you select the type of differential reinforcement that fits the situation. If there’s a clear replacement behavior that meets the same need, DRA or DRI is the likely choice. If the goal is simply reducing the behavior’s occurrence without a specific replacement, DRO works well. If the behavior is acceptable at lower levels, DRL makes sense.
Once you’ve chosen the approach, you set the reinforcement schedule. This means deciding how often and under what conditions the person earns reinforcement. Starting with a rich schedule (frequent reinforcement) and gradually thinning it over time helps the new behavior take hold without creating dependency on constant rewards. You also need a clear plan for what happens when the problem behavior occurs, which in most cases means it receives no reinforcement at all.
Finally, you track the data. Counting how often the target behavior and the replacement behavior occur before, during, and after the intervention tells you whether the plan is working or needs adjustment. Behavior change is rarely instant, so consistent monitoring over weeks gives a much clearer picture than day-to-day impressions.