Differential reinforcement is a core technique in Applied Behavior Analysis (ABA) that works by reinforcing desired behaviors while withholding reinforcement for problem behaviors. Rather than relying on punishment to reduce unwanted actions, it shifts the focus to building up better alternatives. The approach is used across a wide range of challenges, from aggression and self-injury to repetitive behaviors and elopement, and it’s one of the most well-supported strategies in ABA practice.
How Differential Reinforcement Works
Every behavior is maintained by some form of reinforcement. A child who screams during homework might be reinforced by getting a break from the task. A student who calls out in class might be reinforced by getting the teacher’s attention. Differential reinforcement targets this cycle by doing two things at once: it provides reinforcement when a more appropriate behavior occurs, and it removes reinforcement when the problem behavior occurs.
The “removing reinforcement” part is called extinction. If a child’s screaming no longer produces a break from homework, but calmly asking for a break does, the screaming gradually weakens while the calm request strengthens. This dual process, reinforcing what you want and withholding reinforcement for what you don’t, is what makes the technique “differential.”
The Five Types of Differential Reinforcement
There are five main types, each suited to different situations. They share the same logic but differ in what exactly gets reinforced.
DRA: Reinforcing an Alternative Behavior
Differential Reinforcement of Alternative Behavior (DRA) is the most commonly discussed type. It involves teaching and reinforcing a specific replacement behavior that serves the same function as the problem behavior. If a child hits to get access to a toy, the alternative might be handing a picture card to a therapist to request the toy. The replacement behavior needs to “work” for the same reason the problem behavior did, otherwise there’s little motivation to switch.
To make the alternative behavior more appealing, practitioners often adjust the effort involved. The replacement should be easier, faster, or more reliably rewarded than the problem behavior. If asking politely for a break takes 30 seconds but screaming gets an immediate response, the child has no reason to switch. The reinforcement for the alternative needs to be at least as quick and as valuable as what the problem behavior used to produce.
DRI: Reinforcing an Incompatible Behavior
DRI is a specific form of DRA where the replacement behavior is physically impossible to do at the same time as the problem behavior. A child who pulls their hair can’t do so while cuddling a stuffed animal. A student who runs out of the classroom can’t do so while sitting in their chair. Clasping your hands in your lap is incompatible with slapping your ears.
The advantage of DRI is its built-in logic: if the child is doing the reinforced behavior, the problem behavior literally cannot happen. This makes it especially useful for behaviors that are dangerous or disruptive enough that even brief occurrences are a concern.
DRO: Reinforcing the Absence of Problem Behavior
DRO doesn’t target any specific replacement behavior. Instead, it reinforces the child for simply not engaging in the problem behavior during a set time interval. If a child goes five minutes without hitting, they earn a reinforcer. Any other behavior during that window is fine.
Setting the right interval length is critical. The starting point is usually based on how often the behavior currently occurs. If a student shouts out roughly every three minutes during observation, the initial interval might be set at two and a half minutes, just short enough that the child can realistically succeed. As the child consistently meets the criterion, the interval gradually increases. A five-minute window becomes seven, then ten, then fifteen, until the behavior has faded substantially.
DRL: Reinforcing Lower Rates of Behavior
Some behaviors aren’t problems in themselves but become problems when they happen too often. A student raising their hand to ask questions is appropriate. Raising their hand 25 times in a 20-minute work period is disruptive. DRL is designed for exactly this situation: it reinforces the behavior only when it occurs below a set threshold.
In practice, a teacher might tell a child the maximum number of times they can ask for attention during a work session. If the child stays within that limit, they earn a reward. The threshold is set based on both baseline rates and what’s reasonable for the setting. Over time, the allowed number can be gradually reduced until the behavior reaches an acceptable frequency. DRL is the right fit when you want less of a behavior, not zero.
DRH: Reinforcing Higher Rates of Behavior
DRH is the mirror image of DRL. It reinforces a desired behavior only when it occurs more frequently than a set criterion. This type is less commonly discussed but useful when a child already has a positive behavior in their repertoire that simply doesn’t happen often enough, like participating in group activities or completing work tasks independently.
Where Differential Reinforcement Is Used
Differential reinforcement procedures have been applied to a broad range of behavioral challenges. In ABA settings, they’re commonly used for tantrums, aggression, self-injury, elopement (running away), stereotypic or repetitive behaviors, food selectivity, inappropriate verbalizations, and even specific issues like needle phobia and skin picking. The technique is considered a foundational intervention for interfering behaviors, particularly when the child already has some appropriate behaviors in their repertoire that can be strengthened.
Research has also explored differential reinforcement outside of autism-related ABA. A meta-analysis of DRO for tic disorders in children found that the technique was effective in reducing tic frequency across eight studies involving 79 children, though researchers noted the evidence base remains small. The flexibility of differential reinforcement, the fact that it can target virtually any measurable behavior, is part of why it appears across so many contexts.
The Role of Extinction
Most differential reinforcement procedures pair reinforcement for the desired behavior with extinction for the problem behavior. This combination is what drives behavior change. But extinction on its own can produce a temporary spike in the problem behavior before it decreases, sometimes called an extinction burst. A child who used to get a break by screaming might scream louder or longer when it stops working, at least initially.
This is why differential reinforcement is more effective than extinction alone. The child isn’t just losing access to reinforcement for one behavior. They’re simultaneously gaining access to reinforcement for a better one. The replacement behavior gives them a new path to the same outcome, which reduces the frustration that drives extinction bursts. Making the alternative behavior low-effort and ensuring it produces reinforcement quickly and reliably are key strategies for keeping this transition smooth.
How Reinforcement Is Faded Over Time
In the early stages, reinforcement for the replacement behavior is delivered consistently and immediately. Every time the child hands over a picture card instead of hitting, they get what they’re requesting. But this density of reinforcement isn’t sustainable in everyday life. A teacher can’t provide individualized reinforcement every 30 seconds indefinitely.
Schedule thinning is the process of gradually reducing how often reinforcement is provided. It typically follows a dense-to-lean progression: reinforcement starts on a very rich schedule, and once the child demonstrates sustained reductions in problem behavior across several sessions, the gap between reinforcements is slowly increased. For example, a therapist might start with 30 seconds of unavailability between reinforcements, move to one minute after three successful sessions, then continue stretching the interval until a practical, real-world schedule is reached. Each step forward only happens after the child shows they can maintain the replacement behavior at the current level.
The goal is to reach a “terminal schedule” that caregivers and teachers can realistically maintain throughout the day. Rushing this process risks a return of problem behavior, so the pacing is driven entirely by the child’s performance rather than a fixed timeline.
Choosing the Right Type
The choice between DRA, DRI, DRO, DRL, and DRH depends on the specific behavior and the goal. If you need to eliminate a dangerous behavior like self-injury, DRI or DRO are strong options because they target the problem behavior directly. If the child needs a functional replacement skill, like learning to request a break instead of throwing objects, DRA is the natural fit. If the behavior is acceptable in moderation but happens too often, DRL preserves the behavior while reducing its frequency.
In many treatment plans, these procedures overlap or are combined. A child might receive DRA to build a communication skill while DRO is used during specific activities where the problem behavior is most likely. The common thread across all types is the same: make the right behavior pay off better than the wrong one, and the child’s behavior will shift accordingly.