Design and Evaluate Differential Reinforcement Procedures With and Without Extinction

You’ve completed a functional behavior assessment. You know exactly why your client is engaging in problem behavior—what’s maintaining it, when it happens, and under what conditions. Now comes the critical next step: choosing an intervention that reduces the behavior while teaching something better in its place.

This is where differential reinforcement becomes your most valuable tool. Unlike punishment-based approaches, differential reinforcement works by reinforcing desirable behaviors while strategically withholding reinforcement for problem behavior. The procedure is flexible, ethically grounded, and aligns with the ABA principle of teaching skills rather than simply suppressing behavior.

The challenge isn’t whether to use differential reinforcement, but *how*—which type, whether to pair it with extinction, and how to design a plan your team can actually implement with fidelity. This article walks you through the decision points, common mistakes, and ethical guardrails so you can design a differential reinforcement plan that works in your setting.

What Differential Reinforcement Really Means

At its core, differential reinforcement increases desirable responses while decreasing problem behavior—all by manipulating what gets reinforced. Instead of using aversive consequences, you create a situation where the client learns that appropriate behavior pays off better than problem behavior.

The plain version: you reinforce something you want to see more of and withhold reinforcement for something you want to see less of. When done right, the client naturally shifts toward the behavior that actually gets rewarded.

The beauty of this approach is that it teaches. It doesn’t just stop a behavior; it builds something in its place. A child who screams for attention doesn’t just stop screaming—they learn to raise their hand. A student who taps the desk for sensory input doesn’t just suppress the tapping; they get access to a fidget tool instead. Over time, the replacement behavior becomes the client’s default response.

Four Types of Differential Reinforcement

Not all differential reinforcement looks the same. The procedure you choose depends on your goal and what replacement behavior makes sense.

DRA—Differential Reinforcement of Alternative Behavior reinforces a socially acceptable behavior that serves the same function as the problem behavior. If a child hits to gain a toy, you teach them to ask for it instead, then reinforce asking. DRA is the workhorse of ABA intervention—it’s direct, functional, and teaches a skill.

DRI—Differential Reinforcement of Incompatible Behavior reinforces a behavior that physically cannot happen at the same time as the problem behavior. If a student stands up during group work, you reinforce them for keeping their hands busy with a sensory tool while seated. The two behaviors are mutually exclusive.

DRO—Differential Reinforcement of Other Behavior reinforces the absence of the problem behavior for a set time interval. If a client usually taps incessantly, you set a 5-minute timer and reinforce them for not tapping during that window. DRO doesn’t teach a specific replacement skill; it simply marks periods when the problem behavior didn’t occur.

DRL—Differential Reinforcement of Low Rates reinforces the client for engaging in a behavior at a lower rate than baseline. This is useful when the behavior itself isn’t inherently inappropriate—like asking for breaks, which is good, but doing it 30 times per hour isn’t sustainable. You reinforce the behavior when it falls below a threshold.

Alongside these types, you may also use extinction—removing the reinforcer that maintains the problem behavior. If a child’s tantrums are maintained by attention, extinction means withholding attention when tantrums occur. Extinction is often paired with differential reinforcement to make the replacement behavior more attractive than the problem behavior.

Function-Based Selection: The Starting Point

You cannot choose a differential reinforcement procedure in a vacuum. Your selection must be grounded in the function of the behavior—what the client gains or avoids by engaging in it.

The four primary functions are attention, tangibles, escape, and sensory or automatic reinforcement. When you design differential reinforcement, you match the reinforcement you use to the function you identified.

If a child hits to gain attention, you reinforce a replacement behavior (like a hand-raise) with attention. If a child throws materials to escape a difficult task, you reinforce a replacement behavior (like a break request) with escape. If a behavior is maintained by sensory input, you provide an alternative source of that input (like a fidget) when they engage in an acceptable behavior instead.

This function-to-reinforcement match is non-negotiable. If you reinforce an alternative behavior with the wrong consequence, the client won’t switch—because the problem behavior still pays off better. Invest time in your functional assessment, and the differential reinforcement plan becomes much clearer.

With Extinction or Without? Making the Call

Here’s where many clinicians get stuck: Should you pair differential reinforcement with extinction, or can you just reinforce the replacement behavior and let the problem behavior fade on its own?

The answer depends on three things: safety, feasibility, and consistency.

Use differential reinforcement WITH extinction when the reinforcer maintaining the problem behavior can be safely and completely withheld, your team can deliver extinction consistently across all settings and people, and you’ve prepared for potential side effects. Pairing extinction with DRA is powerful because it removes the payoff for problem behavior while immediately rewarding the alternative. The client learns: “Tantrums don’t work anymore, but asking politely does.”

Use differential reinforcement WITHOUT extinction when you cannot safely withhold the maintaining reinforcer, consistency across all change agents is unlikely, the behavior is maintained by automatic reinforcement you cannot easily block, or you want to emphasize a reinforcement-only approach. In these cases, you reinforce the replacement behavior generously and let the relative lack of reinforcement for the problem behavior gradually reduce it. This is gentler and sometimes more sustainable—but it may take longer.

For example, if a child’s behavior is maintained by internal sensory stimulation (like hand-flapping), extinction isn’t practical. You cannot safely block someone’s sensory input. Instead, you teach and reinforce an incompatible behavior (like keeping hands occupied with a fidget) and trust that the available reinforcement steers the child toward the replacement.

If a student shouts for attention in a busy classroom where you cannot guarantee that every adult will ignore shouting consistently, extinction will backfire. Intermittent reinforcement—sometimes getting attention for shouting—actually strengthens behavior. In that setting, you’d focus on lavishly reinforcing a quiet hand-raise and accept that shouting will fade more slowly.

Designing Your Differential Reinforcement Plan: Key Steps

A sound differential reinforcement plan has several moving parts. Here’s how to build one:

Identify and define the target behavior and replacement behavior in observable, measurable terms. Don’t write “aggression” or “good behavior.” Write “hitting with a closed fist directed at a peer; occurs on average 8 times per hour.” And for the replacement: “Raising hand for attention; defined as arm extended above shoulder height, held for at least 2 seconds, without vocalizing.”

Select the DR procedure based on your goal and the function. If you need to teach a specific skill, use DRA or DRI. If you want to reduce rate of an acceptable behavior, use DRL. If you want to reinforce absence, use DRO. Your functional assessment guides this choice.

Choose your reinforcer and reinforcement schedule. The reinforcer must actually matter to your client—it must be motivating and available. Start with continuous reinforcement during the acquisition phase. Once the replacement behavior is reliable, shift to intermittent reinforcement for maintenance. If you’re pairing with extinction, continuous reinforcement for the replacement behavior is usually necessary at the start.

Get quick tips

One practical ABA tip per week.

No spam. Unsubscribe anytime.

Decide on extinction using the three criteria above. If you’re using it, write down exactly what you’ll withhold and how. Brief any staff or caregivers on what to expect—extinction bursts, variability, and possible emotional responses.

Set up your data collection system. How will you measure the target and replacement behaviors? Frequency? Duration? Interval recording? You need an objective measure to know whether the plan is working.

Create decision rules. When will you increase the reinforcement schedule? When will you fade extinction? When will you say the plan isn’t working and needs revision? Concrete decision rules prevent drift and keep the plan data-driven.

Ethical Guardrails and Informed Consent

Differential reinforcement is a low-risk procedure compared to punishment or restrictive interventions, but it still requires informed consent and ongoing ethical oversight.

Inform your client and their caregivers about what you’re doing and why. Explain that you’ve identified a function for the problem behavior and you’re going to reinforce something better instead. If you’re using extinction, describe it plainly: “We’re going to stop giving attention for tantrums so that asking politely becomes the behavior that pays off.” Mention potential side effects—especially extinction bursts, which can unsettle caregivers who don’t expect a temporary increase.

Get explicit, documented consent. This isn’t a verbal “sounds good.” It’s a written agreement showing the client or guardian understands the procedure, the rationale, and the expected outcomes. Include a plan for monitoring and adjusting if things go sideways.

Use the least restrictive, effective treatment. If you can achieve your goal with reinforcement alone, do that. If you can teach a skill instead of just reducing behavior, teach. If you can set the client up to succeed rather than waiting for failure, do it.

Monitor for distress and harm. If a client shows signs of increased emotional distress, aggression, or self-injury during extinction, pause and reassess. The plan might need to be less restrictive, you might need more staff training, or you might need environmental supports to make the replacement behavior more attractive.

Common Mistakes That Undermine Effectiveness

Even well-intended plans can fail if certain pitfalls aren’t avoided.

Inconsistent extinction is the silent killer. If you plan to ignore shouting for attention but staff sometimes respond, you’ve created an intermittent reinforcement schedule—the most powerful way to strengthen behavior. If you cannot guarantee consistency, don’t use extinction; rely on reinforcement alone instead.

Choosing a replacement behavior that doesn’t serve the same function sends a mixed message. If a child hits to escape a difficult task and you teach them to ask for a break, that works. But if you teach them to sit quietly, they’ve lost their escape route and hitting will likely continue or escalate.

Using DRO when the client needs a specific skill. If a child needs to learn functional communication but you only reinforce the absence of problem behavior, they leave the intervention without a skill. DRA or DRI teaches; DRO only suppresses.

Failing to prepare staff for extinction bursts. When you first withhold attention for tantrums, the tantrums often increase temporarily. Caregivers see this and think the plan isn’t working. They resume giving attention, inadvertently reinforcing the burst. Now the behavior is worse and more resistant. Brief everyone in advance: “Expect an increase for the first week or two. That’s normal. Stay consistent.”

Assuming extinction works in isolation. Never use extinction without pairing it with reinforcement for an alternative. Extinction alone creates frustration and teaches nothing.

Data Collection and Evaluation

You cannot know whether your plan is working without data. Set up a simple system before you start.

Choose a measurement method that fits your setting. For high-frequency behaviors, interval recording is often practical. For lower-frequency behaviors, frequency counts work fine. If duration matters, record duration.

Collect data consistently—ideally every day or every session. Plot it visually so you can see trends. Make it a standing agenda item at team meetings: “Are we seeing change? In what direction?”

Use pre-specified decision rules: If the behavior decreases by 20% within two weeks, the plan is moving. If there’s no change in four weeks and fidelity is good, revisit the function, the reinforcer, or the extinction component. If the behavior gets worse, stop and reassess immediately.

Include a fidelity check. Have someone observe implementation and confirm that your team is actually reinforcing the replacement behavior when they should and withholding reinforcement when they should. Fidelity failures are often why plans appear ineffective.

Three Real-World Scenarios

Scenario 1: Attention-Maintained Tantrum

Eight-year-old Maya throws tantrums (screaming, crying, flopping on the ground) to gain adult attention. Functional assessment confirms the behavior is maintained by attention.

• Target behavior: Tantrum (crying, screaming, or lying on floor for more than 10 seconds)
• Replacement behavior: Polite request (“Can you help me?” or hand-raise)
• DR procedure: DRA (reinforce the polite request with attention)
• Extinction: Withhold attention for tantrums (brief, safe ignoring)
• Reinforcement schedule: Every polite request earns 30 seconds of one-on-one attention during acquisition; fade to intermittent as the behavior grows
• Data: Tally frequency of tantrums and polite requests daily
• Decision rule: If tantrums decrease by 50% within three weeks and polite requests increase, continue. If not, reassess function.

This plan is straightforward because the function is clear, the replacement behavior is teachable, and extinction is safe and feasible.

Scenario 2: Sensory-Maintained Tapping

A young adult repeatedly taps their desk throughout the day, likely for sensory stimulation. You cannot and should not try to prevent someone from engaging in self-stimulation internally.

• Target behavior: Audible desk-tapping (more than 15 occurrences per hour)
• Replacement behavior: Using a fidget tool or textured object
• DR procedure: DRI (reinforcing an incompatible behavior that provides similar sensory input)
• Extinction: None—you’re redirecting sensory input to a socially acceptable outlet, not withholding it
• Reinforcement: Frequent positive feedback and access to preferred fidgets
• Data: Frequency of tapping vs. fidget use
• Decision rule: If fidget use increases and tapping decreases, continue. If tapping persists, explore whether the fidget is motivating enough or environmental adjustments are needed.

This plan avoids extinction because extinction of automatic reinforcement is neither feasible nor ethical.

Scenario 3: Low-Rate Behavior

A high schooler requests help 20–30 times per hour—more than any teacher can handle. The behavior itself is appropriate, but the rate is unsustainable.

• Target behavior: Raised hand or verbal request for help
• Goal: Reduce rate to 3–5 requests per hour
• DR procedure: DRL (reinforce lower rates)
• Threshold: Reinforce if help-seeking occurs no more than 5 times per hour
• Reinforcement: Positive feedback, points toward a preferred activity
• Data: Frequency of requests per class period
• Decision rule: Each week, if the student maintains ≤5 requests, the threshold stays. If they exceed it, reset. Once they’re consistently at 5 or below, the threshold can be tightened further if needed.

This plan uses DRL because you’re managing rate, not teaching a new skill or withholding the maintaining reinforcer.

Addressing Potential Side Effects

Extinction and differential reinforcement can produce temporary side effects. Knowing this in advance prevents misinterpretation and maintains plan integrity.

Extinction burst is the most common: a temporary increase in the frequency, intensity, or variability of the problem behavior when extinction first begins. The client is trying harder to earn the reinforcer they used to get. This is normal and usually subsides within days or weeks if extinction is consistent.

Resurgence is the reappearance of an older, previously extinguished behavior when a newer replacement behavior is no longer reinforced. If a child used to hit for attention and you extinguished hitting by teaching them to ask instead, and then attention becomes inconsistently available, hitting may resurface. This signals that maintenance schedules need adjustment.

Emotional distress or aggression can occur, especially during extinction. Some clients become frustrated when their previously effective behavior stops working. Have a plan to support the client, remain calm, and provide positive reinforcement immediately when the replacement behavior occurs.

Spontaneous recovery is the temporary reappearance of the problem behavior after extinction seems to have worked. It’s time-limited and usually fades quickly with consistent extinction.

Brief your team on all of these possibilities before you start. Frame them as expected, temporary, and a sign you’re on the right track—not failures.

Key Takeaways

Differential reinforcement is your front-line intervention for reducing problem behavior while teaching skills—and it aligns with ABA’s ethical foundation of using the least restrictive, most respectful approaches.

Pair differential reinforcement with extinction when you can do so safely, consistently, and with informed consent. Otherwise, rely on reinforcement-based alternatives that are equally ethical and sometimes more sustainable.

Match your DR procedure to your goal and function: teach an alternative with DRA, teach an incompatible response with DRI, reinforce absence with DRO, or manage rate with DRL. Let your functional assessment guide you.

Plan your data collection and decision rules before you start. Train your team thoroughly on extinction side effects and consistency. Monitor for fidelity and unintended effects. Small adjustments made early, based on data, prevent big problems later.

Most importantly, remember that differential reinforcement isn’t just about reducing problem behavior—it’s about building a client’s repertoire, teaching them that appropriate behavior pays off, and respecting their dignity in the process.