Shaping in ABA: Building Behavior One Small Step at a Time

Shaping teaches new behavior by reinforcing small steps toward a goal. Learn how BCBAs set steps, use micro-shaping, and fix stalled progress.

Key takeaway

Shaping is a way to teach a new behavior by reinforcing small steps that get closer and closer to a goal. Behavior analysts call these steps "successive approximations." Each one is a little nearer to the final skill than the one before it.

12 days of PFA & SBT
Watch the full CEU recording

12 days of PFA & SBT

Matt Harrington · 3 CEU · 156 min
Watch on openceu.com →
Free learning pathStep 1 of 3 in Master PFA and SBTSee the whole path →

Experts who cover this

18 recordings on openceu cover shaping.

Shaping is a way to teach a new behavior by reinforcing small steps that get closer and closer to a goal. Behavior analysts call these steps "successive approximations." Each one is a little nearer to the final skill than the one before it.

Shaping helps when a learner cannot do the full skill yet. Instead of waiting for the perfect response, you reward the best try the learner can make today. Then you slowly ask for a bit more. Clinicians use it for talking, eating new foods, sitting at a table, getting a haircut, and much more. They also use it with parents, staff, and even their own health habits.

How shaping works#

At its heart, shaping has three parts. You pick a clear end goal. You find a first step the learner can already hit. Then you raise the bar in small steps as the learner succeeds. The "criterion" is the bar. It is the rule for which responses earn reinforcement right now.

Megan Patterson uses a water goal to show the idea. Someone drinks 20 ounces a day but wants to reach 80. That is too big a jump. So the plan starts at 40 ounces, four days a week, for two weeks. Next comes 60 ounces, five days a week. Only then does the full goal kick in.

So we shape it, the same way we shape any behavior. We pick a small first target we can hit, then raise the criterion in steps. From the talk — Megan Patterson

She also notes that none of her steps say "every single day." A goal with no room for a sick kid or a bad day at work is a goal people drop fast. Room for real life is built into each step. That same logic shows up across nearly every talk on this page. Small steps keep the learner winning. The end goal keeps everyone pointed the right way.

Why so many clinicians love shaping#

Matt Harrington calls shaping his favorite topic in all of behavior analysis. He sees it as the thread that ties together the practical functional assessment (PFA) and skill-based treatment (SBT) process. Shaping shows up in rapport building during caregiver interviews. It shows up when you shape a learner's tolerance of "no" after teaching them to ask for things. It shows up when you move from simple to more complex demands.

I really freaking love shaping. I think shaping as a concept is one of the coolest and most undervalued teaching skills that we have in our field. From the talk — Matt Harrington

He also describes a "progressive" style of discrete trial teaching. Here you slowly add little bits of difficulty, fade out prompts, and add new goals. The aim is for the learner to always be succeeding. In his words, this is shaping in a broad sense, even if it is not used exactly the way some research articles define it.

Michael Carrero calls shaping "a core principle of learning." In another talk, Harrington groups shaping with pairing and transfer of stimulus control. He calls these tools "an ethical analyst's best friend." These clinicians see shaping less as one procedure and more as a way of thinking about all teaching.

Start with the goal and the entry point#

Good shaping starts before the first step. You need to know two things. Where is the learner now? And exactly what do you want them to do in the end? The end goal is called the "terminal behavior." The starting step is the "entry point."

We want to define the terminal behavior precisely. We want to analyze the implementation skill. We want to find the entry point. From the talk — Matt Harrington

Harrington adds two more steps. Build a sequence that gives lots of early reinforcement. Then teach each step with behavioral skills training (BST), or a version of it that fits the setting. He points out that behavior plans should not say "client will do the final goal at 90%" with nothing in between. Plans need prerequisite goals and shaping steps.

Skipping the entry point is a common way programs fail. His toilet training example makes the point clearly.

Remember that shaping, the way shaping works is small steps at the beginning. So if you don't assess where they are now and you place the kid in underwear before they can identify what wet and dry is, well, you are in for a beating when it comes to the success of that program. From the talk — Matt Harrington

The lesson is simple. If the first step is too hard, the learner never gets the early wins that keep the program moving.

Is shaping an art or a science?#

Many clinicians have heard the phrase "shaping is an art more than a science." Harrington admits there is some truth to it. Nobody pulls out a spreadsheet before every high five. Some staff seem to read a client so well that they always know when to raise or lower the bar.

But he argues that shaping rests on a clear decision rule. Every time a response happens, you ask one question.

at its core, all shaping is, is looking at a behavior and deciding based on your past behaviors, should we reinforce this or not From the talk — Matt Harrington

He gives an example. A child is working toward sitting at a table for five minutes. Their last three sits were 10, 12, and 15 seconds. Then they sit for 14 seconds. That is not their best, but it is near the top. Should you reinforce it? Probably, because the overall trend is going up.

Why turn the art into a science at all? Harrington says it makes training others "10 times easier." The gifted staff member who reads clients by feel is hard to copy. A clear rule is easy to teach. That matters for every supervisor training new staff.

The percentile schedule: math for shaping#

The percentile schedule is a formula that tells you when a response is good enough to reinforce. Harrington calls it the tool that makes shaping objective.

the percentile schedule is the equation to objectify the shaping process. From the talk — Matt Harrington

The formula uses a few simple variables. "M" is how many recent responses you look at, such as the last five. "W" is the density of reinforcement, meaning how often you want the learner to earn a reward. "K" is the result. It is the value the next response must beat to earn reinforcement.

The big benefit is that the bar moves on its own. If the learner does well, the bar goes up. If responses drop, the bar drops too. Harrington shows an example where the criterion was three and then moved back to two after a few weaker responses. He calls it hyper-individualized, since every variable fits the specific client.

He is careful to say the spreadsheet is not meant for live sessions. The goal is to understand why the bar changes. Once you know which variables matter, you can adjust your shaping by feel with much more precision.

Reinforcement density and pacing#

How often should a learner earn reinforcement during shaping? Harrington ties the answer to how severe the behavior is and how far along the learner is. An "extinction burst" is a short spike in problem behavior when a reward suddenly stops. For some learners, that spike is not safe.

If you have severe behavior that you can't handle an extinction burst with, and they're on a cab one and a cab two, maybe even up to cab three, you need to have a high density of reinforcement so that it is a slow and steady process. From the talk — Matt Harrington

CAB steps are the numbered skill steps in SBT. Early steps get rich, frequent reinforcement. Later, around CAB five and six, something natural happens. The client moves faster. Clinicians add more branches and more challenges. Reinforcement gets thinner because the client can handle it.

Sean Yocum and Michael Carrero describe the same pattern in their talk on assent. You start small, honor the learner's words, and use a high density of reinforcement. Over time, tolerance grows and participation becomes reinforcing on its own. They call it "just good behavioral design," not compliance.

Micro-shaping: when the steps need to be smaller#

Micro-shaping means breaking a shaping step into even tinier pieces. It is common language in the PFA and SBT world. When a learner gets stuck between two steps, the first question is often whether the steps are small enough.

In the PFA world, we love a good micro shape when you're like, oh, I can't get from cab two to cab three. And then we're like, oh, did you try to micro shape it? From the talk — Matt Harrington

But how small is "micro"? Harrington jokes that the answer is often just "go micro micro." He offers a more exact answer through the percentile schedule. You can set the K value so the chance of reinforcement is close to 100%.

This is, I think, the closest thing we have to micro shaping is, of course, using this math, this K value in the equation to lift the probability of reinforcement occurring to near 100%. From the talk — Matt Harrington

The point of near-certain success is to avoid an extinction burst. He shares a case about a child who ran around the back of a church. Putting that running on extinction would have caused screaming. So the team shaped time in the pew instead. The steps started at seconds and grew over many trials to 20 or more minutes.

Micro-shaping in the classroom#

Micro-shaping is not only for clinic programs. Nikki, a school-based clinician, uses it after a student has a big blowup. She notes a common belief that shaping "in its finest form is quick." Sometimes it cannot be.

Sometimes we have to micro shape is what we call it in skill based treatment, which is we have to take tiny steps. From the talk — Nikki

Her example is a student who has just flipped a desk and is hiding at the back of the room. Asking him to sit back down right away will not work.

You've just flipped the desk. I cannot ask you. And now you're hiding in the back of the room. I cannot ask you to come back and sit at your seat. That is not going to work. From the talk — Nikki

Instead, she tries to move a little closer. If he backs up, she stops. She tells him she will not touch him and just wants to make sure he is safe. Most of the time, he lets her come a bit closer. In another case, she starts drawing a Pokemon badly under the desk. The student comes over to correct her. Now he is talking, moving toward her, and giving assent. From there, she might write numbers on the drawing and ease back into math. Each step is tiny, but each one builds on the last.

What to change when progress stalls#

Shaping programs often hit a plateau. When that happens, the shaping plan itself is the first thing to look at. Harrington lists two main variables you can adjust.

The first is step size. Say your steps are 1, 5, 15, and 30 seconds. You could shrink them to 1, 3, 5, 7, 10, and 11 seconds. The second is the mastery rule. You could move on after one to three perfect data points, or wait for five. Either change affects how fast the learner climbs.

The shaping intervention, though, is is a variable that I would consistently change. From the talk — Matt Harrington

He adds that risk sets the pace. If the behavior could include head banging or self biting, he would change the plan very fast. If the risk is lower, there is more room to wait.

Eric Soltysik used a shaping protocol for a haircut goal. The family had needed to restrain the child many times before. When progress slowed, the team sat down with the family. Did they jump too far? Was the child going through reinforcement strain, meaning the reward became too thin for the work asked? Patterson gives a similar rule for adults. If you miss a step, stay there a little longer. That is not failure. It is how shaping works.

Stepping back is part of the process#

Sometimes a learner tells you, through their behavior, that a step was too big. Harrington shares a case about teaching a boy to tolerate delays before a break. After about a month and a half, they were close to a one-minute delay. He got excited and pushed straight to two minutes.

he clearly told us that he wasn't ready for that jump and we took it all the way back... but we realized that we weren't going back to zero, we made a bunch of forward progress, now we're taking a couple steps back and that is absolutely okay. From the talk — Matt Harrington

The team dropped back to 30 seconds. It felt like "a kick in the gut." But they kept going. In the end, the boy could handle a five-minute delay with five minutes of work and five minutes of break. The team also taught him to say when he was not ready to work. Harrington calls it one of his proudest moments, and he credits his whole team.

The bigger lesson is that a step back does not erase progress. A good shaping plan expects small setbacks. The percentile schedule even builds this in, since the bar drops when responding drops.

Shaping tolerance and relationships#

Shaping is often used to build tolerance. That can mean tolerating a delay, a demand, or simply another person being nearby. Dr. Megan DeLeon describes one of her first clients, back in 2011. A free operant preference assessment is a way of watching what a child chooses on their own. It showed that the girl loved the trampoline. But if an adult joined her, she stopped right away.

So the team built an 18-step shaping program to grow her comfort with an adult nearby. The steps came in three groups, each with its own criteria.

The adult would do the trampoline for five seconds and then say, okay, you know, I'm done and walk away because we were trying to show her it's not that bad. From the talk — Dr. Megan DeLeon

DeLeon explains why leaving early matters. Many people stay with a child until the child pushes them away. That teaches the child that pushing is the only way to get space. Leaving first shows respect and changes the child's learning history. Nicole Stewart takes a similar first step with her clients. She starts with plain toleration and engagement. Can the child tolerate the activity, talk about it, and do it together with an adult?

Shaping instead of escape extinction#

Escape extinction means not letting a learner get out of a task when they show problem behavior. In feeding, it can look like not removing the spoon. Many clinicians now look for ways to avoid it. Shaping is a main alternative.

Dr. Holly Gover worked with Dr. Hanley and colleagues on a food selectivity process that relied heavily on shaping. Their top goal was to avoid escape extinction in any form. That ruled out both keeping the spoon in place and "waiting them out."

So we, like many other shaping papers, task analyzed the act of picking up a piece of food, chewing it and swallowing it. From the talk — Dr. Holly Gover

A task analysis breaks a skill into small steps. Like other desensitization programs, the steps moved in stages. They went from touching the food, to smelling it, to the lips and tongue, and finally to chewing and swallowing. An early version even had a dice game, but it had too many bells and whistles. The team stripped it back.

One girl struggled to swallow. So the team had her chew one tiny noodle 20 times. It simply dissolved. She was shocked and proud, and soon said she ate noodles all the time.

It just took some like creative shaping, task analyzing to just like mentally help her out. From the talk — Dr. Holly Gover

Harrington adds a safety note from his talk on care collaboration. A child who refuses vegetables is very different from a child who gags or struggles to swallow. For those learners, get speech-language pathologists (SLPs) involved right away.

Teaching, not forcing#

Harrington calls this "shaping's dirty little secret." People expect a slow shaping plan to lose the race to a hard extinction plan. In his clinical experience, the opposite is true.

you get so much done faster with shaping in my experience... we want to be the clinicians who teach, we don't want to be the clinicians who force. From the talk — Matt Harrington

He is honest about the limits of this claim. To his knowledge, no study compares treatment length between an extinction approach and a shaping approach. His view comes from practice. He also asks a values question. Extinction can create new behavior through "extinction-induced variability," a burst of new kinds of responses. But is that in line with trauma-informed care?

In another talk, Harrington stresses picking the right problem to solve. Say a learner will not sit for table work. If the issue is motivation, a preference assessment may help. If the learner lacks the skill of waiting for reinforcement, that needs "a totally different shaping process." He also lists shaping as one of several ways to fade prompts during functional communication training (FCT). Matching the tool to the real problem keeps shaping from being used in the wrong place.

Shaping communication#

Shaping is a natural fit for language. Noah McAmis shares a moment from a session with a child who had emerging language. The child and an RBT (a registered behavior technician who works directly with clients) were bouncing on a yoga ball. The RBT paused, waited, and modeled the word "bounce."

The child gave only a rough approximation of the word at first. Within a few minutes, the RBT had shaped that attempt into a full mand, a request, for "bounce." The plan did not change. The same target and data were used. What changed was that someone noticed the moment and used it.

Carrero and Yocum describe shaping communication for a learner whose first "no" was a push. If you never shape that skill, the learner will keep pushing everyone away.

we started with pushing and then through a shaping process. We built functional skills that led to an increased tolerance with delayed reinforcers From the talk — Michael Carrero

The push became "leave me alone," then "I'm not ready." Later, the team adapted again. Now "I'm not ready" could be met with "I hear you, but the appointment is in 30 minutes." Harrington also credits SLPs with a real gift for shaping natural language. He likes building their shaping into ABA plans.

Shaping the form of a behavior, not erasing it#

Shaping can change how a behavior looks without removing it. Behavior analysts call the look or form of a behavior its "topography." Carrero uses hand flapping and other large movements as an example.

Because that's the aspect of shaping, right? So, like, let's say I engaged in super high frequency, like, gross motor movement, right? You don't need me to stop doing the movement, you need me to stop doing it unsafely. From the talk — Michael Carrero

He is firm that "stop hand flapping" should never be the goal. Removing a part of who someone is would be unethical. But if a movement blocks safety or access to places the person wants to go, shaping it can help. The goal is their benefit, not ours.

He and Yocum also point to culture. What counts as socially acceptable depends on where you are. A behavior seen as fine in one country may be judged differently in another. So any plan to shape a behavior's form needs to be individualized. It must also be checked for social validity, meaning it matters to the person and their community.

Shaping when ADHD is in the picture#

Nicole Stewart's talk focuses on shaping reinforcement for learners with ADHD. Her key point is that these learners show more day-to-day variability in response effort. That means how hard a task feels to do can change from day to day.

The classic idea of shaping is that once a learner meets a step, you keep moving up. Stewart says that rule needs to bend.

For people with ADHD. you have to keep that a little bit flexible. From the talk — Nicole Stewart

She lists real reasons a learner might do less today. Maybe they did not sleep well. Maybe they fought with a parent, a sibling, or a friend. On those days, the same task takes more effort. So she reinforces effort and engagement, not just hitting the last best number. She also starts most shaping plans with simple toleration and engagement before asking for skill demands. Her full course, Shaping Behavior Change when ADHD is in the Picture, goes deeper into these strategies.

Shaping caregiver goals#

Shaping works with adults, too. Kendrell "Abby" Richmond says parents often get one training session and a packet. A week later, the packet is forgotten. So she shapes caregiver goals instead. She might ask whether a caregiver can practice a skill twice a day instead of every hour.

Can we do it twice a day? From the talk — Kendrell "Abby" Richmond

After a few weeks at twice a day, the goal might move to four times a day. Step by step, it climbs toward the ideal frequency. Richmond adds that strong rapport makes caregivers more open to feedback along the way.

Harrington makes the same case in his caregiver training talk. He separates a "right now goal" from a "terminal goal." He breaks big goals into small, doable steps so caregivers get early wins. He compares this to the branching skill steps in SBT. Soltysik adds another piece. In his haircut case, he explained to the family why he chose a shaping protocol. When progress stalled, that shared understanding made it easier to problem solve together.

Shaping staff skills#

Mellanie Page brings shaping into staff training. She asks whether new RBTs are thrown in too deep. In their eagerness to share, trainers often teach concepts that new staff do not need yet.

shaping behavior isn't just for kids or clients. We're conditioning skills in our RBTs as well. From the talk — Mellanie Page

Her advice mirrors client programming. Start where the learner is. In most cases, that is very early in the learning process. Then create learning opportunities that move the staff member toward the target level, one step at a time.

During her talk, she asked attendees to rate their own training from one to five. A five meant they had fully planned shaping into their training. Most answers were twos and threes. None were fives. She sees that gap as a chance to revisit training materials with fresh eyes. Ask the same question you would ask for any learner: what is the next small step? Patterson's water example shows the same tool works on your own habits. Shaping is a teaching method, not just a client intervention.

Where experts disagree: should earlier steps still be reinforced?#

The clearest debate on this page is about what happens to earlier, easier responses once a learner moves up a step.

The classic view ties shaping closely to extinction. Carrero puts it plainly.

Think shaping. If I'm utilizing shaping, what do we have to do? Clearly put something on extinction. Successful approximations only get you so far until the next successful approximation is the new response and all other responses are placed on extinction. From the talk — Michael Carrero

In this view, once a learner can do the new step, the old step stops earning reinforcement. That is what moves behavior forward. Even so, Carrero softens it. He calls it a sliding scale. If a child usually says "ball" but can only say "bah" in one moment, he will take it.

Stewart pushes back more directly, especially for learners with ADHD.

unlike shaping that where I feel like they say you're not supposed to reward the previous approximations, everybody's best looks different every single day. From the talk — Nicole Stewart

She gives a parenting example. Even after a full meltdown, she still looks for something to praise, like calming down and coming back. Harrington's percentile schedule sits between these views. It uses a rule, but the rule lowers the bar when responding drops.

A second open question involves feeding. Gover's team aimed to avoid escape extinction completely. But in one study, shaping with pairing worked for only one of two children. The other child needed nonremoval of the spoon (Demchuk et al., 2026). So it is still unclear when shaping alone is enough.

What the research says#

Research supports shaping across many skills and settings. One study showed that shaping can grow simple requests into more complex ones. In FCT, low-effort requests work best at first. But pushing for longer requests too soon can bring problem behavior back. A careful shaping process built complex requests in four children without that return of problem behavior (Ghaemmaghami et al., 2018).

Shaping can also work without prompts or tangible rewards. In a telehealth study, parents in India shaped social eye contact in two preschool children with autism. They used preferred social play as the reward. Both children learned to make eye contact during different activities, and parents rated the approach highly (Strömberg et al., 2025).

Feeding research shows both promise and limits. Pairing alone did not help two children with autism and feeding problems. Adding shaping helped one child accept four new foods. The other child still needed nonremoval of the spoon (Demchuk et al., 2026).

Finally, computer models of teaching echo what clinicians describe. A 2025 study in the journal PLOS Computational Biology modeled teaching with a "teacher" program guiding a "student" program, and found that near-optimal shaping balances reinforcement and extinction carefully. The best approach tracked the learner's rate of learning and moved between easier and harder tasks as needed. That closely matches the idea of a bar that moves both up and down.

FAQ#

What is the difference between shaping and prompt fading? Shaping reinforces closer and closer versions of a behavior. Prompt fading slowly removes help, like a gesture or model. Shaping can be one way to fade prompts, but they are not the same thing.

What is micro-shaping? Micro-shaping means breaking a shaping step into much smaller steps. Clinicians use it when a learner gets stuck between two steps. The goal is to keep success, and reinforcement, close to 100%.

What should I do if a learner fails a shaping step? Stay at that step longer, or go back to an easier one. You can also make the steps smaller or change how many successes are needed to move on. A step back does not erase progress.

Turn this topic into a CEU

You just studied this. Now get credit for it.

Watch 12 days of PFA & SBT with Matt Harrington and earn 3 free BCBA CEUs. Detailed certificate, delivered the moment you finish.

Watch and earn the CEU →Free account · No card · detailed cert
Free learning pathStep 1 of 3 in Master PFA and SBTSee the whole path →

Want the primary literature? Read the Shaping in ABA research roundup on our sister site, Behaviorist Book Club.