Shaping is a technique in operant conditioning where you teach a new behavior by reinforcing small steps that gradually move closer to the target. Instead of waiting for the exact behavior to appear on its own (which might never happen), you reward rough versions of it first, then raise the bar little by little until the full behavior emerges. B.F. Skinner, who coined the term “operant conditioning” in 1937, developed this method as a core tool for training behaviors that are too complex to occur spontaneously.
How Successive Approximations Work
The formal name for the process is the “method of successive approximations.” The idea is straightforward: you start by reinforcing any response that remotely resembles the goal. Once that response is happening reliably, you stop reinforcing it and only reward a version that’s one step closer. You keep tightening the criteria until only the final target behavior earns reinforcement.
Two mechanisms drive this process simultaneously. The first is differential reinforcement: selectively rewarding behaviors that meet your current standard. The second is extinction: withholding reinforcement for behaviors that no longer meet the standard, which causes them to fade. Together, these two forces steer behavior in a specific direction without the learner needing instructions or a model to imitate.
Consider teaching a child to say the word “ball.” The successive approximations might look like this:
- Making any vocal sound
- Making a “b” sound
- Saying “ba”
- Saying “bal”
- Saying “ball”
At first, any vocalization earns a reward. Once the child is vocalizing regularly, only the “b” sound is reinforced. Then only “ba,” and so on. Each step builds on something the child can already do, which keeps frustration low and momentum high.
Why Shaping Works Better Than Waiting
Many behaviors are too complex or too specific to appear naturally. A rat in a laboratory won’t press a lever by chance very often, and a child who has never spoken won’t suddenly produce a full word. Shaping solves this by meeting the learner where they are, rewarding what they can already do, and gradually pulling behavior toward the goal.
Research supports this advantage in concrete terms. A study on preschoolers with autism compared shaping to a “reasonable attempts” approach, where therapists simply reinforced any effort that seemed close enough. Children mastered vocal skills more rapidly when they learned through structured shaping. In the majority of phases, mastery criteria (90% accuracy across three sessions) were reached more quickly during shaping conditions. Two of the three children in the study never achieved mastery under the reasonable-attempts condition at all, while they consistently did so with shaping.
Reinforcement Timing Matters
During the early stages of shaping, continuous reinforcement is essential. That means rewarding every single instance of the behavior that meets your current criterion. This is the fastest way to establish a new response. Once the behavior is solid, you can shift to partial reinforcement, where you reward it intermittently. Partial reinforcement makes the behavior more resistant to fading over time, so the transition from continuous to partial is what locks the behavior in.
Timing also matters within each trial. The reinforcement needs to arrive immediately after the correct response, not a few seconds later. Any delay risks accidentally reinforcing whatever the learner did between the target behavior and the reward.
Shaping in Animal Training
Animal trainers use shaping constantly, often with a clicker that marks the exact moment of the correct behavior. Teaching a dog to bow, for example, might progress through these steps:
- Dipping the head while standing
- Lowering the head halfway to the floor
- Nose close to touching the floor
- Elbows bending
- Elbows on the floor with rear in the air (the completed bow)
The click acts as a precise signal that tells the dog “that exact thing you just did is what earned the treat.” Trainers also use the environment creatively. When teaching a dog to circle an object, you might click for simply moving beside it, then toss the treat so the dog has to travel further around the object to retrieve it, naturally extending the behavior.
Shaping in Therapy and Education
Applied Behavior Analysis, the field that most systematically uses shaping with people, applies it to teach everything from motor skills to social behaviors. Teaching a child to wave goodbye, for instance, starts by reinforcing any arm movement, then lifting the arm, then moving the hand, then a full wave, and finally waving in response to “bye-bye.” Each step is only a small stretch beyond what the child was already doing successfully.
Writing skills follow the same logic. A therapist might first reinforce holding a pencil, then scribbling, then tracing a letter, then forming that letter independently. The key is that the child experiences success at every stage. No step is so large that it feels like failure, which keeps motivation intact and builds a sense of competence along the way.
Shaping vs. Chaining
Shaping is sometimes confused with chaining, another technique for teaching complex behaviors, but they work differently. Shaping refines a single behavior by gradually raising the standard for what gets reinforced. You’re always working on the same action, just expecting a better version of it. Chaining, by contrast, links a sequence of distinct steps into a routine. Each step is a separate behavior that serves as a cue for the next one.
Toothbrushing illustrates the difference well. Shaping would be used to teach the correct way to hold a toothbrush, reinforcing increasingly accurate grip over time. Chaining would then connect all the separate steps of the brushing routine: applying toothpaste, brushing each section, rinsing, and putting the toothbrush away. In practice, the two techniques often complement each other. You shape individual skills to a high standard, then chain them together into a complete sequence.
Common Mistakes in the Process
The most frequent error is requiring too large a jump between steps. If the learner was being reinforced for making a “b” sound and you suddenly withhold reinforcement until they say the full word “ball,” the behavior is likely to collapse. The gap between what they can do and what you’re now asking for is too wide, so they stop trying. Smaller steps take longer to plan but are far more reliable.
Another common pitfall is reinforcing the wrong behavior by accident. If your timing is off by even a second or two, you may end up strengthening a head turn, a foot shift, or some other irrelevant action that happened to follow the target behavior. This is why precise marking tools like clickers are so valuable in animal training and why therapists are trained to deliver reinforcement immediately.
A third mistake is lingering too long on an intermediate step. If you reinforce the same approximation for too many trials, it becomes entrenched, and the learner resists moving to the next level. The general principle is to move forward as soon as the current step is happening reliably, before it becomes a deeply established habit in its own right.