Translated from the Portuguese original. Read the original (Português)
Imitating means reproducing what another person just did. Technically, imitation occurs when the response comes right after the model, has the same form as it, and happens because the model appeared. Observational learning is broader: the observer also learns from what happens to the other person — whether it worked out or not. And the best news for anyone who teaches: imitating is itself a learned behavior. Once it generalizes, it becomes a powerful shortcut for learning everything else.
In behavior analysis (Cooper, Heron & Heward), a behavior is imitative when it meets three conditions: immediacy — it follows the presentation of the model within a few seconds (by convention, 3 to 5 s); formal similarity — the response and the model physically resemble each other and are in the same sensory modality; and the model as the controlling variable — it is the presentation of the model that evokes the response, not an instruction, an object, or some other contingency. Modeling is the procedure of presenting the model. Observational learning, in turn, is the larger umbrella: a change in the observer's behavior as a function of seeing what the other person does and the consequences that produced — the so-called vicarious reinforcement (or punishment). Here the change may take time to appear and does not always have the same form as what was observed.
Three conditions, not one
In everyday speech, we call almost anything that resembles what someone else did "imitation." In technical use, the criterion is narrower — and that is exactly what makes it useful for teaching and for measuring. Notice that the three conditions address different things: when the response came, how it looks, and because of what it occurred.
A vocabulary note is worth making: when the model is vocal and the repetition has point-to-point correspondence — someone says "water" and the child says "water" — verbal behavior analysis calls this echoic, not motor imitation. They are close cousins with different names. See verbal behavior and verbal operants.
Imitation, modeling, and observational learning
These three terms often travel together in conversation, but they describe different things. Explore each one:
Imitation — the behavior of the one who copies
This is the behavior itself: an immediate response, similar to the model, and evoked by it. It is what gets observed and recorded. An adult raises their arms; seconds later, the child raises her arms too. If that gesture was not in her repertoire before, the model has just taught something — without a single verbal instruction.
Modeling — the procedure used by the teacher
This is the arrangement: presenting the model on purpose so that another person learns. It can be live or on video, done by an adult, by a peer (peer modeling), or by the person themselves recorded succeeding (video self-modeling). Modeling is what the teacher does; imitation is what the learner does.
Observational learning — the umbrella
This is the broadest concept. The observer learns not only what the other person did, but what happened to them: a classmate is praised for raising her hand, and the class starts raising their hands too (vicarious reinforcement); another is scolded, and the rest avoid doing that (vicarious punishment). The change may show up hours or days later, and it is not always a copy — sometimes what is learned is precisely not to do something.
Generalized imitation: the shortcut
If every new gesture had to be taught one by one, learning would be painfully slow. What changes the game is generalized imitation: the person starts imitating new models that were never specifically trained or reinforced. Imitation stops being a handful of isolated responses and becomes a class of behavior — "doing what the other person does."
This demonstration is old and classic in behavior analysis: Baer, Peterson, and Sherman (1967) systematically reinforced the similarity between children's behavior and that of a model and observed that the children went on to imitate new models as well, without specific reinforcement for those gestures.
That is why generalized imitation is the kind of repertoire that opens doors: once it is in place, the person starts learning in situations nobody planned — in the lunch line, at the playground, watching a sibling. Gains like this, which put the learner in contact with many new contingencies at once, are what behavior analysis calls a behavioral cusp (Rosales-Ruiz & Baer, 1997).
What makes observation "stick"
Watching is not enough. Albert Bandura, in social learning theory, described four processes that must be present for observing to turn into learning. Tap the cards:
How imitation is taught
When imitation is not in the repertoire — which is common early in development and in some autistic children — it is taught directly, like any other behavior. A typical script:
- Assess first. Find out which gestures the person already makes and which ones she already imitates, to separate what is new from what already exists.
- Start with the easiest. Large movements involving an object (banging the table, shaking a rattle) tend to be simpler than fine gestures or movements the person cannot see herself making, like touching her own face.
- Ensure success with prompts. A light physical assist secures the correct response right from the start — and is then gradually withdrawn, through fading, until only the model remains.
- Reinforce. The imitative response needs to produce something worthwhile for that person; see positive reinforcement.
- Mix mastered and new. Alternating already-learned models with new ones keeps the success rate high and keeps the task from becoming frustrating.
- Test with never-trained models. This is the real test: imitating a gesture that was never reinforced shows that the class has generalized. See generalization and maintenance.
Video modeling
Instead of a live model, a short video is presented of someone performing the skill — it can be another person, a peer, or the person themselves recorded at their best moments (self-modeling). Video has practical advantages: it is always the same, it can be reviewed as many times as needed, and it allows framing exactly what matters.
What the evidence showsThe review by the National Clearinghouse on Autism Evidence and Practice (Steinbrenner et al., 2020) lists both modeling and video modeling among the evidence-based practices for autistic children, youth, and young adults. An earlier meta-analysis of single-case studies (Bellini & Akullian, 2007) had already found favorable effects of video modeling and video self-modeling on social, communication, and functional skills.
Two cautions
Imitation is a tool — and tools do not choose on their own what they are used for.
The first: imitation does not select the content. Whoever imitates copies what is available, not what is desirable. Children learn by observation both to share a toy and to yell — and adults are models all the time, especially when they do not realize they are being watched. The classic study by Bandura, Ross, and Ross (1961), with the Bobo doll, showed that children who watched an adult act aggressively toward the doll later reproduced those actions.
The second: teaching imitation is not teaching someone to maskA systematic review of studies on action imitation (Williams, Whiten & Singh, 2004) found, overall, lower performance on imitation tasks among autistic people — but that is a group average, not an individual destiny: there are autistic people who imitate very well. And the goal of teaching imitation is never to make someone "look normal" or to mask who they are. It is to widen access to learning — to offer one more path for learning whatever that person wants to learn.
Test your understanding
Quick quiz
Three questions to reinforce the concept. You see the explanation right away.
Further reading
- Cooper, J. O., Heron, T. E. & Heward, W. L. Applied Behavior Analysis. 3rd ed. Pearson, 2020 — chapter on imitation, modeling, and observational learning.
- Baer, D. M., Peterson, R. F. & Sherman, J. A. “The development of imitation by reinforcing behavioral similarity to a model”. Journal of the Experimental Analysis of Behavior, 10(5), 405–416, 1967.
- Bandura, A., Ross, D. & Ross, S. A. “Transmission of aggression through imitation of aggressive models”. Journal of Abnormal and Social Psychology, 63(3), 575–582, 1961; and Bandura, A. Social Learning Theory. Prentice Hall, 1977.
- Steinbrenner, J. R. et al. Evidence-Based Practices for Children, Youth, and Young Adults with Autism. National Clearinghouse on Autism Evidence and Practice, FPG Child Development Institute, University of North Carolina, 2020.
- Rosales-Ruiz, J. & Baer, D. M. “Behavioral cusps: a developmental and pragmatic concept for behavior analysis”. Journal of Applied Behavior Analysis, 30(3), 533–544, 1997.
- Bellini, S. & Akullian, J. “A meta-analysis of video modeling and video self-modeling interventions for children and adolescents with autism spectrum disorders”. Exceptional Children, 73(3), 264–287, 2007.
- Williams, J. H. G., Whiten, A. & Singh, T. “A systematic review of action imitation in autistic spectrum disorder”. Journal of Autism and Developmental Disorders, 34(3), 285–299, 2004.
This entry is educational and does not replace assessment, diagnosis or treatment by a qualified health professional. Every person is unique — for a specific case, seek individual guidance.