A skill that only shows up with one therapist, one set of flashcards, and one exact phrasing is not much use in daily life. Generalization is behavior change that spreads beyond the exact conditions of teaching. This lesson separates the two directions it can spread: to new stimuli and to new responses.
Why it matters
Most of what you teach is meant to travel. A greeting learned in the clinic should work at school, and a request for help should not depend on one memorized sentence. Knowing which kind of generalization you are seeing lets you measure it, plan for it, and fix it when it goes too far or not far enough.
When a scenario asks you to name what happened, the fastest route is one question: what is new, the stimulus or the response?
Key ideas
- Stimulus generalization: new stimuli, same response. After a response has been reinforced in the presence of one stimulus, the same response also occurs in the presence of other, untrained stimuli that share features with it. The learner does the same thing in a new situation.
- Response generalization: new responses, same function. After one response has been taught, the learner also emits untrained responses that serve the same function. The learner does something different that works just as well. Clear examples hold the situation constant so the new response form is the only thing that changed.
- Generalization is graded. Responding is usually strongest to the original training stimulus and weaker as stimuli become less similar to it. A plot of this pattern is a generalization gradient. A flat gradient means broad generalization; a steep one means tight stimulus control.
- Generalization and discrimination are two ends of one continuum. Responding the same way to many different stimuli is generalization. Responding differently depending on which stimulus is present is discrimination.
- Setting, person, and material generalization are kinds of stimulus generalization. When a skill shows up in a new room, with a new adult, or with new materials, the response is the same and the surrounding stimuli are new.
- Generalization is not always good. A toddler who calls every man "Daddy" is showing overgeneralization. Differential reinforcement (discrimination training) narrows it. Response generalization can also produce unwanted variations, which differential reinforcement of the desired form can reduce.
- Generalization is different from maintenance. Maintenance asks whether the behavior continues over time after teaching ends. Generalization asks whether it spreads to untrained stimuli or responses. Textbooks group all three under generalized behavior change, but each is a distinct outcome you measure separately.
- Respondent behavior generalizes too. Stimuli similar to a conditioned stimulus can also elicit the conditioned response, as when fear of one dog extends to other dogs.
Definitions
- Stimulus generalization: a response reinforced in the presence of one stimulus also occurs in the presence of untrained stimuli that share physical features with it.
- Response generalization: the learner emits untrained responses that are functionally equivalent to the trained target response.
- Generalization gradient: a graph showing how responding changes as test stimuli become more or less similar to the training stimulus.
- Setting (situation) generalization: the target behavior occurs in a setting or situation different from the one where it was taught.
- Overgeneralization: stimulus generalization that extends to stimuli where the response is not appropriate or will not be reinforced.
Examples from a session
Which generalization is it?
- Stimulus generalization at school: Amara, 6, learns to raise her hand for help with her RBT at the clinic table. The next week she raises her hand for help with her first-grade teacher. Same response, new person and setting.
- Stimulus generalization with materials: Jun, 4, learned to tact "car" with three toy cars. He now says "car" when he sees a picture of a taxi in a library book.
- Response generalization at home: Mrs. Delgado, 68, who is recovering from a stroke, practiced asking "Can you pass the salt?" during family dinners at home. At those same dinners she now also says "Salt, please" and "Could I have the salt?" Neither phrase was taught; each gets her the salt.
- Response generalization at work: Theo, 19, was taught at his job site to greet coworkers by saying "Good morning." He now also says "Hey, how's it going?" to the same coworkers, an untaught greeting that gets the same friendly replies.
- Overgeneralization on an outing: Lena, 3, now says "doggy" to cats, goats, and a horse at the petting zoo. The response is spreading to too many stimuli, so the team reinforces "doggy" only for dogs and teaches new tacts for the other animals.
Notice that the Amara and Jun examples change the antecedent while the behavior stays the same. The Mrs. Delgado and Theo examples keep the situation the same while the behavior takes new forms.
Common mistakes / exam traps
- Forgetting "untrained." If the new stimulus or response was directly taught, its occurrence is acquisition, not generalization.
- Assuming generalization is always the goal. Overgeneralization is a real problem, and discrimination training is how you fix it.
- Ignoring function in response generalization. A new response only counts if it serves the same function as the trained one.
Exam trap
Stimulus generalization vs. response generalization vs. maintenance: ask "what is new?" A new antecedent (a new room, person, or material) with the same behavior is stimulus generalization, not response generalization. A new form of behavior that does the same job is response generalization. Nothing new, just the same behavior in the same setting months later, is maintenance.
Exam trap
Response generalization vs. extinction-induced variability: both produce new, untaught forms. Check the trained response. If it is still being reinforced and the new forms appear alongside it with the same consequence, that is response generalization. If the new forms appear only after the trained response stopped producing reinforcement, that is extinction-induced variability.
Quick check
Ravi, 9, learned to put his backpack in his cubby when he arrives at his after-school program. Teaching ended in March. In June, staff notice he still puts his backpack in the same cubby every day without prompts. Which concept best describes this?