Generalization: Why Skills Must Work at Grandma's House Too

A child who asks for juice clearly at the therapy table — and nowhere else — has learned something real, but not yet the thing that matters most. The gap between "works in session" and "works at Grandma's house" has a name, and closing it is one of the most important jobs in ABA.

What Generalization Means in ABA

Generalization is the ability to use a learned skill across new people, new places, new materials, and new situations — not just under the exact conditions where it was first taught. In applied behavior analysis, a skill is not considered truly learned until it shows up where life actually needs it: at home, at school, in the backseat of the car, and at a grandparent's kitchen counter.

Behavior analysts usually talk about three related ideas:

• Stimulus generalization — the skill works with different people, settings, and materials. A child who learned to wash hands at the clinic sink can also do it at the sink at home, which has a different faucet, different soap, and a step stool.
• Response generalization — the child can produce useful variations of the skill, not one memorized version. "Help me," "can you open this," and handing an adult a stuck container are all forms of the same skill: recruiting help.
• Maintenance — the skill keeps working over time, after active teaching has ended.

A quality ABA program treats all three as part of the teaching plan from day one, not as a bonus that might happen later.

Why Skills Get "Stuck" in the Therapy Room

Behavior is context-sensitive for everyone. Adults speak differently in a job interview than at a family dinner; most of us can recall a phone number at home that vanishes when someone asks for it in a noisy store. Skills attach themselves to the cues that were present when we learned them.

Many autistic learners learn with real precision — and precision cuts both ways. It can make a child wonderfully reliable with routines they know well. It can also mean that incidental details of the teaching situation quietly become part of the skill: the specific chair, the specific therapist's voice, the specific blue cup. Change the cup, and from the child's perspective the situation is genuinely different.

None of this is a flaw in the child. It is a predictable feature of how learning works, which is exactly why behavior analysts plan for generalization instead of hoping for it. Research summaries of ABA consistently describe generalization programming as a core component of well-designed intervention, not an optional add-on.

The Learn, Practice, Use-Anywhere, Keep Model

One practical way to think about a skill's full journey is in four stages.

Stage	What is happening	Example: asking for help
Learn	The skill is taught in a structured setting with clear cues, planned prompts, and frequent reinforcement	Practicing "help, please" with a favorite puzzle at the table with a familiar RBT
Practice	Same skill, deliberately varied — different materials, people, rooms, and phrasings, with prompts fading	Asking for help with zippers, snack containers, and a tablet, from two different technicians
Use anywhere	The skill appears in new places with new people, without rehearsal	Asking Grandma for help with a jacket at her house on a Sunday afternoon
Keep	The skill is checked over time after teaching ends, and everyday life sustains it	Still asking for help six months later, because asking reliably works

A skill can stall at any stage. A child who has strong table performance but nothing at home is stuck between Learn and Practice. A child who used a skill everywhere last spring but rarely uses it now has a Keep problem, and the fix looks different for each.

How therapists build the bridge

The techniques are concrete, and you can watch for them in your child's sessions:

• Multiple exemplars. Teaching with many examples — different cups, different books, different phrasings — so the skill attaches to the concept, not one object.
• Varying instructors and settings. Rotating which team members run a program, and moving practice from the table to the floor, the hallway, the playground.
• Natural environment teaching. Folding practice into play and daily routines the child already cares about, so the teaching situation looks like real life from the start. We cover this approach in depth in our post on natural environment teaching.
• Shifting to natural reinforcement. Early teaching may use frequent, planned rewards. Over time, the plan hands the job to natural consequences — asking for help gets help; greeting a neighbor gets a warm hello back. Skills that everyday life rewards are the skills that last.

If you want a picture of how a session is structured around these ideas, our overview of what happens in an ABA session walks through it step by step.

What This Looks Like in a Real Program

Consider a five-year-old learning to wash hands independently. In the Learn stage, her RBT breaks the routine into steps and practices at the same clinic sink each day. Once the steps are solid, the team deliberately varies things: the other bathroom, a different soap pump, a paper-towel dispenser instead of a hand towel. Her parents get coached on the same steps and the same level of help to offer at home — no more, no less. Within a few weeks, the team is checking the skill at home during an in-home session, and her mom reports it happened unprompted at a restaurant.

Or consider a nine-year-old using a speech-generating device. If the device lives on the therapy table, his requesting skills live there too. So the plan makes the device travel — kitchen, car, park — and involves his older sister as a practice partner, because siblings are people too, and a skill that only works with adults is only half generalized.

Notice what is not on the target list in either example: eye contact for its own sake, or suppressing harmless stimming. Generalization applies to skills that widen a child's access to the world — communication, self-help, play, flexibility — not to making a child appear a certain way. And a generalized skill still belongs to the child: a child who can ask to leave a loud party, and does, is showing the skill working exactly as intended. A child's no counts, at the clinic and at Grandma's house alike.

Common Misunderstandings

"If she really learned it, it should transfer automatically." Occasionally it does, and that is a happy day. But automatic transfer is the exception, for autistic and non-autistic learners alike. Treating generalization as a planned teaching phase — with its own data — is what separates careful programs from hopeful ones.

"The skill disappeared, so therapy is failing." More often, the cues changed. A skill that shows up in one setting and not another is information, not defeat. It tells the team exactly which conditions to add to practice next.

"More drilling at the table will fix it." Usually the opposite. If the table is the problem, more identical repetitions at the table deepen the attachment to the table. The fix is varied practice, not more of the same practice.

"Generalization means performing on demand everywhere." Generalization is about capability and access — the child can use the skill wherever it serves them. It is not a promise of compliance in every setting, and it should never be measured that way.

The Home–Center Handoff

Generalization is a big part of why many families blend settings. Center-based sessions offer structure, planned variety, and peers — a strong environment for the Learn and Practice stages. In-home sessions put skills to work amid real couches, real siblings, and real bedtime routines — the natural home of the Use-anywhere stage.

The handoff between them is deliberate. The BCBA decides which skills are ready to travel, the team aligns on prompts and expectations so the child is not getting two versions of the same lesson, and caregivers get coached directly — because parents and grandparents are the people who will be there when the therapy team is not. Maintenance gets its own attention too: mastered skills are re-checked on a schedule, and anything that has faded gets a brief, focused refresher rather than a full re-teach.

Questions Worth Asking Your Child's Team

• How do you program for generalization from the start, rather than at the end?
• Which of my child's current skills are ready for home practice, and exactly how should I support them?
• How will you vary people, places, and materials as a skill gets stronger?
• How do you check maintenance on mastered skills, and how often?
• What will you do if a skill works in one setting but not another?

A team that welcomes these questions — and answers with specifics from your child's data — is a team thinking beyond the therapy table.

Skills That Travel Are the Point

At Inspire Center for Autism, generalization is built into how programs are structured: BCBA-designed plans delivered 1:1 by RBTs, with parent and caregiver coaching in every program, across both our Mesa center and in-home sessions throughout the East Valley. The whole design assumes that the real test of a skill happens somewhere we are not.

If you would like to see what generalized, real-world skills look like across communication, self-regulation, and daily living, browse examples from our program areas.