Social Skills and Autism: What Helps, and What to Question

Key takeaways

Autism support levels describe support needed, not severity of the person — and they're assigned separately for social communication and restricted behaviors.¹
Asperger's, PDD-NOS and Childhood Disintegrative Disorder were removed from the DSM in 2013. You'll still hear them; they aren't current diagnoses.¹
Social difficulty runs in both directions. Research on the double empathy problem describes autistic and non-autistic people as each finding the other hard to read.²
Autistic people communicate effectively with each other — which reframes the difficulty as a mismatch rather than a deficit.³
Masking has a cost. Autistic adults describe camouflaging as effortful and exhausting, and report doing it under pressure to fit in.⁴
Ask what the goal is before starting any program. "Fewer stims" is a different goal from "can ask for help."
The child's own preferences count. A child who prefers one friend to a group has a social preference, not a social deficit.

What the autism support levels actually mean

DSM-5 describes three levels of support need. They are not a classification of types of autism, and they are not a measure of a person's ability.¹

Level	Description	What it means practically
Level 1	Requiring support	Difficulty initiating interaction; may need help with organization and planning
Level 2	Requiring substantial support	More pronounced communication differences; support needed across many settings
Level 3	Requiring very substantial support	Significant differences in verbal and nonverbal communication; extensive daily support

Levels describe support needed, not the person. DSM-5 assigns them separately for social communication and for restricted, repetitive behaviors, and they change with setting and with how much capacity someone has left.¹

They aren't fixed. Support needs change with setting, demand and how much capacity a person has left. The same child may need Level 1 support at home and Level 3 support in a crowded cafeteria at the end of a school week.

They're assigned twice. DSM-5 rates social communication and restricted, repetitive behaviors separately. A person can be Level 1 for one and Level 3 for the other.¹

They say nothing about intelligence, and nothing about what a person feels or understands. A child requiring very substantial support may have no cognitive impairment at all.

Asperger's and PDD-NOS: why you may still hear them

Neither is a current diagnosis. When DSM-5 was published in 2013, Asperger's Disorder, PDD-NOS and Childhood Disintegrative Disorder were consolidated into a single diagnosis: Autism Spectrum Disorder.¹

Why you still encounter them: people diagnosed before 2013 often keep the terminology, and some — particularly adults diagnosed with Asperger's — identify with it and prefer it. That's their choice and worth respecting.

But if a clinician offers one of these as a new diagnosis today, ask about it. The current framework is autism spectrum disorder with specified support levels. Our page on atypical autism covers the older terminology in more depth.

Why social interaction can be harder — and it goes both ways

The standard framing is that autistic children have difficulty reading social cues. That's part of it, and it's incomplete.

Research on the double empathy problem describes the difficulty as bidirectional — autistic and non-autistic people each find the other hard to read, because they are working from different social expectations rather than one group lacking a skill.² And a study of information transfer found that autistic-to-autistic chains passed information as effectively as non-autistic chains, while mixed chains did less well.³

Why this matters for what you do next. If the difficulty is a mismatch rather than a deficit, then the intervention isn't only "teach the child to communicate differently." It's also: find peers they communicate well with, and help the adults around them adjust.

What genuinely is harder for many autistic children:

Processing spoken language quickly enough to respond in real time
Managing sensory input while also managing conversation — see sensory over-responsivity
Unwritten rules that nobody states
Shifting attention between topics — see monotropism
Predicting what happens next in an unstructured setting

Those are real, and they're worth support. "Doesn't make eye contact" and "prefers one friend to a group" are not on that list.

The goal question: skills, or masking?

This is the section to read before choosing any program.

Masking, or camouflaging, means suppressing natural behavior to appear more typical — forcing eye contact, hiding stims, scripting responses, performing interest. Autistic adults describe it as effortful and exhausting, and report doing it largely because of pressure to fit in and to avoid negative treatment.⁴ It is measurable, and it is largely invisible to the people around them.⁵ See autistic burnout for where sustained masking leads.

Builds a genuine capability	Teaches concealment
How to ask for help	How to look like you don't need help
How to say no, or leave	How to tolerate something distressing quietly
How to join a game you want to join	How to sit still during a game you hate
Using AAC to communicate	Speaking aloud when AAC works better
Recognizing your own overload	Hiding overload until you get home
Finding people you click with	Fitting in with people you don't

The point of all of this is the adult the child becomes, and what they will need is an environment that accommodates them rather than a lifetime of concealment.

The test: does this skill give the child more control over their life, or does it make them easier for other people to be around? Both can be legitimate goals — but they're different goals, and a program should be able to say which it's pursuing.

Questions worth asking any provider: What are the goals, and who chose them? What happens if my child doesn't want to do something? Do you work on stimming, and if so why? How do you know it's working — and what would tell you it isn't?

Social stories are short, personalized narratives describing a situation and what typically happens in it. They're widely used, low-cost, and easy to make.

The evidence is mixed.⁷ They are most defensible as a way to make an unfamiliar situation predictable — describing it in advance — rather than as a way to teach social behavior directly.

What makes them work better: written for the individual child, describing what will happen rather than instructing how to behave, read in advance rather than during, and used for specific situations rather than general conduct.

What makes them worse: a story that tells a child how they should feel, or that frames their natural response as wrong.

Where ABA fits, and what the evidence says

Applied Behavior Analysis is the most commonly funded autism intervention in the US, and the most contested.

What the evidence shows, honestly. The largest meta-analysis of early childhood autism interventions found that when effects were restricted to those immune to selection, detection and placebo-by-proxy bias, no intervention type showed statistically significant effects on any outcome. The same review found that adverse events were inadequately monitored across the literature.⁶

What this means practically: ABA isn't a settled answer, and a provider presenting it as one is overstating. Our page on behavioral interventionists covers the credentials, the evidence and the debate in full, including the questions to ask before starting.

How to support friendships without forcing them

Shared interest beats structured practice. A child who finds one other child who cares about the same thing has a friendship. A child who completes twelve turn-taking sessions has completed twelve sessions.

Interest-based groups rather than social skills groups — coding club, Lego, animals, a specific game
Parallel activity, where children do something alongside each other without needing constant interaction
Shorter, earlier, smaller. One friend for an hour beats a party
A structured activity to hang the interaction on, rather than open-ended free play
Autistic peers. Many autistic children find each other easier company, and the research on autistic-to-autistic communication supports why³

What to be careful about: a child who prefers solitude some of the time is not necessarily lonely. Ask them. Loneliness and low sociability are different things, and only one needs addressing.

Visual supports and AAC

Visual supports — schedules, picture cards, written instructions — reduce reliance on real-time spoken processing and make expectations explicit. Low cost, low risk, widely useful.

AAC (augmentative and alternative communication) covers everything from picture boards to speech-generating apps.

The most important thing to know about AAC: introducing it does not delay or replace speech. That worry is common and the evidence doesn't support it. Communication access comes first; speech may or may not follow, and either outcome is acceptable.

Access matters more than the specific system. A child who can communicate is a child who can ask, refuse, and explain — which is worth more than any social skills curriculum. Our page on gestalt language processing covers a pattern often mistaken for a lack of language.

When a speech-language pathologist helps

A speech-language pathologist works on far more than articulation: comprehension, conversational structure, AAC assessment and setup, and the pragmatic side of language.

Worth asking for if your child has difficulty being understood, wants to communicate more than they currently can, needs AAC, or has feeding or swallowing difficulties alongside. Find a speech-language pathologist.

Play-based approaches

Play is where most children learn to negotiate, take turns, and read each other — and play-based approaches work with that rather than against it. They tend to be better tolerated than table-top instruction, because they follow the child's interest rather than a curriculum.

One caution: play-based doesn't automatically mean child-led. Ask whether the child chooses the activity, or whether the activity is chosen and made to look like play.

Social skills groups: what to ask before joining

Groups can genuinely help — practice with peers, in a supported setting, with someone to scaffold. Ask before enrolling:

Who sets the goals? The best answer includes the child
Are the other children autistic? A group of autistic peers works differently from one autistic child among non-autistic peers
What happens if my child wants to leave?
Is eye contact a target? If yes, ask why
How do you handle stimming?
How do you measure success?

A good group builds confidence. A poor one teaches a child that they're the problem in the room.

What to ask school for

A quiet space available without asking each time
Interest-based lunch clubs rather than unstructured cafeteria time
A peer buddy who genuinely shares an interest
Advance warning of changes, substitutes and assemblies
Not being made to participate in group work as the price of inclusion
Adults who understand that not speaking isn't defiance

These belong in an IEP or 504 plan, and our resources for teachers include material to share.

What parents can do at home

Follow the interest. The thing your child talks about at length is your best route to connection, not a distraction from it.

Let them communicate their way. Scripting, echolalia, AAC, gestures and typing are all communication.

Protect recovery time. A child who masks all day at school needs the evening to be easy. What looks like withdrawal at 4pm is usually depletion — see living with sensory processing differences.

Don't make eye contact the goal. Many autistic people find it uncomfortable and listen better without it.

Model rather than instruct. Narrating your own social reasoning — "I'm going to say no to that, because I'm tired" — teaches more than a lesson.

Signs a program isn't right for your child

Signs worth raising

Your child dreads it, consistently, over weeks
Goals focus on appearance — eye contact, quiet hands, sitting still
You aren't allowed to observe
Distress is treated as an obstacle rather than information
Progress is measured only in compliance
Nobody can tell you what would make them stop
Your child is masking more, and more exhausted at home
Stimming is targeted for reduction with no explanation of why

Any one of these is worth raising. Several together is worth leaving.

Frequently Asked Questions

Is Asperger's Syndrome still a diagnosis?

No. Asperger's, PDD-NOS and Childhood Disintegrative Disorder were consolidated into Autism Spectrum Disorder when DSM-5 was published in 2013. People diagnosed before then often keep the terminology, and some prefer it — but it isn't a current diagnosis, and a clinician offering it as a new one today is worth questioning.

What do autism Levels 1, 2 and 3 mean?

They describe the amount of support a person needs, not their intelligence or their worth. Level 1 requires support, Level 2 substantial support, Level 3 very substantial support. They're assigned separately for social communication and for restricted or repetitive behaviors, and they aren't fixed — support needs change with setting and with how much capacity someone has left.

How can I help my autistic child develop social skills?

Start with what your child wants to be able to do. Interest-based groups usually work better than social skills groups, communication access matters more than any curriculum, and shared-interest friendships form more readily than assigned ones. Ask any program what its goals are and who chose them.

What is the double empathy problem?

Research describing social difficulty between autistic and non-autistic people as bidirectional — each finds the other hard to read, rather than one group lacking a skill. Studies of information transfer have found autistic-to-autistic communication is as effective as non-autistic-to-non-autistic communication, while mixed pairs do less well.

Is social skills training harmful?

It depends entirely on the goals. Training that builds capability — how to ask for help, how to say no, how to join something you want to join — is useful. Training that teaches concealment, like forcing eye contact or suppressing stims, teaches masking, which autistic adults report as effortful and exhausting. Ask which one a program is doing.

Do social stories work?

The evidence is mixed. They are widely used, low-cost and easy to make, and they are most defensible as a way to make an unfamiliar situation predictable in advance rather than as a way to teach social behavior directly. They work better when personalized, descriptive rather than instructional, and used for specific situations.

Should my child do ABA for social skills?

It's the most commonly funded autism intervention in the US and the most contested. The largest meta-analysis found that when effects were restricted to those immune to selection, detection and placebo-by-proxy bias, no intervention type showed statistically significant effects on any outcome — and that adverse events were inadequately monitored.

Will using AAC stop my child from talking?

No. Introducing AAC does not delay or replace speech. Communication access comes first, and speech may or may not follow — either outcome is acceptable.

My child prefers to play alone. Is that a problem?

Not necessarily. A preference for solitude some of the time is not the same as loneliness, and only one of those needs addressing. Ask your child.

Sources

American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5). Arlington, VA: American Psychiatric Publishing; 2013. DSM-5 consolidated Autistic Disorder, Asperger's Disorder, Childhood Disintegrative Disorder and PDD-NOS into a single diagnosis of Autism Spectrum Disorder, and specifies severity as a level of support need rated separately for social communication and for restricted, repetitive behaviors.
Milton DEM. On the ontological status of autism: the ‘double empathy problem’. Disability & Society. 2012;27(6):883–887. doi:10.1080/09687599.2012.710008
Crompton CJ, Ropar D, Evans-Williams CV, Flynn EG, Fletcher-Watson S. Autistic peer-to-peer information transfer is highly effective. Autism. 2020;24(7):1704–1712. doi:10.1177/1362361320919286
Cage E, Troxell-Whitman Z. Understanding the reasons, contexts and costs of camouflaging for autistic adults. Journal of Autism and Developmental Disorders. 2019;49(5):1899–1911. doi:10.1007/s10803-018-03878-x
Hull L, Mandy W, Lai M-C, et al. Development and validation of the Camouflaging Autistic Traits Questionnaire (CAT-Q). Journal of Autism and Developmental Disorders. 2019;49(3):819–833. doi:10.1007/s10803-018-3792-6
Sandbank M, Bottema-Beutel K, Crowley LaPoint S, et al. Autism intervention meta-analysis of early childhood studies (Project AIM): updated systematic review and secondary analysis. BMJ. 2023;383:e076733. doi:10.1136/bmj-2023-076733
Chen T, Yang W, Wang Q. Effects of social stories intervention for children and adolescents with autism spectrum disorders. Medicine. 2020;99(38):e22018. doi:10.1097/md.0000000000022018 — a systematic review protocol, which records that evidence on the effectiveness of social stories is mixed.

Medical disclaimer. This page is for general educational purposes and does not constitute medical advice, diagnosis, or treatment.