Social Skills & Peer-Mediated Supports

See what research suggests this support may help with, how confident the evidence is, and where important limitations remain.

Social ConnectionLast reviewed August 2026

What It Is

Understanding what the support actually involves.

Social skills support is a broad category rather than one standardized intervention. Programs may explicitly teach conversation, friendship, conflict resolution, dating, responding to teasing, joining activities, or other social strategies. Peer mediated interventions instead involve classmates or other peers in creating opportunities for interaction, participation, play, or communication. Programs such as PEERS combine structured social knowledge teaching, practice, role play, and support from caregivers or social coaches. The goals and methods differ substantially across programs, so evidence for one model should not automatically be generalized to every intervention described as social skills training.

Quick Answer

What should you knowbefore digging into the details?

Social skills and peer mediated supports aim to help autistic people navigate relationships, friendship, conversation, peer interaction, school participation, and other social situations. Research suggests that some structured programs, particularly PEERS and peer mediated interventions, can improve social knowledge, practiced social skills, and certain peer interaction outcomes. However, broad social skills group programs produce inconsistent results, real world generalization is less certain, and improvements on social rating scales do not automatically mean greater friendship, belonging, authenticity, or quality of life.

Best-Supported For

Social knowledge and practiced social skills in structured programs such as PEERS, plus peer engagement in some school based peer mediated interventions

How Confident Are We?

Mostly Low certainty, with results varying substantially by program, outcome, informant, and setting

Biggest Limitation

Real world friendship, long term generalization, quality of life, masking, high support autism, and minimally speaking people

Bottom Line

Some structured social programs can help autistic adolescents and young adults learn social information and practice specific interaction skills, and peer mediated approaches can improve engagement with peers in school settings. The evidence is much less consistent when broad social functioning is measured across different programs, informants, and real world environments. Improvements in knowing social rules or performing a taught behavior should not automatically be interpreted as improved friendship, belonging, wellbeing, or quality of life. Social support should help autistic people pursue relationships and participation they actually want rather than teaching them to hide autistic characteristics or perform neurotypical behavior for its own sake.

Full Evidence

What does the research showfor each outcome?

Evidence can be stronger for one goal and weaker for another. Open any rating below to see why AFA reached that conclusion.

Outcome

Social knowledge and practiced skills in structured programs

Evidence certainty: LowFinding: Likely Benefit

Structured programs such as PEERS can improve knowledge of social strategies and the ability to demonstrate some taught social skills in autistic adolescents and young adults.

Why did we rate it this way?

A 2025 systematic review and meta analysis of PEERS included 21 randomized controlled trials and 10 nonrandomized controlled trials.
The review found improvement in participants' social skills knowledge and ability to apply social skills.
Individual randomized trials in adolescents and young adults have also reported improvements in social knowledge and practiced skills.
Effects differ substantially depending on who reports the outcome, including autistic participants, caregivers, teachers, and researchers.
PEERS is one structured program and its results should not automatically be attributed to every form of social skills training.
Knowing a social rule or demonstrating it during assessment does not necessarily show that relationships or everyday participation have improved.

Outcome

Broad social functioning from social skills group programs

Evidence certainty: LowFinding: Mixed Results

Social skills group programs can improve some measured social outcomes, but the overall evidence is inconsistent and becomes less convincing when publication bias and differences between studies are considered.

Why did we rate it this way?

A recent systematic review included 65 randomized social skills group studies and 49 studies in its meta analysis.
Before adjustment, some informants reported small positive effects.
The review identified a high likelihood of publication bias.
After statistical adjustment for publication bias, the review found no overall effect across the included outcomes.
A separate 2024 meta analysis of 36 randomized trials involving 2,796 adolescents with Level 1 autism found a small positive effect on social skills.
Differences in populations, programs, control conditions, outcome measures, and study quality help explain why reviews can reach different conclusions.

Outcome

Peer engagement and interaction in school settings

Evidence certainty: LowFinding: Likely Benefit

Peer mediated interventions can improve social engagement and initiation for some autistic school age children by changing the social environment as well as supporting the autistic child.

Why did we rate it this way?

A 2026 systematic review included 24 peer mediated intervention studies involving autistic primary school children.
Most studies reported positive effects on social outcomes.
A broader review of reviews covering intellectual and developmental disabilities found positive outcomes particularly for peer engagement and initiating interaction.
Peer mediated approaches can create more natural opportunities to practice interaction with classmates rather than limiting practice to an adult directed therapy setting.
Many underlying studies use single case or multiple baseline designs rather than large randomized trials.
Only five studies in the recent primary school review included follow up, and their follow up findings were inconsistent.

Outcome

Friendship, get-togethers, and everyday social participation

Evidence certainty: LowFinding: Possible Benefit

Some structured programs increase social activities and opportunities for friendship, but evidence that social skills intervention consistently produces lasting, satisfying friendships remains limited.

Why did we rate it this way?

Randomized PEERS trials have reported increases in peer get-togethers and other measures of social participation.
A Polish adolescent randomized trial reported increases in social skills knowledge and the number of peer get-togethers, with many effects maintained at six months.
Young adult trials have reported improvement in some measures of social interaction and loneliness.
A small randomized adult PEERS study found improvements in social skills and social knowledge but only limited gains in social engagement.
Friendship quality, mutuality, belonging, authenticity, and satisfaction are measured much less often than social skill scores.
An increase in the number of interactions does not necessarily mean that the relationships are meaningful or desired.

Outcome

Social skills in autistic adolescents and young adults

Evidence certainty: LowFinding: Possible Benefit

Structured social programs can improve some social outcomes for autistic adolescents and cognitively able young adults, but effects vary between programs and populations.

Why did we rate it this way?

A 2025 PEERS meta analysis found positive effects across adolescents and young adults.
A randomized trial involving 82 Taiwanese autistic young adults found improvements across several self reported, coach reported, and observer rated outcomes.
Many effects in that trial were maintained at three and six month follow up.
A 2025 meta analysis of employment related soft skills interventions in transition age autistic people found positive effects on social adjustment, performance, and social skills.
Most adult and young adult studies involve people without significant intellectual disability and who can participate in verbally based group programs.
Evidence for middle aged and older autistic adults remains very limited.

Outcome

Social anxiety and emotional wellbeing

Evidence certainty: LowFinding: Possible Benefit

Some social programs may reduce social anxiety or emotional distress for certain autistic adolescents and young adults, but mental health benefits are not consistent enough to treat social skills training as a primary mental health treatment.

Why did we rate it this way?

A 2024 meta analysis of adolescent randomized trials found a small improvement in anxious and depressive symptoms alongside social skill outcomes.
Several young adult PEERS trials have reported reductions in social anxiety.
Social confidence may improve when people gain strategies and opportunities to connect with peers.
Other research shows that pressure to perform socially typical behavior can itself contribute to stress and camouflaging.
Social anxiety may require direct psychological treatment rather than assuming social skills instruction alone will address it.
Few studies distinguish anxiety caused by uncertainty about social situations from distress caused by rejection, stigma, bullying, or inaccessible environments.

Outcome

Real world generalization and long term maintenance

Evidence certainty: Very LowFinding: Effect Uncertain

It remains uncertain how consistently skills learned in structured social programs transfer into everyday relationships and remain useful over the long term.

Why did we rate it this way?

Some PEERS trials report maintenance of selected outcomes for three to six months.
The recent review of peer mediated school interventions found that only five studies included follow up assessments.
Follow up results in those peer mediated studies were inconsistent.
Many studies measure social behavior shortly after treatment rather than years later.
Teacher, caregiver, clinician, and autistic self report ratings frequently differ.
Real life friendship, social belonging, relationship satisfaction, and long term participation are undermeasured.

Outcome

Masking, camouflaging, and pressure to appear nonautistic

Evidence certainty: InsufficientFinding: Potential Concern

Research does not establish that social skills programs as a category cause camouflaging, but goals centered on performing neurotypical behavior can create a legitimate concern because autistic people consistently report mental health costs associated with sustained masking and social camouflaging.

Why did we rate it this way?

Direct trials rarely measure masking or camouflaging as a potential adverse outcome of social skills intervention.
A systematic review of camouflaging research identifies substantial use of strategies intended to assimilate with social expectations.
Large qualitative studies of autistic adults describe exhaustion, anxiety, reduced authenticity, and other mental health costs associated with sustained camouflaging.
Autistic adults report that social pressure and lack of acceptance can motivate masking.
Some people also describe strategic social adaptation as useful in certain circumstances, so the issue is not simply whether a learned social strategy exists.
Programs should distinguish optional tools for achieving a person's own social goals from demands to suppress harmless autistic traits or appear neurotypical.

Outcome

Long term quality of life and belonging

Evidence certainty: InsufficientFinding: Effect Uncertain

There is not enough evidence to determine whether traditional social skills programs consistently improve long term quality of life, belonging, authentic relationships, or overall wellbeing.

Why did we rate it this way?

Most studies prioritize social rating scales, skill knowledge, or observable interaction behaviors.
Quality of life and belonging are rarely primary outcomes.
The quality and reciprocity of relationships are less frequently measured than frequency of interaction.
Autistic people's own definitions of successful social connection are not consistently used as primary outcomes.
Mental health consequences of masking and pressure to conform are rarely assessed alongside intervention benefits.
Long term participatory research is needed to determine whether learned strategies produce relationships and social lives that autistic people themselves value.

Who Was Actually Studied?

Applicability matters.

Age and life stage

The largest evidence base involves school age children and adolescents. PEERS and related programs also have evidence in young adults. Research becomes much thinner in middle and older adulthood.

Intellectual disability

People with intellectual disability are poorly represented in many verbally based social skills group trials. Peer mediated intervention literature includes broader developmental disability populations, but autism specific subgroup evidence remains limited.

High support / profound autism

High support and profound autism are substantially underrepresented in traditional social skills group research. Some peer mediated studies include children with greater support needs, but subgroup outcomes are rarely reported using modern definitions.

Minimal or no spoken language

Minimally speaking and nonspeaking autistic people are poorly represented in conventional verbally based social skills programs. Peer mediated and AAC supported approaches may be relevant, but the evidence is much smaller.

Adults

Young adult research exists, particularly for PEERS and transition focused programs. Evidence for autistic adults across middle age and older adulthood is limited.

Settings

Programs have been studied in schools, clinics, universities, community groups, workplaces, telehealth programs, and structured peer environments. Generalization from a training setting into everyday relationships remains an important evidence gap.

What We Do Not Know

Important gaps remain.

High support and profound autism are substantially underrepresented.
People with intellectual disability are poorly represented in many social skills group trials.
Minimally speaking and nonspeaking autistic people are rarely included in verbally based programs.
Middle aged and older autistic adults are substantially underrepresented.
Long term friendship quality and relationship satisfaction are rarely measured.
Belonging, authenticity, quality of life, and autistic defined social success are rarely primary outcomes.
Masking and camouflaging are rarely monitored as potential adverse outcomes of intervention.
Teacher, parent, clinician, researcher, and autistic self report outcomes often disagree.
Publication bias may inflate apparent benefits in the social skills group literature.
Generalization from structured sessions to everyday social environments remains uncertain.
Research rarely compares teaching autistic people new social strategies with changing the social environment or educating nonautistic peers.
Peer mediated approaches need more randomized trials and longer follow up.
Different cultures may have different social expectations, and intervention effects may not transfer equally across cultural settings.
Programs often measure conformity to expected social behavior more readily than mutual understanding between autistic and nonautistic people.

Potential Concerns

How a support is implemented matters.

A social support should begin with whether the autistic person actually wants help with a particular social goal.
Teaching an optional social strategy is different from requiring a person to hide autistic characteristics.
Eye contact, facial expression, body language, tone, or other neurotypical social behaviors should not automatically be treated as required markers of successful interaction.
Harmless stimming and other self regulatory behavior should not be suppressed merely to make someone appear more socially typical.
Social success should not be defined only by how well the autistic person adapts to nonautistic expectations.
Peers, schools, workplaces, families, and communities also have responsibility for creating accessible and accepting social environments.
Programs should support refusal, boundaries, recognizing unsafe relationships, and self advocacy alongside friendship or conversation skills.
Increasing interaction frequency is not automatically beneficial if the interactions are unwanted, exhausting, unsafe, or inauthentic.
Bullying or social rejection should not be framed solely as a failure of the autistic person's social skills.
Social strategies can be useful tools when freely chosen, but sustained masking to avoid stigma can carry significant psychological costs.
Intervention goals should prioritize connection, safety, confidence, participation, and personally meaningful relationships rather than appearing less autistic.

Autistic Perspectives & Masking

Research evidence and lived experience answer different questions.

Autistic people describe wanting friendship, connection, acceptance, and practical strategies for navigating social situations, while also describing the heavy cost of feeling required to perform nonautistic behavior. Camouflaging research shows that some autistic people consciously learn and use social strategies because they can make certain situations easier, but sustained masking can also be exhausting and damaging to wellbeing. These findings support offering social strategies as optional tools while also changing environments and expectations so that authentic autistic communication is accepted.

Themes Reported

Many autistic people genuinely want friendship, relationships, community, and social connection.

Learning explicit social information can be useful when social expectations are otherwise unclear or unpredictable.

Autistic people differ greatly in how much social interaction they want and what kinds of relationships they value.

Camouflaging can sometimes help people navigate school, work, healthcare, or social situations.

Sustained camouflaging is also associated with exhaustion, anxiety, reduced authenticity, and poorer mental health.

Autistic adults describe social pressure and fear of negative reactions as important reasons for masking.

Some autistic people deliberately choose to unmask and seek relationships where autistic communication is accepted.

The goal should not be to eliminate every learned social strategy, but to ensure that its use is voluntary and serves the autistic person's own goals.

Successful social support should include acceptance and adaptation from nonautistic communication partners as well as tools for autistic people.

Lived experience is shown separately and is not mathematically combined with scientific evidence certainty ratings.

Sources Reviewed

See the research behind the review.

Systematic Review and Meta Analysis2024

A Systematic Review of Implicit Versus Explicit Social Skills Group Programs in Different Settings for School-Aged Autistic Children and Adolescents

Sixty five studies were included and 49 entered the meta analysis. The review found a high possibility of publication bias and, after statistical adjustment for that bias, no overall effect across the assessed outcomes. Small effects remained for some informants.

View Source →

Systematic Review and Meta Analysis2024

Social Skills Interventions for Adolescents With Level 1 Autism Spectrum Disorder: A Systematic Review With Meta Analysis of Randomized Controlled Trials

Thirty six randomized trials involving 2,796 participants were included. Social skills interventions produced a small positive pooled effect on social skills and smaller improvements in behavioral and anxious or depressive outcomes.

View Source →

Systematic Review and Meta Analysis2025

Impact of Social Knowledge and Skills Training Based on UCLA PEERS on Social Communication and Interaction Skills of Adolescents or Young Adults With Autism

Twenty one randomized and 10 nonrandomized controlled trials were included. PEERS improved social knowledge, ability to apply social skills, and emotional intelligence, while effects differed by informant and cultural setting.

View Source →

Systematic Review2026

Peer-Mediated Social Skills Interventions in Primary School Settings for Autistic Children

Twenty four school based peer mediated intervention studies were included. Most reported positive social outcomes, but only five included follow up assessment and longer term findings were inconsistent.

View Source →

Systematic Review of Reviews2023

Peer-Mediated Interventions for Students With Intellectual and Developmental Disabilities: A Systematic Review of Reviews

Forty three reviews representing 357 unique studies and more than 4,000 participants were examined. Peer mediated interventions most consistently supported peer engagement and initiation of social interaction.

View Source →

Randomized Controlled Trial2023

Effectiveness, Durability, and Clinical Correlates of the PEERS Social Skills Intervention in Young Adults With Autism Spectrum Disorder

Eighty two cognitively able autistic young adults were randomized to PEERS or treatment as usual. Several social and social anxiety outcomes improved, with many effects maintained at three and six months.

View Source →

Randomized Controlled Trial2023

Effects of Social Skills Training for Adolescents on the Autism Spectrum: A Randomized Controlled Trial of the Polish Adaptation of PEERS

Twenty nine adolescents were randomized to PEERS or a waitlist. Social skills, social knowledge, and peer get togethers improved, and most measured effects were maintained over six months.

View Source →

Randomized Controlled Trial2023

Fostering Friendship and Dating Skills Among Adults on the Autism Spectrum: A Randomized Controlled Trial of PEERS for Young Adults

A small randomized trial of 15 young adults found improvement in social skills, cognition, and knowledge but limited gains in social engagement. Effects were maintained at six months.

View Source →

Participatory Lived Experience Study2021

Autistic Adults' Experiences of Camouflaging and Its Perceived Impact on Mental Health

Qualitative responses from 277 autistic adults described both perceived usefulness and substantial mental health costs of social camouflaging, particularly when masking was frequent or sustained.

View Source →

Systematic Review2025

A Systematic Review of Social Camouflaging in Autistic Adults and Youth: Implications and Theory

The review synthesized research on autistic social camouflaging and concluded that camouflaging may be associated with important mental health risks and social pressures.

View Source →

Systematic Review and Qualitative Meta Synthesis2025

Consolidating a Framework of Autistic Camouflaging Strategies: An Integrative Systematic Review

Twenty eight studies involving more than 2,600 autistic people older than 16 were synthesized to describe different camouflaging strategies and the social contexts that shape their use.

View Source →

Mixed Methods Lived Experience Study2026

Unmasking: A Mixed-Methods Study of Autistic College Students' Efforts to Stop Masking and Camouflaging

This recent study examined autistic college students' deliberate efforts to reduce masking and reclaim more authentic self presentation, adding emerging evidence on what autistic people themselves want from social environments.

View Source →

Evidence Atlas

Evidence should help inform decisions, not make the decision for you.

Return to Evidence Atlas Read Our Methodology Contact AFA

Important: The Support Evidence Atlas is an educational decision support resource and does not provide individualized medical advice or prescribe a particular therapy, intervention, service, or support. Decisions should reflect the individual's needs, preferences, goals, abilities, environment, circumstances, and appropriate professional guidance when relevant.

Autism Foundation of America

What should you knowbefore digging into the details?

Important gaps remain.

High support and profound autism are substantially underrepresented.

People with intellectual disability are poorly represented in many social skills group trials.

Minimally speaking and nonspeaking autistic people are rarely included in verbally based programs.

Middle aged and older autistic adults are substantially underrepresented.

Long term friendship quality and relationship satisfaction are rarely measured.

Belonging, authenticity, quality of life, and autistic defined social success are rarely primary outcomes.

Masking and camouflaging are rarely monitored as potential adverse outcomes of intervention.

Teacher, parent, clinician, researcher, and autistic self report outcomes often disagree.

Publication bias may inflate apparent benefits in the social skills group literature.

Generalization from structured sessions to everyday social environments remains uncertain.

Research rarely compares teaching autistic people new social strategies with changing the social environment or educating nonautistic peers.

Peer mediated approaches need more randomized trials and longer follow up.

Different cultures may have different social expectations, and intervention effects may not transfer equally across cultural settings.

Programs often measure conformity to expected social behavior more readily than mutual understanding between autistic and nonautistic people.

How a support is implemented matters.

A social support should begin with whether the autistic person actually wants help with a particular social goal.

Teaching an optional social strategy is different from requiring a person to hide autistic characteristics.

Eye contact, facial expression, body language, tone, or other neurotypical social behaviors should not automatically be treated as required markers of successful interaction.

Harmless stimming and other self regulatory behavior should not be suppressed merely to make someone appear more socially typical.

Social success should not be defined only by how well the autistic person adapts to nonautistic expectations.

Peers, schools, workplaces, families, and communities also have responsibility for creating accessible and accepting social environments.

Programs should support refusal, boundaries, recognizing unsafe relationships, and self advocacy alongside friendship or conversation skills.

Increasing interaction frequency is not automatically beneficial if the interactions are unwanted, exhausting, unsafe, or inauthentic.

Bullying or social rejection should not be framed solely as a failure of the autistic person's social skills.

Social strategies can be useful tools when freely chosen, but sustained masking to avoid stigma can carry significant psychological costs.

Intervention goals should prioritize connection, safety, confidence, participation, and personally meaningful relationships rather than appearing less autistic.

Research evidence and lived experience answer different questions.