Beyond Social Skills: Building True Community for Autistic Children and Adolescents

Overview & Clinical Significance

Loneliness is not a peripheral concern in autism services — it is a central one. Epidemiological data consistently show that autistic adolescents and young adults experience rates of loneliness and social isolation that far exceed those of their neurotypical peers, and this isolation carries measurable mental health consequences. Depression and anxiety are among the most common co-occurring conditions in autism, and emerging evidence points to the absence of genuine social community as a significant contributing factor rather than merely a correlate.

RuthAnne Rehfeldt's presentation targets a gap that behavior analysts rarely name directly: ABA has historically been effective at teaching discrete social behaviors — making eye contact, asking questions, joining a group — but substantially less effective at helping individuals become genuine members of a social community. These are meaningfully different goals. Teaching a behavior and cultivating belonging require different program designs, different reinforcement systems, and a fundamentally different understanding of what success looks like.

For supervising BCBAs and program directors, this distinction has urgent clinical implications. If a learner can initiate conversations but has no reciprocal friendships, if they can follow group norms but are never sought out by peers, then the behavioral skills being trained are failing to produce the outcomes that matter most for quality of life. This presentation asks clinicians to zoom out — to consider community membership as a measurable treatment goal and to redesign programming accordingly.

The evolutionary framing is deliberate. Belonging to a social group is not a preference; it is a biological drive shaped by millions of years of selection pressure. The social motivation hypothesis in autism research has generated substantial discussion precisely because it implicates core neurological systems, not just skill deficits. Whether or not one accepts that framing, the practical reality is the same: isolated individuals experience worse outcomes, and behavior analysts must accept responsibility for whether their programming actually addresses that isolation.

This course is classified as Supervision CEU because the programming gaps Rehfeldt identifies are most often introduced — and can most effectively be corrected — at the supervisory level. BCBAs who understand the limitations of current social skills curricula are positioned to redirect their behavior technicians, modify program templates, and advocate for systemic changes within their organizations.

Background & Context

The history of social skills training in ABA is long and productive. Early work demonstrated that reinforcement-based procedures could reliably establish discrete social behaviors in children with autism who had previously shown little social engagement. Peer-mediated interventions, social stories, video modeling, and naturalistic teaching procedures expanded the toolkit considerably. Yet across decades of research, a consistent pattern emerged: skills acquired in structured contexts frequently failed to generalize to natural social environments and even more frequently failed to produce durable peer relationships.

This generalization failure is not simply a procedural problem solvable by adding more training trials in more varied settings. It reflects something more fundamental about what social skills training has traditionally targeted. When a program teaches a child to ask a peer a question, it is shaping a topography. When the goal is community membership, topography is necessary but insufficient. The social environment must also change — peers must respond, relationships must develop, shared history must accumulate.

Rehfeldt's work draws on the evolutionary science of prosocial behavior, which distinguishes between social behaviors performed for immediate reinforcement and genuinely prosocial behavior that involves taking perspective, sharing resources, and contributing to group welfare. This distinction maps onto clinical concerns directly. A learner who has been taught to share materials because sharing was reinforced on discrete trials may not engage in genuine generosity — the motivated, flexible version of sharing that makes other people want to be around someone.

The depression and anxiety data deserve particular attention. Both conditions are dramatically overrepresented in autistic adolescents and young adults, and longitudinal data suggest that social isolation mediates a significant portion of this risk. BCBAs supervising adolescent programming need to understand this pathway. Anxiety is often treated as a behavioral excess requiring reduction; depression is often treated as mood state outside ABA's scope. Rehfeldt reframes both as potential downstream consequences of programming that failed to produce community membership.

For supervisors, the background context also includes the organizational reality of most ABA programs. The majority of service delivery involves one-to-one instruction, often in home or clinic settings that structurally limit exposure to natural peer communities. Designing community has to be intentional because the typical service environment does not create it by default.

Clinical Implications

The most immediate clinical implication of Rehfeldt's framework is the need to audit current social programs for whether they are targeting community membership or merely social behaviors. These are overlapping but non-identical constructs. A useful heuristic: after twelve months of social skills programming, does the learner have friends? Not behavioral proxies for friendship — actual relationships with specific peers who seek them out, who share private jokes and history, who include them in non-programmed activities. If the answer is no, the program needs revision regardless of skill acquisition data.

One concrete implication involves the selection of social skills targets. Many programs work from commercially available social skills curricula that identify targets based on normative development or social validity ratings from neurotypical raters. Rehfeldt's framework suggests an alternative starting point: What prosocial behaviors most reliably build reciprocal relationships in the specific social contexts this learner inhabits? This requires analyzing the learner's actual social ecology — what the peer group values, how relationships are initiated and maintained in that specific culture, what contributions earn genuine social inclusion in that community.

Behavioral flexibility is the other major clinical theme. Rigid behavioral repertoires create friction in social interactions even when discrete social skills are present. An adolescent who responds to novel social situations with scripted language or inflexible behavioral patterns may execute the component behaviors of social interaction while still being perceived as odd or difficult by peers. Programming for behavioral flexibility means deliberately exposing learners to variability, reinforcing multiple different topographies for the same social function, and teaching them to read and respond to contextual cues that shift appropriate behavior.

From a supervision standpoint, BCBAs should consider establishing community membership as a program-level goal with operationally defined criteria — not just annual survey data about friendships, but active tracking of peer relationship data. Who does this learner spend unstructured time with? Who initiates contact with them? How many reciprocal relationships do they maintain across contexts? These metrics are harder to collect than correct responses on discrete trial data sheets, but they are more meaningful measures of treatment success.

Supervising BCBAs also carry responsibility for educating caregivers about this broader framing. Families are often focused on discrete skill deficits — their child won't make eye contact, won't start conversations, won't respond to questions. Helping families understand that the goal is community membership, not skill acquisition per se, shifts how they interpret progress and what they advocate for across service settings.

Ethical Considerations

Code 2.01 of the BACB Ethics Code for Behavior Analysts requires that BCBAs provide services in the best interest of clients, which includes selecting goals that are meaningful and likely to improve quality of life. Designing social programs that reliably produce skill acquisition without producing community membership arguably violates this standard if clinicians are aware of the gap and continue programming as before without modification.

Code 2.14 addresses treatment efficacy: BCBAs are obligated to use evidence-based practices and to monitor whether interventions are producing meaningful change. When social skills training consistently fails to generalize to genuine peer relationships, continued implementation without critical examination of the programming approach is ethically questionable. The ethics code does not permit clinicians to continue ineffective treatment simply because the skill data look good.

A subtler ethical issue involves the goals of social skills training itself. There is ongoing discussion in the autism self-advocacy community about whether ABA's social goals are designed to help autistic individuals thrive or to make them appear more neurotypical. Rehfeldt's community membership framework partially addresses this tension: genuine community membership, as she conceptualizes it, involves belonging to communities that value the learner as they are — not masking or camouflaging autistic characteristics. BCBAs should engage thoughtfully with this distinction. Code 1.05 on cultural humility and Code 1.07 on the dignity of clients are both relevant here.

Consent and assent deserve explicit attention. Adolescent learners should have meaningful input into what social goals are targeted and what communities they are being prepared to join. This is not merely an ethical nicety; it is functionally relevant. Motivation to engage in social behavior is substantially shaped by whether the individual values the outcomes that behavior is meant to produce. An adolescent who is indifferent to or actively dislikes the peer group their program is designed to integrate them into will be a poor candidate for the intervention regardless of skill level.

Finally, Code 2.09 on involving stakeholders in treatment planning means that the family's vision of social success must be engaged explicitly. Some families prioritize safety and comfort over social integration; some prioritize integration above all else. Both are legitimate values, and the behavior analyst's role is to inform these decisions with accurate data about outcomes, not to substitute clinical judgment for family values.

Assessment & Decision-Making

Assessing for community membership requires expanding beyond standard social skills assessments. Most commercially available tools measure social skill knowledge or discrete behavior frequency — do they know the rule, can they perform the behavior on demand. These tools do not measure whether the learner is embedded in a social community, whether peers seek them out, or whether the learner experiences their social environment as rewarding.

A clinically useful assessment framework for community membership would include at minimum: a social network analysis (who does this learner interact with, who initiates contact, how many relationships are bidirectional), an ecological analysis of the learner's primary social contexts, and a preference/motivation assessment specific to social activities and peer relationships. Standard preference assessments used in ABA focus on tangibles and activities; applying preference assessment methodology to social contexts — which peers, which activities, which types of interaction — provides crucial information for program design.

Functional assessment methodology is also directly applicable. Social isolation in autistic adolescents is not always purely a skills deficit. In some cases, avoidance of social situations is negatively reinforced — the learner escapes demands, unpredictability, or past aversive social experiences. In other cases, the learner has abundant motivation to connect but lacks effective behavior to initiate and sustain connection. These functional profiles require different treatment approaches, and a behavioral assessment that treats all social deficits as skill deficits will misallocate treatment resources.

Decision-making around social program goals should involve a structured review of generalization data at least quarterly. Are skills that are mastered in structured contexts appearing in natural environments? Are peer relationships forming? Are indicators of depression or anxiety changing as social programming is implemented? BCBAs who are supervising social programs should establish clear decision rules: if community membership metrics have not improved after a defined period, the program design must change, not just the implementation effort.

For adolescents in particular, the assessment must include the peer community itself. Programming in isolation — teaching the autistic learner skills without any intervention in the peer environment — is unlikely to produce community membership. Peer-mediated interventions have strong evidence precisely because they modify the social environment, not just the learner's behavior. Assessment should identify which peers are candidates for peer support training and what the barriers to natural peer engagement are in the specific settings where the learner spends time.

What This Means for Your Practice

Start by pulling up the social program goals for three of your current adolescent learners. For each, ask whether the goal is a behavioral skill or community membership. If all of your social goals target skill topographies without any goal tracking whether peer relationships are forming, your programs need revision.

Building community into ABA programming requires structural changes, not just new targets. Review service delivery models for adolescent clients: how much of their programming occurs in one-to-one settings, and how much occurs in natural peer contexts? For learners who spend the majority of their intervention hours in one-to-one sessions, the structural conditions for community membership are absent by design.

Consider adopting a prosocial behavior lens when selecting social skills targets. Rather than teaching scripted initiations or rote reciprocal exchanges, identify the behaviors that generate genuine positive responses from peers in this learner's community — humor, shared interest expression, collaborative play, contributions to group activities. These are harder to teach than scripted exchanges but far more likely to produce lasting relationships.

Document community membership outcomes alongside skill acquisition data. Create a simple tracking system for peer relationship indicators — number of peer-initiated interactions per week, number of reciprocal friendships maintained, caregiver report of social inclusion in non-programmed activities. Share these data with families at treatment review meetings to shift the conversation from skill counts to quality of life outcomes.

Finally, Rehfeldt's framing has supervisory implications. Share this framework with your BTs and make explicit that the goal of social programming is belonging, not compliance. BTs who understand the community membership goal will make better in-the-moment clinical decisions — they will create opportunities for genuine social exchange rather than just prompting scripted social behaviors.