Social skills training (SST) is a structured approach to learning and practicing interpersonal behaviors. A program typically identifies a meaningful interaction goal, demonstrates a possible response, gives the participant opportunities to rehearse it, provides specific feedback, and supports practice in everyday settings. Common targets include initiating an exchange, listening and responding, asking for clarification, negotiating a disagreement, expressing a preference, joining a group activity, and recognizing when an interaction requires a different strategy.
The scientific question is whether this training changes the interactions that matter to the person, in the environments where those interactions occur, and for long enough to improve daily life. The evidence supports learning some rehearsed behaviors and provides context-specific evidence for several structured programs. Effects on friendship quality, workplace relationships, independent functioning, subjective well-being, and enduring participation are more variable. An effective evaluation therefore distinguishes performance in a training session from performance in everyday relationships Spence, 2003; Kurtz and Mueser, 2008.
What social skills training means in psychology
Social skills training is a family of behavioral, educational, and sometimes cognitive-behavioral interventions organized around acquiring, refining, or using social behaviors. It is not a single manual, a universally standardized treatment, or a psychiatric diagnosis. Social skills are actions and response patterns; social competence concerns whether a person can pursue meaningful goals and maintain mutually workable interactions in particular contexts. Confidence reflects expectations about one's ability, while social anxiety concerns fear and avoidance associated with possible evaluation. These domains can overlap without being interchangeable.
A person may know how to ask a follow-up question yet be unable to do so when anxious; another may perform well in a role-play but have no welcoming social opportunities; a third may communicate effectively through direct, explicit language while being judged by an observer who assumes indirect hints or sustained eye contact are necessary. Training must specify which outcome is being targeted and whose judgment defines success. A norm that serves one culture, profession, or relationship can be unhelpful elsewhere.
The term SST is especially prominent in psychosocial rehabilitation for psychosis, autism-related social learning programs, child and adolescent interventions, and professional communication education. The populations, outcomes, and comparison treatments in these traditions differ. Consequently, a result obtained with adults receiving community mental health services cannot be generalized automatically to healthy adults who want to improve conversation skills, or to children in classrooms Turner et al., 2018; Storebø et al., 2019.
How the training model developed
Behavioral social-skills approaches grew from social learning, applied behavior analysis, and cognitive-behavioral work. Researchers investigated how observation, practice, feedback, consequences, emotional arousal, and the interpretation of social information affect the performance of interpersonal behavior. Rather than merely telling a participant to communicate better, an intervention identifies observable actions that can be practiced and evaluated.
An influential account by Susan Spence described instructions, modeling, behavioral rehearsal, feedback, and reinforcement as central procedures, often combined with social perception and interpersonal problem solving. It also emphasized that competing behaviors, emotion regulation, and environmental contingencies can prevent newly learned responses from appearing outside treatment Spence, 2003. Early methodological reviews already warned that success during instruction could be mistaken for generalization and that outcome measures must have social validity Gresham and Lemanek, 1983.
Contemporary work broadens the unit of analysis from the individual response to reciprocal interactions. An utterance acquires its function through timing, relationship, social position, language, setting, and what the other person does next. This makes participant-chosen goals, communication accessibility, peer involvement, contextual assessment, and outcomes beyond standardized performance tasks central to the design of training.
The mechanisms: acquisition, performance, and opportunity
Skill acquisition and response flexibility
Instruction provides an understandable representation of a behavior. Modeling shows several ways it might look. Rehearsal makes it possible to practice without the full stakes of an unfamiliar interaction. Feedback identifies what happened and offers one or two changes for the next attempt. Repetition under varied conditions can help a person retrieve and adapt a response rather than recite a fixed script.
Learning a useful phrase is only one component of competence. Social interaction is contingent: the partner may agree, misunderstand, change the subject, refuse, or communicate nonverbally. Training that rehearses several plausible responses supports flexibility better than a program that rewards a single prescribed conversational formula. The broader educational evidence supports structured opportunities for practice and feedback, although the relative value of each ingredient remains difficult to isolate Reith-Hall and Montgomery, 2023.
Emotional and cognitive interference
A behavior can be learned but difficult to perform under threat, scrutiny, distraction, sensory overload, fatigue, or high conflict. Some programs therefore combine interpersonal rehearsal with coping strategies, deliberate pacing, attention to the partner's response, problem solving, and ways to request clarification. A participant who already communicates effectively when relaxed may need interventions directed at anxiety, environmental support, or social opportunity rather than additional lessons in basic conversational mechanics. SST should not substitute for assessment and treatment of a condition that actually explains the difficulty.
Reinforcement and social opportunity
Successful practice is easier to maintain when peers, coworkers, caregivers, or conversation partners respond in ways that make the trained behavior useful. A learner may practice entering a discussion correctly yet face repeated exclusion, racism, bullying, or inaccessible communication rules. These environments restrict participation regardless of technical competence. Good programs therefore examine the surrounding interaction system, adapt tasks and opportunities, and avoid placing every responsibility for successful communication on the learner.
Transfer from practice to lived interaction
Generalization occurs when a trained behavior appears in a setting, with people, or in circumstances beyond the original exercise. Maintenance means that change persists after structured teaching is reduced or withdrawn. Both are distinct from immediately repeating a demonstrated skill. A 2024 systematic review of social-training studies involving autistic participants specifically examined differences among trained tasks, assessment materials, environments, levels of functioning, and durability. Its heterogeneous findings reinforce the importance of defining transfer rather than assuming it Nejati et al., 2024.
The principal methods used in social skills training
Behavioral instruction and task analysis
The facilitator describes the social situation, the participant's goal, and the features of one or more workable responses. A vague aim such as being more likable is replaced with a behavior that is observable and relevant: for example, request clarification when an instruction is ambiguous, or state a preference before agreeing to a plan. Complex actions may be divided into components, such as checking whether an invitation is welcome, identifying an opening, making a request, and responding respectfully to the answer.
Task analysis helps with teachability and measurement, but excessive atomization can fragment a conversation into artificial moves. The learner should understand the purpose of a behavior and be able to choose whether it fits the interaction. A script can be a scaffold, while conversational competence ultimately requires judging when to depart from it.
Modeling and video demonstration
A therapist, teacher, peer, or video demonstrates different versions of the target behavior. A useful model reveals both words and contingencies: how a person changes phrasing when the first request is misunderstood, acknowledges a boundary, or closes a conversation when the other participant is busy. Video demonstrations can make sequences repeatable and accessible to learners who benefit from reviewing examples at their own pace.
Evidence about modeling is population- and method-specific. A meta-analysis of single-case studies found favorable results for peer-mediated and video-modeling approaches in samples of autistic children, but single-case effect-size estimates and limited study contexts cannot establish that video is the best format for every learner Wang, Cui, and Parrila, 2011. The mechanism is plausible; the choice of medium should follow personal needs, accessibility, and comparative evidence.
Role-play and behavioral rehearsal
In a rehearsal, participants practice a situation that resembles a real interpersonal challenge. The scenario might involve joining an unfamiliar conversation, requesting accommodation, responding to a misunderstanding, declining an unwanted favor, or clarifying a disagreement. Rehearsal works best when it includes the possibility that the other person reacts unexpectedly. A staged exercise that always rewards the same response can teach compliance with the exercise instead of interactional judgment.
Role-play may occur individually with a facilitator, with peers, or with trained simulated partners. The task can vary gradually in difficulty, and the learner can stop, repeat, or experiment with alternatives. Because rehearsals reduce immediate consequences, they can create conditions for deliberate practice. Their value ultimately depends on subsequent transfer, not on whether the participant performs the scenario exactly as the instructor prefers Spence, 2003.
Feedback, reinforcement, and self-observation
Feedback should be timely, specific, understandable, and connected to the person's chosen objective. Saying that an interaction was awkward gives little usable information. Identifying that the request was clear but the next step remained ambiguous allows the participant to try a concrete revision. Video or audio review can be useful when consent, privacy, and cultural interpretation are addressed. Self-observation can reveal what felt difficult, what actually happened, and whether the participant's predictions matched the interaction.
Encouragement and reinforcement may support practice, but rewarding obedience to a narrow style of eye contact, facial expression, or politeness risks shaping conformity rather than agency. The facilitator should attend to communication access, listener behavior, emotional safety, and mutual respect. Feedback from people directly affected by the interaction can help identify whether the behavior solves an authentic problem.
Problem solving and perspective-taking exercises
Some programs teach people to identify the objective of an interaction, imagine several plausible explanations for another person's response, generate options, estimate consequences, act, and review the result. The objective is flexible judgment under uncertainty. Perspective-taking tasks should recognize that guesses about someone else's feelings can be wrong and that direct questions may be more accurate than mind-reading. Multiple interpretations, accessible clarification, and respect for disagreement are preferable to fixed assumptions about facial expressions or social cues.
Between-session practice and environmental supports
Homework, community practice, family involvement, peer support, and workplace adaptations are attempts to connect training with actual life. A useful practice task might be to ask for clarification in a class, arrange an activity with a consenting peer, or use a prepared boundary statement in a low-stakes exchange. The program should then examine what occurred: how the partner responded, which barriers remained, and whether the chosen action was useful. Repeated opportunities, realistic contexts, and supportive partners may matter as much as rehearsal intensity.
How programs are delivered
Individual training
One-to-one sessions permit careful task selection, pacing, accessibility adjustments, and practice of personally sensitive situations. They may be appropriate when a group is overstimulating, unsafe, socially inaccessible, or irrelevant to the person's aims. Individual sessions also make it easier to work on anxiety, communication supports, or highly specific daily-life targets. Their limitation is that the interaction with one facilitator may fail to represent peers, groups, or naturally occurring relationships.
Group programs
Groups provide multiple interaction partners and opportunities to negotiate turn-taking, disagreement, cooperation, and role changes. Peers may offer perspectives that a single instructor cannot. Some groups include parallel parent, caregiver, or peer sessions to support practice outside the meeting. Groups also introduce variables that require active safeguarding: status hierarchies, social exclusion, disclosure pressure, unequal participation, and conflicts between participants. Attendance or enthusiasm alone is not evidence of functional benefit.
Peer-mediated, community, and workplace approaches
Training can occur with peers in a classroom, volunteer group, employment setting, or community activity. Natural settings offer opportunities to practice the actual task and receive feedback from relevant partners. They also require consent and support from the environment. For example, a trainee can practice asking questions in meetings, while colleagues simultaneously learn to allow adequate processing time, explain expectations clearly, and make space for different forms of participation.
Simulation and digital delivery
Simulated patients, recorded scenarios, online role-play, and virtual agents can increase access to practice. A systematic review of 18 studies in health professional education found weak indications that standardized patients were superior to some alternative simulations for certain learning measures, but none of the included comparisons evaluated patient outcomes. The findings therefore support cautious use of simulation for learning rather than a claim of proven superiority in everyday communication Rutherford-Hemming, Herrington, and Ngo, 2024.
A digital dialogue system may provide repetitions and immediate prompts; its resemblance to authentic interaction is limited by the quality of the scenario, the evaluation method, and the absence of reciprocal human relationships. Effectiveness, confidentiality, access, and generalization require separate evaluation. The presence of an interactive technology is not itself evidence of a therapeutic outcome.
What the evidence shows across populations
Psychosis and schizophrenia: gains are outcome-dependent
SST has a substantial research history in psychosocial rehabilitation. A 2008 meta-analysis of 22 controlled studies involving 1,521 participants reported a large mean effect on content-mastery tests (d = 1.20), moderate effects on performance-based social and daily-living measures (d = 0.52) and community functioning (d = 0.52), and smaller effects on general symptoms and relapse. This proximal-to-distal pattern demonstrates why a training examination and a long-term life outcome should not be treated as interchangeable Kurtz and Mueser, 2008.
A later meta-analysis of 27 randomized trials involving 1,437 participants found small advantages for negative symptoms in psychosis compared with treatment as usual and active comparison conditions, commonly around g = 0.2–0.3, with findings on social outcomes that varied across comparisons. The authors called for better effectiveness studies and implementation research Turner et al., 2018. These results address defined clinical populations and structured interventions rather than social confidence in the general population.
A Cochrane review of 13 randomized trials involving 975 people with schizophrenia reported some favorable findings compared with standard care, but rated the evidence for important outcomes very low certainty. Comparisons with discussion-group controls did not show the same advantages; most included studies were conducted in China, limiting the transfer of conclusions to other health systems Almerie et al., 2015. Together, the reviews suggest that specific functions can improve while certainty depends heavily on comparator, outcome, study quality, and setting. SST in this context belongs within individualized, professionally supported care rather than as a substitute for treatment of psychosis.
Autism: practical social learning and the limits of normative targets
Autistic people vary in communication profile, sensory needs, language processing, interests, and preferences. Social differences are reciprocal: an interaction may fail because communication styles are mismatched or because the environment is unaccommodating. A useful intervention starts with the autistic person's goals, such as forming chosen friendships, navigating inaccessible expectations, protecting boundaries, or communicating at work. Success need not require performing a nonautistic style of gaze, gesture, or small talk.
A Cochrane synthesis of five randomized studies with 196 participants aged 6–21 found some evidence for improvements in overall social competence and friendship quality, while evidence for several other outcomes was weak or absent. Studies were small, mostly US-based, and concentrated on younger participants; the review highlighted bias and restricted generalizability Reichow, Steiner, and Volkmar, 2012.
A small randomized study of the PEERS program with 22 autistic young adults reported improvements in social-skills knowledge, engagement and assessed skills, with many gains maintained at short follow-up; the modest sample limits broad conclusions Laugeson et al., 2015. A broader 2024 review including 52 experiments examined transfer across environments and measures and found that age, intervention format, intensity, and context were associated with heterogeneous transfer patterns Nejati et al., 2024.
The UK National Institute for Health and Care Excellence advises considering individual or group social learning programs for autistic adults with identified social-interaction difficulties, with modeling, feedback, discussion, explicit guidance, and adaptations to individual needs. This is a conditional, person-centered recommendation, not a mandate that every autistic adult undergo social training NICE, CG142. Autistic communication differences as a topic have their own scientific context, explored separately in Autistic Communication Differences in Adults: Language and Mutual Understanding.
ADHD in children and adolescents: limited evidence for broad improvements
Interpersonal difficulties associated with ADHD may arise through attention regulation, impulsive responding, inconsistent performance, difficulty waiting, or challenges in environments with demanding expectations. These causes matter because teaching conversational knowledge may not resolve the mechanism that interferes with its use. The large 2019 Cochrane review included 25 randomized trials and 2,690 children and adolescents. It found no clinically relevant benefit on teacher-rated social skills at the end of treatment (standardized mean difference 0.11, confidence interval approximately 0.00–0.22), with low or very low certainty across outcomes. Parent ratings sometimes appeared more favorable, demonstrating how conclusions change with the informant and outcome Storebø et al., 2019.
A newer school-based systematic review and meta-analysis reported an overall negligible effect (ES = 0.09) across a subset of eligible ADHD studies, despite variation in individual estimates and program formats Bussanich et al., 2025. It would therefore be misleading to offer general SST as an established stand-alone solution for ADHD-related social difficulties. Assessment should consider the setting, supports, attention and emotion regulation, bullying, school practices, and whether the child can use a skill that is already known.
Professional communication: improvements in trained performance, uncertainty about downstream outcomes
Teaching communication in health care and social work often uses interviewing practice, standardized scenarios, observed consultations, and feedback. A Campbell systematic review of 15 studies of social work students found indications of improvements in some communication and empathy-related skills, but the primary studies were heterogeneous and frequently at moderate to serious or incompletely assessed risk of bias. It was not possible to establish one superior format or draw confident conclusions about real-world professional outcomes Reith-Hall and Montgomery, 2023.
Research on standardized patients shows a similar distinction between learning outcomes and practical consequences. The systematic review by Rutherford-Hemming and colleagues found some advantages for certain skill outcomes over selected simulation methods, while noting the absence of measured patient outcomes in those comparisons Rutherford-Hemming, Herrington, and Ngo, 2024. These studies support realistic practice and specific feedback as reasonable pedagogical components, while leaving the durability of downstream effects open.
Everyday social difficulties and social anxiety: an evidence gap
People seeking SST outside defined clinical or educational programs form a heterogeneous population. One person wants conversational fluency; another wants to speak up at work; another experiences fear of negative judgment; another faces exclusion rather than a skill deficit. There is no single well-established effect estimate for a generic course delivered to all such adults. It is therefore more accurate to examine the individual goal and the evidence for the relevant problem than to advertise a universal improvement rate.
Social anxiety can reduce participation even when interpersonal knowledge is adequate. Anxiety-focused assessment and psychological treatment may be relevant when persistent fear, avoidance, and impairment are present; that decision belongs to an appropriately qualified professional. The distinction between anxious self-evaluation and observed conversational behavior is explored separately in Social Anxiety in Conversations: Self-Focus, Silence, and Fear of Judgment.
What counts as evidence of effectiveness?
Skill knowledge
Knowledge tests ask whether a participant can identify or describe a response. They are relatively easy to administer and may show large gains after direct instruction. Knowledge is an intermediate endpoint. Recognizing the most appropriate reply in a multiple-choice item cannot demonstrate that the reply will be available during a stressful conversation.
Performance under observation
Structured role-plays and standardized interactions can assess whether a participant executes a behavior, how flexibly they respond, and whether the performance changes over time. Ideally, raters are trained, the tasks represent the person's goals, and observers are unaware of assignment or assessment time when feasible. Even a valid performance test captures only part of a relationship's natural variability.
Ratings by the participant and others
Self-report can capture perceived confidence, strain, acceptance, and satisfaction; caregiver, teacher, peer, or clinician reports can capture behavior observed in distinct settings. Ratings are vulnerable to expectations, response styles, unequal opportunities to observe, and cultural assumptions. The disagreement between parent and teacher ratings in ADHD trials shows why using multiple informants is informative rather than a nuisance Storebø et al., 2019.
Naturalistic behavior and meaningful participation
A stronger outcome asks whether the person can now make a needed request, sustain a chosen relationship, take part in a valued activity, avoid unwanted concessions, or access support. These outcomes can be evaluated using ecological observation, consensual diaries, repeated real-world behavioral measures, or independently corroborated examples. They may be influenced by social access, peers, discrimination, workplace expectations, and opportunity as well as by the person's behavior. Such influences should be measured or acknowledged instead of attributed automatically to the intervention.
Maintenance, generalization, and quality of life
The most demanding evidence follows participants after the course and examines novel situations, partners, and contexts. An apparent improvement on the rehearsed scenario may disappear when the facilitator is absent. Conversely, meaningful participation may improve through accommodations or changed opportunities even without conspicuous changes in conventional skills ratings. Long-term follow-up, social validity, and the participant's own priorities establish whether the outcome represents a real benefit Gresham and Lemanek, 1983; Nejati et al., 2024.
How research designs can mislead
Many intervention trials cannot conceal group assignment from participants or facilitators. When both the teacher and outcome rater know who received training, enthusiasm or expectations can affect ratings. Waiting-list and no-treatment controls establish a different comparison from active, credible alternatives. Outcome switching, small samples, missing follow-up, incompatible scales, and selective publication can magnify apparent benefits. A pooled effect estimate can also hide sharp differences between schoolchildren, clinical rehabilitation participants, and professional trainees.
Studies should specify the program components, duration, delivery setting, participant selection, comparator, preregistered outcomes, adherence, unwanted effects, and the length and completeness of follow-up. Ideally, they should collect both independent performance measures and meaningful real-world outcomes. Review-level certainty should reflect the quality of these components. The contrasting psychosis review results and the weak overall ADHD findings make clear that evidence cannot be summarized honestly as a simple statement that SST works or fails Almerie et al., 2015; Bussanich et al., 2025.
Designing a sound, individualized training program
Begin with a specific and voluntary goal
The participant should be able to describe an interaction they wish to handle differently and why it matters. Useful goals include explaining a need to a colleague, joining a hobby group, asking for accessible instructions, or repairing a routine misunderstanding. Avoid objectives like acting normal, appearing extroverted, or pleasing everyone. The goal should distinguish what the individual can influence from the responsibilities of other people and institutions.
Assess the baseline and the setting
Identify whether the issue is missing knowledge, limited practice, anxiety, sensory demands, language comprehension, unwelcoming peers, risk of exploitation, or inconsistent access to opportunities. Observe or discuss more than one setting when possible. Ask what the person already does successfully and when a difficulty occurs. A social-skills difficulty is a description of a challenge, not an independent diagnostic category.
Choose representative behaviors and multiple options
For a target such as asking for clarification, the facilitator may model a concise question, a request to write instructions down, or a follow-up message. The trainee should decide which forms are safe and culturally workable. Scripts may reduce cognitive load early in learning, while later practice should permit spontaneous variants and explain the tradeoffs among them. This helps prevent a rigid rule from being used in situations for which it was never intended.
Rehearse in graduated, varied conditions
Begin with conditions the person can tolerate, then vary the partner, topic, setting, and type of response. Incorporate ambiguous or unsuccessful outcomes so that the participant can learn recovery and adjustment. After each attempt, review the goal, available information, and what was learned. Training should allow a participant to decline exercises that expose private experiences or create undue distress.
Transfer with support and measure what changes
Agree on optional between-session opportunities that are realistically available. Consider support from peers, family, or coworkers when the participant consents. Collect a brief baseline, a measure soon after training, and follow-up observations that relate to the original goal. The intervention may need adjustment when a person succeeds in rehearsal but experiences barriers in everyday life. The process is iterative, and the meaningful endpoint is participation and agency.
A worked example: clarification in a workplace meeting
Suppose an employee understands their assignment only after the meeting has ended. The goal is to obtain accurate task instructions without unnecessary delay. Training might model different options: ask what success looks like during the meeting, request a written summary, or clarify responsibilities afterward. The employee rehearses those options with a supportive partner, receives feedback about whether the request was clear, and chooses which to try in an actual meeting. The employer's responsibility is also considered: inaccessible jargon, penalization of questions, or unclear role allocation may require organizational change. Evaluation records whether the employee received the needed information, how stressful the experience was, and whether future meetings became more workable. This outcome is more meaningful than rating the employee's eye contact during the rehearsal.
Cultural variation, disability access, and communication rights
Interaction rules differ across languages, communities, institutions, and relationships. Pauses, conversational overlap, direct requests, emotion displays, distance, forms of address, and gaze can carry different meanings. A training program that treats one culturally dominant style as universally correct risks teaching the participant to suppress effective personal communication. Trainers should ask which contexts matter and adapt scenarios to the social rules actually operating there.
Disability-related communication access can require written materials, alternative and augmentative communication, interpreters, sensory accommodations, predictable structure, breaks, slower pacing, or explicit instructions. These supports are part of effective intervention design. NICE guidance explicitly advises attention to individual sensitivities, clarity, routine, and personally appropriate formats in interventions involving autistic adults NICE, CG142.
The ethical objective is to extend choices, not to teach masking for the convenience of observers. A person has the right to communicate directly, refuse unwanted contact, use a preferred mode of expression, and decide whether a social convention is worth adopting. Social outcomes must be judged against reciprocal relationships and the participant's welfare rather than a checklist of conventional gestures.
Power, coercion, and interpersonal safety
Training in negotiation, politeness, disagreement, or emotional expression takes place within unequal systems of power. A learner should not be instructed to resolve threats, harassment, abuse, or coercive control through better listening or more accommodating speech. There are situations in which disengagement, documentation, organizational intervention, advocacy, or access to protection is the appropriate response. Practicing safety-related refusals and requests for help can be legitimate targets; asking the endangered person to manage the aggressor's emotions is not a safe goal.
In education or employment, group exercises must respect confidentiality and consent. Participants should be free to decline potentially exposing scenarios, especially those involving trauma, discrimination, sexuality, or coercion. Trainers should establish how concerns will be handled and should not pressure someone to disclose experiences publicly. NICE's autism guidance includes consideration of anti-victimization work when appropriate, illustrating why social learning should include safety and informed decision-making NICE, CG142.
When social skills training may be an unsuitable first step
Repeated communication difficulty may reflect a language disorder, hearing impairment, inadequate access to communication technology, significant social anxiety, a mood or psychotic disorder, trauma-related responses, cognitive load, exclusion, or a hostile setting. These possibilities call for targeted assessment rather than an automatic assumption of deficient skills. For example, a person who avoids meetings because of harassment requires workplace protection. Someone whose comprehension depends on written information requires access to that information. Someone with severe clinical symptoms may need coordinated care in which skills work is only one optional component.
A responsible provider explains their qualifications, approach, likely demands, evidence for the target population, expected measures, privacy arrangements, and procedures for responding to distress. There is no scientifically justified universal number of sessions or single format for all goals and all populations. The program should be reconsidered if it produces fatigue, shame, forced conformity, deteriorating well-being, or no meaningful improvement despite repeated effort.
How strong is the evidence? An interpretive guide
Established: Instructions, modeling, practice, and feedback are recognizable components of behavioral SST, and controlled research shows that certain trained behaviors and proximal performance outcomes can improve in specific settings. This establishes the feasibility of skill acquisition, not a universal long-term effect Spence, 2003; Kurtz and Mueser, 2008.
Preliminary: Particular group, peer-supported, simulation-based, and structured autism-related programs have shown promising outcomes for selected participants and measures. Sample size, setting, comparison condition, and participant preferences limit the generality of those estimates Laugeson et al., 2015; Rutherford-Hemming, Herrington, and Ngo, 2024.
Limited: Evidence for durable improvements in untrained daily interactions, independent quality-of-life outcomes, and generic social-skills courses for all adults is comparatively sparse or heterogeneous. Findings from one diagnostic, developmental, or professional group should remain attached to that group Reith-Hall and Montgomery, 2023; Nejati et al., 2024.
Contested: Reviews disagree about how much training benefits different populations, partly because they include different trials, comparators, and outcomes. Low-certainty Cochrane findings in ADHD and schizophrenia caution against broad efficacy claims even where individual studies report improvements Storebø et al., 2019; Almerie et al., 2015.
Questions readers frequently ask
Can adults learn social skills through training?
Adults can learn and refine specific interpersonal behaviors, including rehearsed conversational responses. Whether a program produces lasting changes in a particular person's relationships depends on the initial difficulty, the training format, the setting, opportunities to use the skills, and how outcomes are measured. Evidence from clinical rehabilitation or professional education should not be presented as a universal estimate for all adults.
Does role-play make a person better at real conversations?
Role-play can provide a structured opportunity to practice, receive feedback, and explore alternative responses. Performance during a role-play is a learning measure. Improvement in untrained conversations requires separate observation or follow-up and may also depend on the behavior of other participants in the interaction.
Is SST a treatment for social anxiety?
SST can address identified behavioral targets, but social anxiety often involves fear, avoidance, and self-evaluation that require their own assessment and intervention. A person may have adequate conversational skills and still experience significant anxiety. Training should be chosen according to the mechanism of difficulty rather than used automatically because someone feels awkward.
Is SST recommended for autism or ADHD?
Guidance and research differ. NICE supports consideration of individually appropriate social learning for autistic adults who identify relevant difficulties; this is a matter of preference and individualized planning. Systematic reviews of ADHD interventions in young people have found uncertain or negligible broad effects on important measures. Neither diagnosis implies that every person needs the same program NICE, CG142; Bussanich et al., 2025.
How long does training take to work?
Programs vary greatly in frequency, length, participant age, and complexity. Acquiring a rehearsal skill can occur before a new behavior becomes comfortable in daily life. A responsible answer specifies the target, baseline, training dose, and follow-up period, rather than promising a fixed timeframe for social competence.
How can I tell if a program is working?
Choose one or two meaningful goals before starting; measure a relevant behavior or participation outcome in daily life; review progress with the participant and, if appropriate, a trusted partner; and assess maintenance after sessions end. Include unwanted effects and whether environmental obstacles have changed. A high course-completion rate or better performance on an instructor's quiz is insufficient by itself.
Conclusion
Social skills training is most defensible when it is a voluntary, goal-specific process of demonstration, rehearsal, feedback, and supported use in real settings. Its strongest educational rationale is the teachability of particular behaviors; the scientific challenge is establishing that those behaviors become flexible, durable, and beneficial in ordinary life. The clearest evaluations separate learning from generalization, assess outcomes that participants value, and account for the people and institutions with whom communication occurs. That standard makes SST both more useful and more scientifically accountable.
Related Articles
Cluster Communication, Conflict and Social Skills — the broader scientific framework for communication, reciprocity, social competence, and conflict.
Assertive Communication: Skills, Examples, and How to Practice — a separate examination of expression, requests, and boundaries.
Autistic Communication Differences in Adults: Language and Mutual Understanding — neurodevelopmental communication and reciprocal understanding.
Social Anxiety in Conversations: Self-Focus, Silence, and Fear of Judgment — why the experience of social difficulty can diverge from behavioral skill.
Feeling Unheard in Marriage: Listening, Responsiveness, and Why Conversations Stop Connecting — relationship context and reciprocal responsiveness.
Social Skills in Psychology: Definition, Components and Assessment — Define social behaviors, competence, contexts and assessment methods.
How to Improve Social Skills as an Adult: Practice and Evidence — Explains skill acquisition, role play, feedback, rehearsal and generalization.
Social Problem-Solving Skills: How People Generate and Evaluate Responses — Explains problem definition, alternatives, consequences and collaborative response.
References
Almerie, M. Q., Al Marhi, M. O., Jawoosh, M., Alsabbagh, M., Matar, H. E., Maayan, N., and Bergman, H. (2015). Social skills programmes for schizophrenia. Cochrane Database of Systematic Reviews, CD009006. https://doi.org/10.1002/14651858.CD009006.pub2
Bussanich, G., Harrison, J. R., Peltier, C., Patel, A., Mehta, P., and Patel, K. (2025, online publication). School-based social skills interventions for youth with ADHD: A systematic review and meta-analysis. Journal of Attention Disorders. https://doi.org/10.1177/10870547251364578
Gresham, F. M., and Lemanek, K. L. (1983). Social skills: A review of cognitive-behavioral training procedures with children. Journal of Applied Developmental Psychology, 4(3), 239–261. https://doi.org/10.1016/0193-3973(83)90021-7
Kurtz, M. M., and Mueser, K. T. (2008). A meta-analysis of controlled research on social skills training for schizophrenia. Journal of Consulting and Clinical Psychology, 76(3), 491–504. https://doi.org/10.1037/0022-006X.76.3.491
Laugeson, E. A., Gantman, A., Kapp, S. K., Orenski, K., and Ellingsen, R. (2015). A randomized controlled trial to improve social skills in young adults with autism spectrum disorder: The UCLA PEERS program. Journal of Autism and Developmental Disorders, 45, 3978–3989. https://doi.org/10.1007/s10803-015-2504-8
National Institute for Health and Care Excellence. (2012, updated). Autism spectrum disorder in adults: Diagnosis and management (CG142). https://www.nice.org.uk/guidance/cg142/chapter/Recommendations
Nejati, V., Peyvandi, A., Nazari, N., and Abadi, F. (2024). The effectiveness of social training in individuals with autism spectrum disorder (ASD): A systematic review and transfer analysis. Scientific Reports, 14, 32131. https://doi.org/10.1038/s41598-024-83953-9
Reichow, B., Steiner, A. M., and Volkmar, F. (2012). Social skills groups for people aged 6 to 21 with autism spectrum disorders (ASD). Cochrane Database of Systematic Reviews, CD008511. https://doi.org/10.1002/14651858.CD008511.pub2
Reith-Hall, E., and Montgomery, P. (2023). Communication skills training for improving the communicative abilities of student social workers. Campbell Systematic Reviews, 19, e1309. https://doi.org/10.1002/cl2.1309
Rutherford-Hemming, T., Herrington, A., and Ngo, T. P. (2024). The use of standardized patients to teach communication skills: A systematic review. Simulation in Healthcare, 19(1S), S122–S128. https://doi.org/10.1097/SIH.0000000000000766
Spence, S. H. (2003). Social skills training with children and young people: Theory, evidence and practice. Child and Adolescent Mental Health, 8(2), 84–96. https://doi.org/10.1111/1475-3588.00051
Storebø, O. J., Elmose Andersen, M., Skoog, M., Joost Hansen, S., Simonsen, E., Pedersen, N., Tendal, B., Callesen, H. E., Faltinsen, E., and Gluud, C. (2019). Social skills training for attention deficit hyperactivity disorder (ADHD) in children aged 5 to 18 years. Cochrane Database of Systematic Reviews, CD008223. https://doi.org/10.1002/14651858.CD008223.pub3
Turner, D. T., McGlanaghy, E., Cuijpers, P., van der Gaag, M., Karyotaki, E., and MacBeth, A. (2018). A meta-analysis of social skills training and related interventions for psychosis. Schizophrenia Bulletin, 44(3), 475–491. https://doi.org/10.1093/schbul/sbx146
Wang, S.-Y., Cui, Y., and Parrila, R. (2011). Examining the effectiveness of peer-mediated and video-modeling social skills interventions for children with autism spectrum disorders: A meta-analysis in single-case research using HLM. Research in Autism Spectrum Disorders, 5(1), 562–569. https://doi.org/10.1016/j.rasd.2010.06.023