Improving Children’s Social Skills With Virtual Reality: A Tailored Single-Session Training in Special Education

Abstract

This within-child study examined whether a tailored, single-session virtual reality (VR) training can enhance social skills among children enrolled in special education programs for behavior problems. Forty-nine children (ages 7–13; 85.7% boys; 57.1% ASD, 20.4% ADHD) completed one 20-min VR session. For each child, teachers selected one skill for children to practice with: staying calm (anger regulation), asking to join a group (peer entry), or saying no (assertiveness). Relative to baseline, in-VR observations showed substantial immediate improvements of the targeted skill (η²_p = .35), but teacher reports two weeks later showed no comparable improvement in the classroom (η²_p = .07). Teachers also reported no changes in children’s aggressive behavior or emotional problems two weeks later. Children reported moderate levels of immersion and perceived efficacy, low-to-moderate emotional engagement, and high appreciation of the training, VR, and trainer-child relationship. Overall, findings suggest that interactive VR is an attractive tool to practice socially skilled behavior, but that transfer to real life may require more sessions, explicit bridging strategies, and/or more emotionally engaging VR.

Plain Language Summary

In special education programs for behavior problems, many children find social situations challenging. For example, it can be difficult to stay calm after losing a game, to ask to join a group, or to say no when another child is being pushy. Interactive virtual reality (VR) may offer a safe and fun way to practice these situations with virtual peers. This study examined whether a single session of training in VR may be enough to improve the social skills of 49 children in special education (ages 7-13, mostly boys, many meeting criteria for an autism spectrum disorder or ADHD). For each child, teachers selected one skill for children to practice with: staying calm (anger regulation), asking to join a group (peer entry), or saying no (assertiveness). During the session, children first learned why the skill matters and what they could say or do. Then they practiced the skill three times in VR. For example, a child might practice staying calm while a virtual peer bragged about winning a game against them. Right after the session, children showed clear improvement in the specific skill they practiced, when tested in another VR scenario. However, teachers did not notice changes in childrens socially skilled behavior in class two weeks later, nor in their aggression or emotional problems. Children reported that the training felt somewhat effective. They felt somewhat immersed in the VR world and experienced some emotional involvement. They also reported they really liked the training, the VR activities, and their relationship with the trainer. Overall, interactive VR appears to be a safe and fun way for children to practice social skills. However, to change behavior in the classroom, children may need more than one session, clear strategies to bridge what they learned in VR to their daily life, or VR practice that feels more emotionally engaging.

Keywords

social skills virtual reality special education behavior problems emotional problems

Children enrolled in special education programs for behavior problems often experience difficulties with social skills (Frostad et al., 2007), including regulating anger, entering peer groups, and responding assertively. Such limited social skills place children at increased risk for various negative outcomes, including aggressive behavior (Glenn et al., 2021) and emotional problems (Obradović et al., 2009). It is therefore important that special education services have tools to improve these skills with their students. Interactive virtual reality (VR)—in which children can practice responding to realistic social situations with virtual peers—may offer this: it can be used flexibly, tailored to individual needs, supports repeated behavioral practice, and is engaging for children (Alsem et al., 2023). This study examined whether such a single-session VR training improves children’s social skills and reduces aggressive and emotional problems.

VR is an emerging and promising intervention method, yet empirical evidence for reducing child mental health problems remains limited (Blanco et al., 2024; Halldorsson et al., 2021). Available evidence suggests that VR can improve children’s social skills (meta-analysis of 15 studies; Zhang et al., 2023), reduce aggressive behavior (one randomized controlled trial; Alsem et al., 2023), and may alleviate emotional problems (systematic scoping review of 11 studies; Blanco et al., 2024). Moreover, single-session VR interventions show initial promise, although this evidence is restricted to anxiety outcomes (Blanco et al., 2024), highlighting the need to examine single-session VR-based social skills training.

VR’s comparative advantage over traditional, typically group-based social skills training may stem from its capacity to immerse children in realistic practice contexts, eliciting the emotions that accompany real-world social interactions and enabling rehearsal under affective load (Alsem et al., 2023). Moreover, children may be motivated to engage with VR, as technology appeals to them (Weisz et al., 2019). Nevertheless, the literature has primarily evaluated multi-session, comprehensive VR intervention packages (Roncero et al., 2025). Such comprehensive intervention packages may not be feasible in an education setting. Evidence of brief, single-session VR training remains limited, despite the practical relevance of flexible application in special education.

This within-child study examined whether a tailored, single-session VR training improved children’s: (a) VR-observed social skills immediately post-training; (b) teacher-rated social skills two weeks post-training; and (c) teacher-rated aggressive behavior and emotional problems two weeks post-training, each compared with pre-training assessments. We also assessed children’s experiences of the training: their appreciation, immersion, emotional engagement, therapeutic relationship, and perceived efficacy. All data, code, and materials (including all items and coding sheets) are publicly available via the Open Science Framework at: https://osf.io/u35jm.

Method

Participants

Participants were 49 children (85.7% boys, M_age = 10.41, SD = 1.46) aged 7-13, recruited from two Dutch special-education primary schools serving children who cannot participate in regular education due to severe social, emotional, and/or behavioral difficulties. Teachers reported that most children (71.4%) had a DSM diagnosis, including autism spectrum disorder (57.1%), attention-deficit hyperactivity disorder (20.4%), post-traumatic stress disorder (4.1%), and developmental language disorder (2.0%).

Each school’s special education professional selected classrooms with sufficient time to participate (k = 6; N = 60). Teachers approached caregivers and children aged ≥12, the majority of whom provided written informed consent (consent rate: 88%). Four children withdrew before training; they did not differ from participants on pre-training teacher-reported social skills, aggressive behavior, or emotional problems, F (3, 48) = 0.09, p = .963, η² < .01. Power analysis at .80 indicated N > 45 to detect small-to-medium within-subject improvements (g = .38; Zhang et al., 2023). This study was approved by the University of Amsterdam Ethics Review Board (FMG-8128).

Procedure

Data were collected at schools in Spring 2024. Teachers completed paper questionnaires rating children’s social skills, aggressive behavior, and emotional problems, both pre-training and two weeks post-training. They also selected one of three social skills that children could practice within VR: anger regulation (n = 26), peer entry (n = 12), or assertiveness (n = 11). Children were individually tested in a quiet room by two of six trained graduate students (one interacting, one observing) in 60-min sessions. The experimenter introduced the VR environment—comprising classroom and schoolyard contexts with experimenter-controlled teachers and peers (Figure 1)—and familiarized children with talking, walking, and grabbing objects (for details on the VR, see Alsem et al., 2023). Children were informed that they could exit the VR at any time, and the experimenter was trained to monitor potential side effects (none occurred, consistent with our previous VR studies). The VR protocol then proceeded as follows: (1) a 5-min pre-training assessment (one scenario); (2) a 20-min social skills training (three scenarios); and (3) a 5-min post-training assessment mirroring the pre-training scenario. Last, children reported their VR training experience.

Figure 1.

The virtual reality environment

Social Skills Training

We developed three tailored, single-session VR trainings targeting anger regulation, peer entry, or assertiveness. These intervention targets were selected by participating teachers and special education professionals using a literature-based list of social skills (Spence, 2003). All trainings were grounded in cognitive-behavioral principles (Marques et al., 2024) and informed by existing interventions (Alsem et al., 2023; Vliek et al., 2019). VR scenarios were developed based on prior pilot work incorporating children’s input (Verhoef et al., 2021) and refined with feedback from special education professionals. For example, an initial anger staircase was replaced with the schools’ emotion pyramid to align with existing practices, resulting in the final training protocols.

The three VR trainings shared an identical structure: First, the experimenter normalized skill practice using a structured worksheet, explained the function of emotions (children drew the target emotion), and introduced a brief cue phrase for the rehearsal of anger regulation (“I’m stepping out for a bit”), peer entry (“Can I join you?”), or assertiveness (“Stop! I don’t like this”). Second, children rehearsed the skill in three VR scenarios (Table S1). For anger regulation, for example, children lost a competitive can-knockdown game to a bragging peer. The experimenter observed and, as needed, coached children (disabling the voice transformer), then provided feedback after each scenario. Children could color a star per scenario to earn a signed diploma. All children received this diploma.

Measures

Social Skills (Observed in Virtual Reality)

We assessed children’s social skills immediately pre- and post-training using standardized, skill-matched scenarios (Table S1). Pre/post scenarios were conceptually similar but perceptually different (e.g., losing different games). Children were asked to treat the virtual school as real and to behave naturally. The experimenter followed scripted interactions and ended each scenario upon a response or after 30 s of nonresponse. The observer recorded children’s verbatim responses and behaviors (see Table S2). After children left, the observer and experimenter independently rated skill-use adequacy on 1-5 scales tailored to the targeted skill (e.g., for anger regulation: 1 = displays aggressive behavior and 5 = uses skill and remains calm; see Supplemental Material for coding). Interrater reliability was adequate at pre-training (κ = .95) and post-training (κ = .79); discrepancies were resolved by consensus.

Social Skills (Teacher-Rated)

We assessed classroom social skills using two items per skill (Alsem et al., 2022; Driestar Onderwijsadvies, 2016). Teachers rated past-week frequency on a scale from 1 (never) to 5 (very often); for instance, for anger regulation: “This week this student managed to do something about his/her anger” (see Supplemental Material for all items). Item means yielded pre/post scores for anger regulation (α = .80/.86), peer entry (α = .90/.87), and assertiveness (α = .74/.75). We used a variable including only each child’s target skill for the analyses.

Aggressive Behavior and Emotional Problems (Teacher-Rated)

We assessed classroom aggressive behavior and emotional problems using three items each (Alsem et al., 2022; Goodman, 1997). Teachers rated past-week frequency on a scale from 1 (never) to 5 (very often); for instance, for aggressive behavior: “This week this student had a fight with someone” (see Supplemental Material for all items). Item means yielded pre/post scores for aggressive behavior (α = .88/.82) and emotional problems (α = .78/.74).

Children’s Virtual Reality Training Experience

We assessed children’s VR training experience using 17 items rated from 1 (totally disagree) to 5 (totally agree), with mean scores for appreciation (3 items; α = .48; item-level analyses yielded similar results), immersion (4 items; α = .79), emotional engagement (3 items; α = .67), trainer-child relationship (3 items; α = .60), and perceived efficacy (4 items; α = .66; see Supplemental Material for all items). Children also provided overall grades (1–10) for the training and for the VR.

Analyses

Analyses were conducted in SPSS 29.0. We used four separate repeated-measures ANOVAs to test whether children improved in their social skills (observed and teacher-rated), aggressive behavior, and emotional problems. Effect sizes are reported as partial eta squared (η²_p) and interpreted as small (.01), medium (.06), or large (.14; Cohen, 2013). We summarized children’s VR training experience with descriptive statistics.

Results

Data Screening

Table 1 reports descriptive statistics. We found no outliers (i.e., z > 3.29). Most variables met normality (z < 1.96 for skewness and kurtosis), except observed social skills, for which we used Friedman’s ANOVA. Missing data were minimal (0.8%) and handled via pairwise deletion.

Table 1.

Descriptive Statistics and Zero-Order Correlations of Pre- and Post-Training Social Skills (Observed and Teacher-Rated) and Teacher-Rated Aggressive Behavior and Emotional Problems

		M	SD	Range	1	2	3	4	5	6	7
1.	Observed skills (pre)	3.37	1.58	1.00–5.00	-
2.	Observed skills (post)	4.39	1.12	1.00–5.00	.49^***	-
3.	Teacher-rated skills (pre)	2.49	0.90	1.00–4.50	.09	−.03	-
4.	Teacher-rated skills (post)	2.72	0.95	1.00–5.00	−.09	−.06	.68^***	-
5.	Aggressive behavior (pre)	2.50	1.10	1.00–5.00	.11	.11	−.09	−.27^†	-
6.	Aggressive behavior (post)	2.61	1.14	1.00–5.00	−.02	−.04	−.12	−.26^†	.78^***	-
7.	Emotional problems (pre)	2.39	1.00	1.00–4.67	−.06	−.14	−.33^*	−.21	.12	.08	-
8.	Emotional problems (post)	2.51	0.99	1.00–5.00	−.25^†	−.28^†	−.31^**	−.22	.02	.16	.76^***

Note. ^†p < .10. ^*p < .05. ^**p < .01. ^***p < .001.

Improvements in Children’s Social Functioning

Results indicated large gains in VR-observed social skills, χ² (1) = 19.20, p < .001, η²_p = .35, but not in classroom social skills two weeks later, F (1, 47) = 3.42, p = .071, η²_p = .07 (Table 1). Teachers also reported no changes in classroom aggressive behavior, F (1, 47) = 0.71, p = .404, η²_p = .02, or emotional problems two weeks later, F (1, 47) = 1.25, p = .269, η²_p = .03.

Children’s Experience of the Virtual Reality Training

Children rated the training highly (M = 4.31 out of 5, SD = 0.59) and gave high grades to the training (M = 8.78 out of 10, SD = 1.49) and VR (M = 8.45, SD = 1.88). Children reported moderate immersion (M = 3.33 out of 5, SD = 1.01) and perceived efficacy (M = 3.54; SD = 0.80), and low-to-moderate emotional engagement (M = 2.23, SD = 1.05). They highly appreciated the trainer-child relationship (M = 4.13, SD = 0.68).

Discussion

This within-child study indicates that a single-session, tailored VR training can strengthen children’s targeted social skill when assessed within the training context. However, these immediate gains did not generalize to teacher-reported behavior over two weeks. The large in-VR improvement in children’s social skills aligns with prior work showing that interactive VR can scaffold children’s social behavior and emotion regulation (Alsem et al., 2023; Arts et al., 2025). The lack of classroom change suggests that more is needed to facilitate transfer to everyday settings.

Dose is likely critical: effective social skills and CBT-informed interventions for youth typically provide multiple, spaced opportunities for rehearsal, feedback, and consolidation (Alahmari et al., 2025; Alsem et al., 2023). Yet, additional explanations are possible. Generalization often depends on explicit bridging strategies—such as goal setting, teacher prompts, or reinforcement in class—which were not embedded here. Children also reported low-to-moderate emotional engagement, whereas sufficient engagement is crucial to realistically practice new skills (Sukhodolsky et al., 2016). Although the VR training targeted the teacher-identified priority skill for each child, the VR scenarios were not individualized to children’s daily experiences. Scenarios that more closely mirror children’s everyday struggles may provide a more engaging and relevant practice environment, potentially enabling transfer even after one session. Future research could experimentally manipulate dose, bridging strategies, and scenario tailoring, to identify components that promote generalization.

Methodologically, this study benefitted from standardized VR assessments, strong interrater reliability for VR-observed skills, and skill-tailored protocols. Nonetheless, several limitations warrant caution. First, the absence of a control group precludes causal inference, leaving expectancy effects and measurement reactivity unaddressed. Second, coders of VR-observed skills were not blind to hypotheses (although the verbatim transcripts and behavior logs provide transparent evidence, see Table S2). Third, teacher-reported outcomes comprised only two or three items per construct, and indications of actual aggressive behavior may be relatively infrequent within two weeks; both factors may have limited sensitivity to detect subtle changes. Fourth, although the sample size was sufficient to detect overall effects of the VR training, it was insufficient to compare the effectiveness of the three skill-specific protocols. Last, subtle changes in children’s classroom behavior may not have been detected by teachers; incorporating self- and peer-report measures could provide a more comprehensive and sensitive assessment.

In sum, a single, tailored VR session can yield immediate, context-specific gains in children’s social skills and is well-received by students in special education. To achieve meaningful classroom change and downstream effects on aggression and emotional problems, the training likely requires greater dose, bridging strategies, and/or more personalized, emotionally engaging VR scenarios.

Supplemental Material

Supplemental Material - Improving Children’s Social Skills With Virtual Reality: A Tailored Single-Session Training in Special Education

Supplemental Material for Improving Children’s Social Skills With Virtual Reality: A Tailored Single-Session Training in Special Education by Sophie C. Alsem, Anouk van Dijk, and Bram O. de Castro in Clinical Child Psychology and Psychiatry.

Footnotes

ORCID iDs

Sophie C. Alsem

Anouk van Dijk

Bram O. de Castro

Ethical Considerations

This study was approved by the Ethics Review Board at the University of Amsterdam (FMG-8128). Caregivers and children age ≥12 provided written informed consent.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Anouk van Dijk and Bram de Castro were supported by grants from the Dutch Research Council (#VI.Veni.211G.076 and #VI.Vici.453-15-004, respectively).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The data are available via the Open Science Framework at: .

Supplemental Material

Supplemental material for this article is available online.

Author Biographies

Sophie C. Alsem holds a PhD in Developmental Psychology and works as a child psychologist in mental health care. Research interests include the effectiveness of interventions for children with behavioral problems and virtual reality-based treatment.

Anouk van Dijk is an assistant professor in Child Development and Education. Research interests include treatment resistance in adolescents with disruptive behavior problems, effectiveness of interventions, and social-cognitive mechanisms underlying aggressive behavior.

Bram O. de Castro is a Professor in Child Development and Education. His research focuses on understanding and changing the mechanisms that maintain behavior problems, such as social information processing, family interaction patterns, and peer social networks.

References

Alahmari

F. S.

Alhabbad

A. A.

Alshamrani

H. A.

Almuqbil

M. A.

(2025). Effectiveness of social skills training interventions for children with autism spectrum disorder: A systematic review and meta-analysis. Saudi Medical Journal, 46(3), 226–237. https://doi.org/10.15537/smj.2025.46.3.20240788

Alsem

S. C.

Keulen

Verhulp

E. E.

van Dijk

De Castro

B. O.

(2022). Capturing mechanisms of change: Weekly covariation in anger regulation, hostile intent attribution, and children’s aggression. Aggressive behavior, 48(2), 232–240. https://doi.org/10.1002/ab.22019

Alsem

S. C.

Van Dijk

Verhulp

E. E.

Dekkers

T. J.

De Castro

B. O.

(2023). Treating children's aggressive behavior problems using cognitive behavior therapy with virtual reality: A multicenter randomized controlled trial. Child Development, 94(6), e344–e361. https://doi.org/10.1111/cdev.13966

Arts

De Castro

B. O.

Luteijn

Elsendoorn

Maric

Vissers

C. T.

(2025). Virtual reality training to improve socio‐emotional functioning in adolescents with developmental language disorders: A multiple baseline effectiveness study. Social Development, 34(1), Article e12784. https://doi.org/10.1111/sode.12784

Blanco

Roberts

R. M.

Gannoni

Cook

(2024). Assessment and treatment of mental health conditions in children and adolescents: A systematic scoping review of how virtual reality environments have been used. Clinical Child Psychology and Psychiatry, 29(3), 1070–1086. https://doi.org/10.1177/13591045231204082

Cohen

(2013). Statistical power analysis for the behavioral sciences. Routledge.

Driestar Onderwijsadvies (2016). ZIEN! Leerling 5-8 Sociale vaardigheden. Driestar Onderwijsadvies & ParnasSys. http://www.driestar-educatief.nl/getmedia/43547bd0-aa8f-4821-82d4-9f383e83ba1c/Handleiding-ZIEN-Leerling-5-8-Sociale-vaardigheden-nov-2016.

Frostad

Pijl

S. J.

(2007). Does being friendly help in making friends? The relation between the social position and social skills of pupils with special needs in mainstream education. European Journal of Special Needs Education, 22(1), 15–30. https://doi.org/10.1080/08856250601082224

Glenn

D. E.

Michalska

K. J.

Lee

S. S.

(2021). Social skills moderate the time‐varying association between aggression and peer rejection among children with and without ADHD. Aggressive Behavior, 47(6), 659–671. https://doi.org/10.1002/ab.21991

10.

Goodman

(1997). The Strengths and Difficulties Questionnaire: A research note. Journal of child psychology and psychiatry, 38(5), 581–586. https://doi.org/10.1111/j.1469-7610.1997.tb01545.x

11.

Halldorsson

Hill

Waite

Partridge

Freeman

Creswell

(2021). Annual research review: Immersive virtual reality and digital applied gaming interventions for the treatment of mental health problems in children and young people: The need for rigorous treatment development and clinical evaluation. Journal of Child Psychology and Psychiatry, 62(5), 584–605. https://doi.org/10.1111/jcpp.13400

12.

Marques

M. M.

Wright

A. J.

Corker

Johnston

West

Hastings

Zhang

Michie

(2024). The behaviour change technique ontology: Transforming the behaviour change technique taxonomy v1. Wellcome Open Research, 8, 308. https://doi.org/10.12688/wellcomeopenres.19363.2

13.

Obradović

Burt

K. B.

Masten

A. S.

(2009). Testing a dual cascade model linking competence and symptoms over 20 years from childhood to adulthood. Journal of Clinical Child & Adolescent Psychology, 39(1), 90–102. https://doi.org/10.1080/15374410903401120

14.

Roncero

Moreno-Fernández

R. D.

Fernández-Moreno

Á.

(2025). Effectiveness of virtual reality interventions for aggression, anger and impulsiveness: A multilevel meta-analysis. Aggression and Violent Behavior, 81, Article 102034. https://doi.org/10.1016/j.avb.2025.102034

15.

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

16.

Sukhodolsky

D. G.

Smith

S. D.

McCauley

S. A.

Ibrahim

Piasecka

J. B.

(2016). Behavioral interventions for anger, irritability, and aggression in children and adolescents. Journal of Child and Adolescent Psychopharmacology, 26(1), 58–64. https://doi.org/10.1089/cap.2015.0120

17.

Verhoef

R. E.

van Dijk

Verhulp

E. E.

de Castro

B. O.

(2021). Interactive virtual reality assessment of aggressive social information processing in boys with behaviour problems: A pilot study. Clinical Psychology & Psychotherapy, 28(3), 489–499. https://doi.org/10.1002/cpp.2620

18.

Vliek

Overbeek

de Castro

B. O.

(2019). Effects of topper training on psychosocial problems, self-esteem, and peer victimisation in Dutch children: A randomised trial. PLoS One, 14(11), Article e0225504. https://doi.org/10.1371/journal.pone.0225504

19.

Weisz

J. R.

Kuppens

M. Y.

Vaughn-Coaxum

R. A.

Ugueto

A. M.

Eckshtain

Corteselli

K. A.

(2019). Are psychotherapies for young people growing stronger? Tracking trends over time for youth anxiety, depression, attention-deficit/hyperactivity disorder, and conduct problems. Perspectives on Psychological Science, 14(2), 216–237. https://doi.org/10.1177/1745691618805436

20.

Zhang

Luo

(2023). Using virtual reality interventions to promote social and emotional learning for children and adolescents: A systematic review and meta-analysis. Children, 11(1), 41. https://doi.org/10.3390/children11010041

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.72 MB