35% vs 67%
Retention or graduation of autistic students compared with their peers, as reported in a recent systematic review.
Koops et al., 2025 [8]Pilot randomized trial · Jordan
Virtual Classroom Training for Autistic University Students
Does immersive VR improve academic interaction more than the same practice on a screen? A pilot randomized comparison in Jordan.
Answering, asking, discussing and presenting are expected in most university courses, yet no training tools are designed for autistic students. This 16-week project tests whether practising in a virtual classroom, where the student controls class size, noise and pace, improves their participation in real classes.
Duration: 16 weeks (4 months), including 2 months of development
01
The problem
More autistic students reach university, but far fewer finish. Speaking up in class is part of almost every course, and there is nothing designed to help them practise it. [8]
77.8% vs 61.6%
Autistic vs non-autistic students who had considered dropping out (63 and 268 UK students).
Cage & McManemy, 2022 [3]1 of 12
Only one study from Jordan among assistive-technology studies on autism in Arab countries, and it involved children aged 5–13. None addressed adults or university students.
Al-Hendawi et al., 2023 [1]02
What the evidence shows
Three findings point to the same gap: simulation can work, but its effects rarely reach real life.
A
Simulation training can change real outcomes
25.0%virtual interview training (n = 48)
0.0%services as usual (n = 23)
Competitive integrated employment within six months in a randomized trial of 71 autistic youth (χ² = 6.9, p < 0.01), after 15 interview-practice sessions. The training was delivered on a standard computer, not a VR headset, and the authors note the trial was not fully powered.
Smith et al., 2021, Autism [14]B
Immersive VR gains are rarely shown in daily life
00.51.01.5
Immersive VR, controlled trials
0.29
VR/AR, controlled trials (9)
0.45
VR/AR, uncontrolled studies (26)
0.74
Effect sizes shrink as designs become more rigorous. In the immersive-VR review, every controlled trial targeted social or affective skills, only 3 of 15 studies reported follow-up, and none examined in depth whether gains transferred to real-life situations.
Effect size (d or g): 0.2 small, 0.5 medium, 0.8 large. Horizontal lines show 95% confidence intervals; for Karami et al. these were derived from the reported standard errors, and the interval for the nine controlled trials includes zero.
Li et al., 2023 [10]; Karami et al., 2021 [6]The transfer gap, seen in a single trial
A controlled immersive-VR trial with 94 autistic children (aged 6–12, 28 sessions). Measures close to the trained task improved; adaptive-behavior measures of daily life stayed near zero.
−0.500.51.0
Measures close to training
Affective expression
0.89
Social reciprocity
0.66
Faces test
0.57
Daily-life adaptive behavior (ABAS-II)
Leisure
0.12
Social
0.04
Community use
−0.02
Communication
−0.02
Self-direction
−0.12
This is why this program measures its primary outcome in the real world, not inside the headset.
Ip et al., 2018 [4], effect sizes (Cohen's d) as reported in Li et al., 2023 [10]C
VR is an established way to rehearse speaking in front of others
−1.39effect of VR exposure on public-speaking anxiety vs control (11 studies, N = 508)
23participants in the only randomized trial with autistic adults
In general populations, VR exposure reduced public-speaking anxiety with a large effect, only marginally below rehearsing before a real audience. For autistic adults, the only randomized trial we found was small (23 participants aged 18–25) and reported higher self-confidence and lower salivary cortisol than controls. The method is established; its use with autistic students has not yet been tested at scale.
Reeves et al., 2022 [12]; Kumazaki et al., 2019 [9]Central question
Does practising in an immersive virtual classroom improve how students take part in real classes more than the same practice on an ordinary screen?
We found no published trial testing this with autistic adults in an Arabic-speaking setting. Whatever the result, it tells universities whether headsets are worth the investment.
What makes this project different
| Existing studies | This program | |
|---|---|---|
| Population | Mostly children; Arab-region studies mainly with schoolchildren | Autistic university students in Jordan |
| Comparison | 26 of 35 studies in one meta-analysis had no control group | A dose-matched standard-screen arm isolates the effect of immersion |
| Outcome | Lab tests and in-VR performance; real-life transfer rarely examined | A real seminar task outside the headset, blindly rated |
| Follow-up | Only 3 of 15 studies reported follow-up | Post-program test and 2-week follow-up |
| Language and context | Only 12 studies across Arab countries, one from Jordan | Jordanian Arabic and English, for Jordanian students |
03
The program
One module, built and tested properly: academic interaction in class, delivered in 12 sessions, twice a week for 6 weeks. Further modules can follow on the same platform.
In this project
Academic interaction in class
Students take part in a virtual class: answering when called on, asking the lecturer a question, joining a group discussion and giving a short presentation. Class size, attention, questions directed at the student and background noise can all be adjusted. Practice starts with a small, friendly class and helper notes, and the supports are removed step by step.
Real-world task: taking part in a small live seminar, then contributing in a real course: a question, an answer or a short presentation.
Student
- Class size3 → 30
- Attentionattentive → distracted
- Questions to younone → frequent
- Noisequiet → busy
Inside the virtual classroom
Concept renders of the planned software, seen from the student's point of view inside the headset. They show the intended look; they are not screenshots of existing software.
How the training plays
Designed as a mission-based game for university students: clear levels, a free choice of challenge, and progress measured only against oneself.
- 1
Study room
3 friendly classmates; answer a question you know
- 2
Seminar room
10 students; ask the lecturer a question
- 3
Discussion table
Group discussion; share a view and disagree politely
- 4
Lecture hall
30 students; a short presentation with questions
- 5
The real class
Take part in a real class
Choose your challenge
A ringing phone, a latecomer, a tough question. The student picks what to add, and each challenge earns more points.
Strategy cards
Unlocked with progress and used during class: ask for the question again, take time to think, return to your notes.
Personal best
Compared only with your previous attempt, which you can replay. No leaderboards between students.
Start with your passion
The first discussions and talks are on a topic the student loves, building from their strengths.
Every session: a five-step cycle
- 1PreviewGoal, settings, how to stop
- 2ModelMore than one valid strategy
- 3RehearseAdaptive difficulty, pause and retry
- 4ReflectReplay with the facilitator
- 5BridgeA real task for the week
- Personal sensory settings
- Shaped by student feedback
- Eye contact and tone never scored
04
Trial design
A two-arm randomized controlled trial over 16 weeks that isolates the effect of immersion: both arms receive the same content and the same time with a facilitator, and the only difference is the headset.
Weeks 1–8 · 2 months
Software development
Building the virtual classroom in both versions, with ethics approval and recruitment in parallel, and prototype testing with a few students.
Weeks 7–8
Baseline
24 students, a baseline seminar task, then randomization.
Weeks 9–14
Program delivery
12 sessions, twice a week, with a real-world task each week.
Weeks 15–16
Testing and analysis
Primary outcome in week 15, follow-up in week 16, then the final report.
Participant flow in the randomized trial
Recruitment: enrolled autistic university students
Eligibility screening and informed consent
Baseline assessment (T0)
Randomization 1:1, n = 24
A · n = 12Immersive VR12 sessions over 6 weeks
B · n = 12Standard screenSame content and time
T1Week 15: primary outcome
T2Week 16: 2-week follow-up
Timeline (16 weeks)
Week 1481216
Software development (2 months)
Ethics approval and registration
Recruitment
Prototype testing with students
Baseline and randomization
Program delivery (12 sessions)
Primary outcome test
Follow-up
Analysis and report
- Week 8Software ready, randomization complete
- Week 14All 12 sessions complete
- Week 15Primary outcome measured
- Week 16Final report and paper draft
| Participants | Enrolled university students aged 18 or over with an autism diagnosis who consent to take part. Students with photosensitive epilepsy are excluded. |
|---|---|
| Randomization | 1:1, computer-generated, run by a team member not involved in sessions or rating. |
| Primary outcome | Number of verbal contributions (answers, questions, discussion turns and clarification requests) in a real 15-minute seminar task at week 15, with team members acting as lecturer and classmates, video-rated by two raters blind to group. Presentation quality within the task is rated with the PSCR, its nonverbal item adapted so eye contact is not scored. [13] |
| Secondary outcomes | Public-speaking anxiety (PRPSA, translated and back-translated into Arabic), cybersickness (SSQ), usability (SUS), maintenance at 2 weeks, and the number of real contributions in lectures during follow-up (questions, answers or presentations), as logged by the student. [11][7][2] |
| Feasibility criteria | Pre-specified: retention ≥ 80%, at least 9 of 12 sessions attended by 75% of students, no serious adverse events, rater reliability (ICC ≥ 0.75). |
| Sample size | 12 per arm, the accepted rule of thumb for pilot trials. The aim is to estimate the effect, not prove it, and to size a larger trial. [5] |
| Analysis | ANCOVA on the week-15 score adjusted for baseline, intention-to-treat, reporting the effect size with a 95% confidence interval. |
The full trial protocol
Written to the international SPIRIT 2025 standard (34 items), with a TIDieR intervention description and schedule of assessments. 15 pages.
05
Who is involved
Project team
Design the project, run the sessions, act as lecturer and classmates in the live seminar, blindly rate the recordings, and analyze the results.
Developer
Builds the virtual classroom in both versions (headset and standard screen) with identical content over 2 months (weeks 1–8), and supports them during delivery.
Students
24 autistic university students, each attending 12 sessions; a few also help test the prototype.
Equipment: 1–2 VR headsets, a high-performance computer, a standard computer and a quiet room.
06
Ethics and safety
- Institutional review board approval before recruitment
- Accessible informed consent in Arabic and English
- Separate consent for each optional data type
- Encrypted local storage under Jordan's Personal Data Protection Law
- Taking part or withdrawing never affects a student's grades
- Seated mode by default and an always-available stop control
- Pre-exposure screening and cybersickness monitoring every session
- All adverse events and withdrawals recorded
- No automated diagnosis or emotion inference
07
Optional extensions
The core project is complete in itself within 16 weeks. Any of the following can be added later on the same platform, each as a separate project, depending on resources and priorities.
Job interviews
A virtual interviewer in Jordanian Arabic and English. It has the strongest evidence of a real outcome: in a US trial with a computer-based simulator, 25% found work vs 0% of controls. [14]
Self-advocacy
Practice requesting academic adjustments, with a personal access profile, measured by genuine requests submitted.
Campus and transit
A 360° replica of the actual campus and Amman routes, measured by independence on an unfamiliar real route, plus a campus sensory map.
A larger trial
More students, a third usual-support group and 3- and 6-month follow-up, with a sample size calculated from this project's results.
08
Expected impact
Scientific
The first controlled trial in Jordan of virtual training in academic interaction for autistic students, and a direct estimate of the added value of immersion. At least one peer-reviewed paper.
Institutional
A ready Arabic–English training platform that stays at the university after the project, directly benefiting students.
Funding
Results within 4 months, ready to support an external grant application.
09
Risks and mitigation
| Risk | Mitigation |
|---|---|
| Slow recruitment due to under-diagnosis or non-disclosure | Recruitment starts during the build, extending to nearby universities if needed |
| Development runs past 2 months | Only one module is built, the screen version shares the same code, and features are frozen in week 7 |
| Ethics approval takes longer than planned | Submit a complete file in week 1; the build continues in parallel |
| Cybersickness or sensory discomfort | Seated default, short blocks with breaks, a standard-screen fallback |
| Raters learning a student's group | Raters did not train the student, and ratings are made from video |
References
- Al-Hendawi, M., Hussein, E., Al Ghafri, B., & Bulut, S. (2023). A scoping review of studies on assistive technology interventions and their impact on individuals with autism spectrum disorder in Arab countries. Children, 10(11), 1828. PMC10670675
- Brooke, J. (1996). SUS: A "quick and dirty" usability scale. In P. W. Jordan et al. (Eds.), Usability evaluation in industry (pp. 189–194). Taylor & Francis.
- Cage, E., & McManemy, E. (2022). Burnt out and dropping out: A comparison of the experiences of autistic and non-autistic students during the COVID-19 pandemic. Frontiers in Psychology, 12, 792945. doi:10.3389/fpsyg.2021.792945
- Ip, H. H. S., Wong, S. W. L., Chan, D. F. Y., et al. (2018). Enhance emotional and social adaptation skills for children with autism spectrum disorder: A virtual reality enabled approach. Computers & Education, 117, 1–15. doi:10.1016/j.compedu.2017.09.010
- Julious, S. A. (2005). Sample size of 12 per group rule of thumb for a pilot study. Pharmaceutical Statistics, 4(4), 287–291. doi:10.1002/pst.185
- Karami, B., Koushki, R., Arabgol, F., Rahmani, M., & Vahabie, A.-H. (2021). Effectiveness of virtual/augmented reality-based therapeutic interventions on individuals with autism spectrum disorder: A comprehensive meta-analysis. Frontiers in Psychiatry, 12, 665326. doi:10.3389/fpsyt.2021.665326
- Kennedy, R. S., Lane, N. E., Berbaum, K. S., & Lilienthal, M. G. (1993). Simulator Sickness Questionnaire: An enhanced method for quantifying simulator sickness. International Journal of Aviation Psychology, 3(3), 203–220.
- Koops, I., Duckett, P., & Gerace, A. (2025). Investigating the outcomes and quality of support programs for autistic young adults in higher education: A systematic review. Autism in Adulthood, 7(4), 403–420. doi:10.1089/aut.2024.0078
- Kumazaki, H., Sumiyoshi, T., Yamagata, B., Muramatsu, T., Mimura, M., & Kikuchi, M. (2019). Impact of autism-focused public speech training using simple virtual audience [Conference abstract]. INSAR 2019 Annual Meeting. insar.confex.com
- Li, C., Belter, M., Liu, J., & Lukosch, H. (2023). Immersive virtual reality enabled interventions for autism spectrum disorder: A systematic review and meta-analysis. Electronics, 12(11), 2497. doi:10.3390/electronics12112497
- McCroskey, J. C. (1970). Measures of communication-bound anxiety. Speech Monographs, 37(4), 269–277.
- Reeves, R., Curran, D., Gleeson, A., & Hanna, D. (2022). A meta-analysis of the efficacy of virtual reality and in vivo exposure therapy as psychological interventions for public speaking anxiety. Behavior Modification. doi:10.1177/0145445521991102
- Schreiber, L. M., Paul, G. D., & Shibley, L. R. (2012). The development and test of the Public Speaking Competence Rubric. Communication Education, 61(3), 205–233.
- Smith, M. J., Sherwood, K., Ross, B., et al. (2021). Virtual interview training for autistic transition age youth: A randomized controlled feasibility and effectiveness trial. Autism, 25(6), 1536–1552. doi:10.1177/1362361321989928