Does VR Safety Training Actually Work? What the Research Shows (2026)

Last Updated: October 12, 2026

VR safety training works. Independent research from universities, Fortune 100 companies, and peer-reviewed journals consistently shows measurable improvements in knowledge retention, hazard recognition, and post-training confidence compared to classroom-only methods. A 2025 quasi-experimental study of 200 industrial workers by Qawqzeh and colleagues, published in Scientific Reports, found the VR-trained group scored 30% higher on safety awareness than the classroom control group. PwC reported that VR learners completed training 4x faster and were 275% more confident applying what they learned. Whether those gains justify the cost depends on your workforce size, your injury rates, and how you measure “working.”

What do the systematic reviews and meta-analyses show?

Two systematic reviews pooled the VR safety training literature and landed in the same place from different directions. VR outperforms traditional training on knowledge acquisition and knowledge retention, and the evidence base is thinner than most vendor decks imply. Scorgie and colleagues reviewed 52 articles and ran two meta-analyses. Stefan, Mortimer and Horan reviewed 136 studies. Read together, they establish that VR wins on what people learn, and that hardly anyone has measured whether it changes what people do once they are back on the floor.

The meta-analysis result

Scorgie, Feng, Paes, Parisi, Yiu and Lovreglio published their review in Safety Science in 2023. They screened the VR safety training literature from 2013 to 2021 outside medicine and the military, ending with 52 articles across 14 industry domains. Construction and fire safety were the two most common since 2018. Their two meta-analyses concluded that VR safety training outperforms traditional training on both knowledge acquisition and knowledge retention. That is the single most cited quantitative statement available on this question, and it comes from a pooled analysis rather than one vendor’s pilot.

Two findings in the same paper cut the other way, and they are worth knowing before you quote the headline. Only 36% of the studies reported long-term retention data. Only 9.6% explicitly grounded their VR application in a learning theory. The field is measuring the right outcome less than half the time.

What the literature does not establish

Stefan, Mortimer and Horan reviewed 136 VR safety training evaluations published between 2016 and August 2021 for the journal Virtual Reality, then sorted every outcome measure into Kirkpatrick’s four levels of training evaluation. The distribution is lopsided: 72.06% of studies evaluated learning and 66.18% evaluated reaction, while very few evaluated behaviour or results. In plain terms, the research tells you trainees score better on the test and like the training more. It largely does not tell you their behaviour changed at the job site, and it does not tell you the incident rate moved.

Humulo’s position on this: anyone selling you VR on the strength of an injury-reduction percentage is extrapolating past what the peer-reviewed literature currently supports. Ask for the measurement plan instead. If a vendor cannot describe how you would detect a behaviour change at 90 days, the number in their deck is someone else’s case study.

The newest fire-specific evidence

Muffato, Mazzella di Bosco, Zuzzi, Pellegrini and Meneghetti published a fire extinguisher training experiment in Frontiers in Psychology in 2026 that compared three conditions: passive video-slide learning, basic VR with avatar instruction, and dual-mode VR adding informational panels. With 111 participants, the basic VR group made 25.2% fewer errors than the video-slide group and completed the procedure 38.1% faster. The dual-mode group was 33.0% faster. Both VR conditions scored significantly higher overall. Higher measured stress was associated with a 2.02% increase in errors per unit, which is the kind of detail that matters when the real event is not calm.

A second sample of 16 working adults showed no significant difference from the student group, which is the part most fire training studies skip. It suggests the result travels from a university lab to a workforce. This is the strongest recent independent support for VR fire extinguisher training as a procedural method, and it was published by researchers with no commercial stake in any platform.

Evidence at a glance

The full methodology and results of the Central Washington University study are published separately, including what it does not establish.

What does the peer-reviewed research actually say?

Most of the skepticism around VR training comes from a fair question: is the evidence solid, or is it vendor marketing dressed up as science? Here is what holds up to scrutiny.

A quasi-experimental study by Qawqzeh and colleagues, published in Scientific Reports in August 2025 (DOI 10.1038/s41598-025-14173-y), divided 200 industrial workers into VR-trained and classroom-trained groups. The VR group showed a statistically significant 30% increase in safety awareness scores, along with higher risk perception and self-efficacy ratings. The researchers used Mann-Whitney U tests, not just self-reported surveys, which matters if you care about methodology.

PwC ran a controlled study comparing VR, classroom, and e-learning for soft skills training. VR learners finished 4x faster than classroom learners. They were 275% more confident in applying their skills afterward. And they reported being 3.75x more emotionally connected to the content, which PwC measured through survey instruments designed to capture engagement depth.

At Central Washington University, Dr. Hongtao Dang and Dr. Jennifer Serne ran an efficacy study using Humulo’s VR safety training modules (later published in peer-reviewed ASCE proceedings). The study split participants from CWU’s Safety and Health Management Program into control and experimental groups. Results: 100% of VR participants said the technology improved their comprehension. 100% wanted VR included in future training. Quantitative testing at 30 days showed the VR group outperformed the classroom-only group on knowledge retention. The sample was small (37 participants total), but the direction was unambiguous.

What about large-scale, real-world deployments?

Lab studies are one thing. What happens when you roll VR training out to thousands of employees across actual facilities?

Walmart deployed VR training to over 1 million associates across its U.S. locations, working with Strivr. The numbers they reported: VR trainees scored higher on content retention tests 70% of the time. Satisfaction scores were 30% higher than standard training. And sessions that took 90 minutes in a classroom were completed in roughly 20 minutes through VR. Those efficiency numbers get a CFO’s attention fast.

Boeing used VR to train technicians on a 50-step cargo door seal procedure that only a handful of people in the company knew how to do. Training time dropped 75%. First-time task accuracy improved 33%. VR-trained workers made 40% fewer errors during assembly. When you are training on procedures where mistakes cost six figures, those margins matter.

Based on Humulo’s deployment data across enterprise clients, we see similar patterns in safety-specific applications. Forklift certification training that used to require a full day of classroom instruction plus floor time now takes roughly half that, and trainees score higher on post-assessment tests. The difference is most visible about 30 days out, when the retention gap between VR-trained and classroom-only workers widens.

How does VR training affect actual injury rates?

This is the question that matters most to EHS managers, and it is also the hardest to answer cleanly. Injury rates depend on dozens of variables beyond training method: workforce turnover, facility age, production pressure, reporting culture.

That said, the Bureau of Labor Statistics reported 2.5 million nonfatal workplace injuries and illnesses in the U.S. in 2024, along with 5,070 fatal work injuries. Those numbers have been trending down, but slowly. OSHA still estimates that employers pay roughly $1 billion per week in direct workers’ compensation costs alone.

Research compiled by the National Safety Council indicates a 32% reduction in workplace accidents at organizations using simulation-based training programs. That figure includes VR alongside other simulation methods, so it is not a pure VR metric. But the direction is consistent with what individual companies report after adopting immersive training.

The mechanism makes intuitive sense if you think about how adults learn. Passive lectures produce low retention. The National Training Laboratories’ learning model suggests that practice-based methods outperform passive ones by wide margins. VR puts trainees inside a scenario where they have to act, make decisions, and experience consequences without anyone getting hurt. That is hard to replicate with a PowerPoint deck and a sign-off sheet.

Does OSHA actually accept VR as compliant training?

Yes, with caveats. OSHA’s August 10, 2020 letter of interpretation (DOL-OSHA-DEP-2020-007) addressed this directly. The short version: OSHA evaluates training effectiveness on a case-by-case basis, regardless of delivery method. VR training can satisfy OSHA requirements as long as the training covers all required content elements for the specific standard in question.

The sticking point is hands-on requirements. Some OSHA standards, like 29 CFR 1910.134 for respiratory protection or 29 CFR 1910.147 for lockout/tagout, require hands-on demonstration of competency. VR can supplement these requirements, but most compliance officers will still want to see trainees physically perform the procedure on actual equipment.

The practical approach most of our clients take: use VR for initial training, familiarization, and refresher cycles, then require a shorter hands-on assessment with real equipment. This cuts total training time significantly while keeping you on solid regulatory ground. You can read more about this approach in our OSHA-compliant VR training guide.

How does VR compare to classroom and e-learning on specific metrics?

The PwC study is the cleanest head-to-head comparison available because it tested all three modalities with the same content. Here is how they stacked up:

Source: PwC, “The Effectiveness of Virtual Reality Soft Skills Training in the Enterprise,” 2020.

That cost parity number at 375 learners is worth lingering on. If you are training fewer than 375 people total, classroom training may still be cheaper. But most manufacturing and warehousing operations cycle through hundreds or thousands of employees annually for safety refreshers. At that scale, VR’s per-session cost drops significantly. For a deeper comparison, see our VR vs. classroom safety training analysis. For a detailed breakdown of VR training versus traditional e-learning on cost, retention, and ROI, see our VR safety training vs e-learning comparison.

What types of safety training work best in VR?

Not all training topics benefit equally from VR. The technology excels when the real-world scenario is dangerous, expensive to replicate, or rare enough that trainees cannot get adequate practice through normal operations.

Humulo’s CWU study results were particularly strong for the high-advantage categories. The study used fire extinguisher training modules, and the VR group’s comprehension gains held up at the 30-day retest, which suggests the immersive format is especially effective for procedural skills that trainees rarely use in real emergencies. Explore our full VR safety training module lineup to see what is available for each topic.

What does VR safety training actually cost?

Pricing varies widely depending on hardware, content licensing, and whether you build custom modules or use off-the-shelf scenarios. Here is a realistic range based on current market pricing:

Compare that to the average cost of a single OSHA recordable injury: the National Safety Council puts direct costs at roughly $42,000 per medically consulted injury, and the indirect costs (lost productivity, retraining, investigation time) typically run 2-4x the direct cost. Preventing even a handful of recordable incidents per year can justify the entire VR training budget. For more on the ROI math, read our guide to training methods that improve retention.

Retention comparison: VR vs. other training methods

Putting the research together, here is how VR stacks up against common training delivery methods on the metrics that matter most for safety programs:

The most important column comparison for EHS managers: VR simulation matches hands-on training for retention (both 75% on the NTL pyramid) but eliminates hazard exposure during the training itself. For high-risk procedures like confined space entry (29 CFR 1910.146) or energized electrical work (29 CFR 1910.331-335), that trade-off is hard to argue against.

Why VR sits in the practice-by-doing tier

VR safety training sits squarely in the “practice by doing” tier at 75% retention. A worker wearing a headset is not watching a video of someone operating a forklift. They are operating the forklift. They are reaching for the fire extinguisher, pulling the pin, aiming at the base. Their hands move. Their body turns. They make decisions in real time and see consequences play out.

This is why VR outperforms lecture-based and e-learning approaches for procedural safety content. The learner is active, not passive. Edgar Dale’s original insight was straightforward: the closer you get to direct, purposeful experience, the more the learner retains. VR gets you close to direct experience without the risk of actual injury, chemical exposure, or equipment damage. For hazardous procedures governed by OSHA standards (29 CFR 1910.146 for confined space, 29 CFR 1910.157 for fire extinguisher use), that distinction matters. You want workers to practice the real procedure. You do not want them practicing it incorrectly on live equipment. For a full breakdown of costs, retention data, and ROI between these approaches, see our VR safety training vs e-learning comparison.

What are the actual limitations?

I would be doing you a disservice if I did not address the downsides honestly.

VR cannot fully replace hands-on training for physical skills. A forklift operator still needs seat time on a real forklift. A worker doing confined space entry still needs to physically enter a permit space under supervision. VR is a supplement that improves the classroom portion, not a wholesale replacement for everything.

Motion sickness remains a real issue for a subset of trainees, typically 5-15% depending on the content type. Stationary experiences (like fire extinguisher training) cause fewer problems than locomotion-heavy scenarios. Hardware has improved significantly, but you will still have some employees who cannot tolerate VR for more than a few minutes.

Content quality varies enormously between vendors. A poorly designed VR module is worse than a good classroom session. Look for vendors who can show you third-party efficacy data, not just flashy demos. The CWU study is one example of what independent validation looks like.

IT infrastructure requirements are real. Device management, firmware updates, content distribution, charging logistics. Plan for at least one person at each site who owns the VR training hardware.

Frequently asked questions

Does VR safety training actually improve knowledge retention?

Yes. Multiple studies confirm it. The Central Washington University study found VR-trained participants outperformed the classroom-only group on knowledge retention tests administered 30 days after training. PwC found VR learners were 275% more confident applying their skills. Walmart reported VR trainees scored higher on content retention tests 70% of the time. The retention advantage appears most significant for procedural and scenario-based training, less so for purely informational content.

For a deeper look at the research behind these numbers, see our VR safety training statistics reference page, which collects the 30-day and 90-day retention data across multiple studies.

Is VR training OSHA compliant?

OSHA’s 2020 letter of interpretation confirms that VR can be used for safety training, evaluated on a case-by-case basis. The training must cover all required content elements for the relevant standard (e.g., 29 CFR 1910.147 for LOTO, 29 CFR 1910.178 for powered industrial trucks). Where standards require hands-on demonstration, VR should supplement rather than replace physical practice.

How much does VR safety training cost per employee?

At scale, per-session costs typically range from $15-$50 after the initial hardware investment. PwC’s research found that VR training reaches cost parity with classroom training at around 375 learners. For organizations running hundreds of safety training sessions annually, VR usually costs less per trainee than recurring classroom instruction with instructors, facilities, and lost production time.

What evidence exists for VR training reducing workplace injuries?

The National Safety Council has reported 32% reductions in workplace accidents at organizations using simulation-based training. A 2025 Scientific Reports study found a 30% increase in safety awareness among VR-trained industrial workers. Boeing reported 40% fewer errors during assembly tasks after implementing VR training. Direct injury rate reduction is harder to isolate because workplace safety depends on many factors beyond training method.

Which industries benefit most from VR safety training?

Manufacturing, warehousing, construction, oil and gas, and utilities see the strongest results because their hazards are difficult, expensive, or dangerous to replicate in training environments. Government and military operations also see high adoption rates. Industries with primarily office-based or low-hazard environments get less value from VR and may be better served by e-learning for general safety awareness.

How much faster is VR training compared to classroom instruction?

PwC’s research found VR learners completed training 4x faster than classroom learners while achieving equal or better learning outcomes. For EHS managers, this means less time pulling workers off the floor for training sessions, which reduces the operational disruption that makes scheduling annual recertification so painful.

How long do VR training retention gains last compared to classroom training?

The Central Washington University efficacy study measured retention at 30 days post-training and found VR-trained participants maintained stronger recall of safety procedures than classroom-only participants. The emotional and motor encoding that VR provides resists the Ebbinghaus forgetting curve more effectively than lecture-based instruction, where learners lose up to 70% of content within 24 hours without reinforcement.

Is there a meta-analysis on VR safety training?

Yes. Scorgie and colleagues published a systematic review with two meta-analyses in Safety Science in 2023, covering 52 articles from 2013 to 2021 across 14 industry domains. Both meta-analyses found that VR safety training outperforms traditional training on knowledge acquisition and knowledge retention. A second systematic review by Stefan, Mortimer and Horan covered 136 studies published between 2016 and August 2021.

What does the research not prove about VR safety training?

It does not prove behaviour change or injury reduction. When Stefan, Mortimer and Horan sorted 136 VR safety training studies into Kirkpatrick’s four evaluation levels, 72.06% measured learning and 66.18% measured reaction, while very few measured behaviour or organisational results. Scorgie and colleagues separately found that only 36% of the studies they reviewed reported long-term retention data. The evidence supports better test scores and better retention. Claims about recordable rates are extrapolation, not findings.

Where to start

If you manage safety training for a manufacturing, warehousing, or industrial operation with 100+ employees, VR is worth a serious look. The evidence base is solid and growing. Start with one or two high-impact modules (forklift, fire extinguisher, or LOTO are common starting points), run a pilot with a single site, and measure before-and-after assessment scores.

You can schedule a demo of Humulo’s enterprise VR training platform to see what a deployment looks like for your specific training requirements. We will walk through your current training gaps, show you how the modules map to your OSHA obligations, and give you real numbers to take to your leadership team.