AR and VR in Education: Beyond the Hype
Headsets in classrooms generate headlines, but the research tells a more nuanced story — one where immersive technology transforms some kinds of learning dramatically and struggles with others entirely.
In 2016, Google Expeditions sent 2 million students on virtual field trips in a single week. Teachers reported that children who had never shown interest in science were pressing their faces against cardboard viewers to stand on the surface of Mars. For a brief, electric moment, it seemed that virtual reality would reshape education the way the printing press had reshaped literacy.
That moment stalled. Google shuttered the Expeditions programme in 2021. Headsets gathered dust in storage cupboards. Critics declared VR in education another expensive fad. Yet dismissing immersive technology entirely would be the wrong lesson to draw. The question that matters is not whether VR and AR work in education — it is when they work, and why, and for whom.
The Embodied Cognition Argument
The theoretical case for immersive learning rests on a well-established body of cognitive science. Embodied cognition research, developed across decades by researchers at institutions including the Max Planck Institute and the University of Edinburgh, holds that understanding is not purely a function of abstract symbol manipulation. The body's involvement in an experience — its posture, its movement, its sensory engagement — shapes how knowledge is encoded and retrieved.
Traditional classroom instruction is remarkably disembodied. A student learning about the human circulatory system reads text, studies diagrams, maybe watches a video. Virtual reality changes the phenomenology entirely. In a well-designed VR experience, you do not look at the heart — you stand inside it, watching valves the size of garage doors open and close, feeling the conceptual rhythm of the thing rather than memorising its labelled parts.
A 2019 meta-analysis published in the British Journal of Educational Technology, examining 38 controlled studies of VR learning environments, found a medium-to-large effect size for spatial understanding tasks and procedural learning. The effect was smaller and less consistent for declarative knowledge — facts and dates and vocabulary. This distinction turns out to be the key to understanding where immersive technology earns its cost and where it does not.
Where the Evidence Is Strongest: Spatial and Procedural Tasks
Medical training was among the earliest serious applications of VR, and the evidence for it is compelling. A landmark Stanford study found that surgical residents who trained using VR simulation were 29% faster in the operating room and made 6 times fewer errors than residents who had trained using traditional methods. The spatial demands of surgery — understanding three-dimensional anatomical relationships, developing fine motor calibration — map perfectly onto what immersive environments do well.
Architecture and engineering students show similar gains. Designing a building in CAD software gives you a floor plan; walking through a VR model of the same building before it is built lets you understand ceiling heights, circulation flows, and the way natural light moves through a space. Firms including Autodesk and Trimble have built entire training platforms around this insight, with measurable reductions in design errors during construction.
Aviation has used flight simulation — the grandfather of VR training — for decades. The International Air Transport Association estimates that roughly 80% of commercial pilots' initial qualification training can now be conducted in advanced simulators without compromising safety standards. The savings are substantial: an hour in a full-flight simulator costs roughly $500; an hour in an actual aircraft costs several thousand dollars, depending on type.
Augmented Reality: Lower Barrier, Different Affordances
Augmented reality overlays digital information onto the physical world rather than replacing it entirely. The technical barrier is lower — a smartphone or tablet suffices where VR requires a dedicated headset — and the pedagogical affordances are distinct. AR does not immerse; it annotates.
This annotation function turns out to be extraordinarily useful for maintenance, repair, and assembly tasks. Boeing has deployed AR headsets in its aircraft manufacturing facilities since 2018. Technicians installing wiring harnesses — a task that requires following a precise sequence through thousands of possible connections — used to consult paper manuals or laptop screens. With AR, the correct wire and connection point are highlighted in their field of view. Error rates dropped by 25%; installation time fell by a third.
The same principle applies in educational settings wherever procedure matters. Medical students learning to insert intravenous lines, engineering students learning to assemble circuit boards, chemistry students learning to operate unfamiliar apparatus — in each case, AR's ability to overlay instruction onto the actual physical task, in real time, addresses the gap between knowing the steps and being able to execute them.
"The question isn't whether VR is more engaging than a textbook — of course it is. The question is whether that engagement translates into durable learning. And the answer is: it depends entirely on whether the design of the experience aligns with the cognitive demands of the content."
— Jeremy Bailenson, Founding Director, Stanford Virtual Human Interaction Lab
The Failure Modes: When Immersion Becomes an Obstacle
The failures are as instructive as the successes. Several large-scale deployments of VR in K-12 settings have produced disappointing outcomes, and the common thread is what educational psychologists call extraneous cognitive load. When students are navigating an unfamiliar interface, managing their physical movements in a headset, and attempting to learn content simultaneously, the cognitive resources available for actual learning are depleted. The novelty effect — which inflates short-term engagement metrics — masks this load during initial use and disappears within a few sessions.
Language learning provides a cautionary case. Several startups raised substantial funding on the premise that VR conversation practice would accelerate second-language acquisition. The hypothesis was reasonable: immersive simulated environments would reduce anxiety and create authentic communicative contexts. The results were mixed at best. Speaking a language well requires rapid lexical retrieval and grammatical automaticity — skills built through repeated, low-stakes practice, not through the novelty of a headset. Students learned to navigate VR environments; they did not learn to speak faster.
There are also equity concerns that the industry has been slow to address. High-quality VR headsets remain expensive. Schools in under-resourced areas cannot deploy them at scale. Introducing immersive technology in well-funded schools while it remains inaccessible elsewhere risks widening the educational technology gap rather than narrowing it.
The Design Imperative
What separates effective immersive learning experiences from expensive novelties is almost always the quality of the pedagogical design, not the sophistication of the technology. The most rigorous work in this space comes from researchers like Richard Mayer at UC Santa Barbara, whose cognitive theory of multimedia learning provides a framework for understanding when visual and interactive elements enhance comprehension and when they distract from it.
Effective VR and AR learning experiences tend to share several characteristics. They make the unique affordances of immersion do real pedagogical work — if the content could be learned just as well from a textbook, there is no reason to add the complexity of a headset. They scaffold the experience so that learners are not overwhelmed by navigating a new interface while also processing new content. They embed retrieval practice and reflection rather than treating the immersive experience as an endpoint in itself. And they are designed for the specific cognitive demands of the specific content being taught.
The University of Maryland's Human-Computer Interaction Lab has produced some of the most careful work in this area, demonstrating through controlled studies that spatial memory — where you are in a virtual space when you encounter information — can serve as a powerful encoding cue. Students who learned a list of words while virtually navigating through a series of rooms could recall them significantly better when asked to mentally retrace their route. This is a twenty-first-century application of the method of loci, the memory technique used by ancient Greek orators. The technology is new; the underlying cognitive principle is millennia old.
What the Next Decade Will Bring
The trajectory of immersive learning technology depends on developments in several overlapping areas. Hardware costs are falling. The Meta Quest 3 costs roughly a third of what a comparable headset cost five years ago, and industry analysts expect further reductions as production scales. Wireless, untethered headsets with higher-resolution displays and more accurate positional tracking are becoming the norm rather than the exception.
Artificial intelligence is beginning to enable adaptive immersive experiences — VR environments that respond to a learner's performance, adjusting difficulty and focus in real time. This matters enormously because one of the consistent findings of learning science is that personalisation improves outcomes, and traditional classroom instruction is inherently one-size-fits-all.
Remote and hybrid learning, accelerated by the pandemic and unlikely to fully reverse, creates new demand for immersive alternatives to physical presence. A student who cannot attend a laboratory session can now, with increasing fidelity, conduct a virtual version. The fidelity is not yet identical — handling a physical beaker and rotating a virtual one involve different sensorimotor experiences — but the gap is closing.
What is unlikely to change is the fundamental insight that technology serves learning rather than replacing it. The most effective deployments of VR and AR in education are those designed by people who understand both the cognitive science of how people learn and the specific capabilities of immersive media. Where those two kinds of expertise come together, the results can be genuinely transformative. Where they do not, the headsets end up in the storage cupboard, and the money ends up looking for a new hype cycle to fund.