More immersion is not more learning

The claim that a more immersive lesson teaches more is the commercial argument for educational virtual reality, and it is the claim this second of ten units tests. The evidence does not support immersion as a general cause of better learning. It supports something narrower. Immersion reliably changes how a lesson feels. The pooled effect on learning is small and depends on what it is compared against. Well-controlled studies find the same people reporting more presence while learning less (Makransky, Terkildsen and Mayer, 2019).

A designer who believes that immersion causes learning buys a headset instead of a clearer explanation. A designer who believes that it changes attention, affect and load spends the budget on instructional support inside the lesson, and tests transfer rather than preference. A third spends it on getting every learner into the lesson, because a shared headset and a metered connection put part of the class outside it.

A fair comparison is hard to build, and the first one found nothing

An immersive lesson compared with a desktop lesson is not a controlled experiment on immersion. The conditions differ in the display, the interaction, the content, the duration, the novelty and the learner's expectation, so they cannot be matched on everything except the medium (Parong and Mayer, 2020).

Parong and Mayer (2020) show the problem inside a single study. They gave a biology lesson on the same topic to two groups, one as an interactive animated journey in immersive virtual reality and the other as a desktop slideshow. The immersive group did worse on transfer tests, reported more emotional arousal and more extraneous cognitive load, and showed less engagement on electroencephalography.

The oldest controlled comparison found nothing. Moreno and Mayer (2002) taught college students about botany through an agent-based game. They presented it on a desktop display, on a head-mounted display used while sitting, and on one used while walking. The head-mounted conditions raised presence ratings, but the medium did not change retention or transfer. What changed performance was the form of the explanation, with narration beating identical on-screen text. The studies claiming the opposite have not answered it.

The pooled effect is positive, and small

The most cited evidence for the claim is a meta-analysis rather than a study. Merchant and colleagues (2014) pooled experimental and quasi-experimental research on virtual reality instruction in schools and universities. They separated games, simulations and virtual worlds, and reported positive effects on learning outcomes, with games ahead of simulations and virtual worlds. The effect belongs to particular applications rather than to immersion as a property of a display.

Two later meta-analyses narrowed the question and shrank the answer. Wu, Yu and Gu (2020) restricted inclusion to comparisons between a head-mounted display and a less immersive condition: pooling 35 randomised or quasi-experimental studies gave a small effect size of 0.24, larger for school-age learners. Coban, Bolat and Goksu (2022) pooled 105 results from 48 primary studies covering 3,179 students, and reported a small overall effect of g = 0.38, larger against traditional instruction than against computer-assisted instruction. A small positive average permits the claim that immersion sometimes helps; it does not license applying a pooled figure to a single lesson.

Immersion buys motivation and presence, and can cost learning

Makransky and Lilleholt (2018) gave 104 university students a virtual science simulation in immersive or desktop virtual reality and modelled their responses. The immersive version scored higher on 11 of the 13 measured variables, with the largest differences on presence and on the emotional and motivational measures rather than on learning. Immersion bought motivation and presence, and a motivated learner is easily read as a better learner.

The cleanest demonstration is Makransky, Terkildsen and Mayer (2019). They added immersive virtual reality to a science laboratory simulation in a two-by-two design with 52 university students, and measured cognitive load with electroencephalography. The immersive condition produced substantially more presence, with an effect size of 1.30, and significantly less learning, with an effect size of 0.80. Cognitive load was also higher. Presence and learning moved in opposite directions in the same people.

Parong and Mayer (2018) reached the same conclusion from the other side. Students who studied a biology lesson from a desktop slideshow outperformed students who studied it in immersive virtual reality. The immersive group reported more motivation and interest. Summarising each section of the immersive lesson closed the gap. The lesson design, not the display, carried the problem.

The comparison condition decides the effect

Hamilton and colleagues (2020) reviewed immersive virtual reality studies reporting quantitative learning outcomes and coded their designs. Most found an advantage, but the gains were largest where the comparison was a conventional lecture rather than an equivalent non-immersive activity. Most interventions were short and few examined retention. The review judged the methods used to evaluate learning often inadequate.

Jensen and Konradsen (2017) reviewed 21 experimental studies of head-mounted displays, quality-assessed with the Medical Education Research Study Quality Instrument and found below average on it. They report that the displays help for spatial and visual knowledge, for observational and psychomotor skills and for affective skills such as managing a stressful situation, but that elsewhere the displays give no advantage and are sometimes counterproductive. Immersion may be a cause for a specified class of skill, and that narrowed claim is testable.

The contested question: is immersion a cause or a context?

One position holds that immersion causes learning. Merchant and colleagues (2014), Wu, Yu and Gu (2020) and Coban, Bolat and Goksu (2022) all report positive pooled effects. Makransky and Mayer (2022) argue for an immersion principle within multimedia learning. On this reading the effect is real but small, and the design task is to use immersion where it is strongest.

A second position holds that the pooled effect is largely an artefact of weak comparisons. Hamilton and colleagues (2020) show it shrinking when the non-immersive condition is a fair equivalent. Parong and Mayer (2018) and Makransky, Terkildsen and Mayer (2019) produce the reversal under controlled conditions. Jensen and Konradsen (2017) confine the advantage to particular skill classes.

A third position, where the evidence now points, holds that immersion is a context rather than a cause. Its immediate effect is on affect, attention and cognitive load (Makransky and Lilleholt, 2018; Parong and Mayer, 2020). Whether that helps learning depends on the instructional support built into the lesson (Petersen et al., 2023). The testable question is then not whether immersion improves learning but which instructional design makes an immersive lesson teach.

Evidence that would settle the dispute can be specified. A study would hold the instructional method constant across the two conditions, so that only the medium varies. It would measure transfer rather than retention or preference, and measure cognitive load and affect in the same participants. It would test again after a delay, because an immediate advantage may be novelty. It would report sickness and dropouts, because the people who leave are the ones the measure misses. Klingenberg, Bosse, Mayer and Makransky (2024) hold the method constant and measure transfer across a two-by-two design with 177 students, and still find no transfer effect. No study in this unit's list meets all five conditions at once.

The unit's thesis has a strong counter-argument of its own. Makransky and Mayer (2022) compared a virtual field trip in a head-mounted display with a 2D video for 102 middle-school students inside a six-lesson unit, and found the head-mounted group ahead on retention at both immediate and delayed tests. That design holds the instructional content constant and tests an outcome rather than a preference, so it meets most of the conditions listed above and still favours immersion. The thesis survives as a statement about the average and the uncontrolled comparison, not as a rule for a well-designed immersive lesson.

What a designer should do while the dispute is unresolved. Treat immersion as the setting for a lesson rather than as the lesson, and budget for the instructional support that carries the learning. Test transfer and delayed retention in your own context rather than importing a pooled effect.

Where this is heading

The field is moving from media comparison towards instructional design inside immersion. Petersen and colleagues (2023) prompt individual or collaborative generative activities inside an immersive biology lesson for 164 high-school students and find that the collaborative version raises learning. Makransky and Mayer (2022) read their own field-trip result as an immersion principle in multimedia learning. The direction is towards treating immersion as the setting for a lesson rather than as the lesson, so the next real advance is more likely to be a controlled study than a headset.

Sources

• Moreno, R. and Mayer, R. E. (2002). Learning science in virtual reality multimedia environments: role of methods and media. Journal of Educational Psychology. doi:10.1037/0022-0663.94.3.598 Compared botany lessons on a desktop and two head-mounted conditions: presence rose with the headsets, but retention and transfer did not, while narration beat on-screen text. It is the founding controlled comparison in which immersion adds nothing.
• Merchant, Z., Goetz, E. T., Cifuentes, L., Keeney-Kennicutt, W. and Davis, T. J. (2014). Effectiveness of virtual reality-based instruction on students' learning outcomes in K-12 and higher education: a meta-analysis. Computers and Education. doi:10.1016/j.compedu.2013.07.033 Pooled studies of virtual reality instruction across games, simulations and virtual worlds, reporting positive learning effects with games ahead. It is the most cited evidence for the claim.
• Jensen, L. and Konradsen, F. (2017). A review of the use of virtual reality head-mounted displays in education and training. Education and Information Technologies. doi:10.1007/s10639-017-9676-0 Reviewed 21 experimental studies, quality-assessed with the Medical Education Research Study Quality Instrument, finding the study quality below average on that instrument, especially for the studies that evaluated educational products. Reported benefits only for spatial and visual knowledge, observational and psychomotor skills and affective skills. It is the clearest statement that immersion helps for particular skill classes.
• Parong, J. and Mayer, R. E. (2018). Learning science in immersive virtual reality. Journal of Educational Psychology. doi:10.1037/edu0000241 Compared an immersive virtual reality biology lesson against a desktop slideshow, then tested adding a generative strategy. Students learned less from the immersive lesson, and summarising each section closed the gap. It is a well-controlled comparison in which the less immersive medium wins.
• Makransky, G. and Lilleholt, L. (2018). A structural equation modelling investigation of the emotional value of immersive virtual reality in education. Educational Technology Research and Development. doi:10.1007/s11423-018-9581-2 Administered a virtual science simulation to 104 university students in immersive or desktop virtual reality. The immersive version scored higher on 11 of 13 variables, with the largest differences on presence and motivation rather than learning. It names the substitution in which immersion buys motivation read as learning.
• Makransky, G., Terkildsen, T. S. and Mayer, R. E. (2019). Adding immersive virtual reality to a science lab simulation causes more presence but less learning. Learning and Instruction. doi:10.1016/j.learninstruc.2017.12.007 Added immersive virtual reality to a science lab simulation with 52 university students, measuring cognitive load with electroencephalography. The immersive condition produced more presence (effect size 1.30) and less learning (effect size 0.80) with higher load. It is the clearest demonstration that more presence can come with less learning.
• Wu, B., Yu, X. and Gu, X. (2020). Effectiveness of immersive virtual reality using head-mounted displays on learning performance: a meta-analysis. British Journal of Educational Technology. doi:10.1111/bjet.13023 Meta-analysed 35 randomised or quasi-experimental studies comparing head-mounted display virtual reality against less immersive conditions. Reported a small effect size of 0.24, larger for school-age learners than for university students. It is the meta-analytic counterweight that shows how small the effect is.
• Hamilton, D., McKechnie, J., Edgerton, E. and Wilson, C. (2020). Immersive virtual reality as a pedagogical tool in education: a systematic literature review of quantitative learning outcomes and experimental design. Journal of Computers in Education. doi:10.1007/s40692-020-00169-2 Reviewed studies reporting quantitative learning outcomes from immersive virtual reality and coded their designs. Found the field dominated by short interventions with weak comparisons and the gains largest against a conventional lecture. It shows that the pooled effect depends on what immersion is compared against.
• Parong, J. and Mayer, R. E. (2020). Cognitive and affective processes for learning science in immersive virtual reality. Journal of Computer Assisted Learning. doi:10.1111/jcal.12482 Tested whether positive affect or cognitive load explained the outcome of comparing immersive virtual reality with a desktop slideshow. They did worse on transfer, reported more arousal and extraneous load, and showed less engagement on electroencephalography. It turns the dispute into a mechanism question about affect and load.
• Coban, M., Bolat, Y. I. and Goksu, I. (2022). The potential of immersive virtual reality to enhance learning: a meta-analysis. Educational Research Review. doi:10.1016/j.edurev.2022.100452 Pooled 105 results from 48 primary studies covering 3,179 students. Reported a small overall effect of g = 0.38, larger against traditional instruction than against computer-assisted instruction. It is the most recent pooled estimate and the evidence that the headline effect shrinks once the comparison is fair.
• Makransky, G. and Mayer, R. E. (2022). Benefits of taking a virtual field trip in immersive virtual reality: evidence for the immersion principle in multimedia learning. Educational Psychology Review. doi:10.1007/s10648-022-09675-4 Had 102 middle-school students take a virtual field trip to Greenland through a head-mounted display or as a 2D video, within a six-lesson climate change unit, and found the head-mounted group higher on presence, enjoyment and retention at immediate and delayed tests. It is the strongest recent evidence that immersion can help inside an instructional sequence.
• Petersen, G. B., Stenberdt, V., Mayer, R. E. and Makransky, G. (2023). Collaborative generative learning activities in immersive virtual reality increase learning. Computers and Education. doi:10.1016/j.compedu.2023.104931 Had 164 high-school students view an immersive biology lesson, some prompted to carry out generative activities during breaks. The collaborative activity raised learning more than the individual activity, relative to no breaks. It shows that the fix for immersion's learning problem is instructional rather than technological.
• Klingenberg, S., Bosse, R., Mayer, R. E. and Makransky, G. (2024). Does embodiment in virtual reality boost learning transfer? Testing an immersion-interactivity framework. Educational Psychology Review. doi:10.1007/s10648-024-09956-0 Assigned 177 high school students to a two-by-two design crossing immersion, a headset against a screen, with interactivity, direct manipulation against a laser pointer. Immersion raised presence, agency and body ownership but moved no transfer measure, the newest controlled test of the unit's thesis.