Immersion and presence are different constructs, and only one of them is measured directly

The claim that a more immersive system produces a stronger sense of presence is the founding assumption of immersive interface design. It is an assumption rather than a finding, because the two terms it joins are not the same kind of thing. Immersion names what a system provides. Presence names what a person does in response. The most direct access to presence remains the person's own report (Witmer and Singer, 1998). This first of ten units settles a vocabulary the nine that follow inherit.

The distinction carries an unwelcome consequence. A study reporting higher presence in a more immersive system has measured a self-report response to a comparison as well as an experience of a place. The field's default instruments cannot separate the two (Usoh, Catena, Arman and Slater, 2000; Skarbez, Brooks and Whitton, 2021). Presence is not unreal, because people demonstrably respond to rendered places as places. The claim is narrower: a presence difference cannot be attributed to the immersion difference the designer introduced.

The founding paper separated two words the market sells together

Slater and Wilbur (1997) set the vocabulary in the paper the field still cites. They proposed that immersion can be assessed objectively as a characteristic of the technology. The properties are how inclusive, extensive, surrounding and vivid the illusion is. Further properties are how far the tracked body matches the participant's movements and whether the environment responds autonomously. Presence they placed elsewhere, as a state of consciousness in which a person has a sense of being in a place.

The dimension list ranks two systems differently, which is why immersion is a profile rather than a score. Slater (2009) describes a Cave as a room whose four to six walls are back-projected stereo screens, and a head-mounted display as two small screens close to the eyes, updated as the head moves. The Cave surrounds the participant but is not inclusive, because looking far enough to one side brings the real room into view, while the headset shuts it out entirely.

The default instrument measures a state, and it nearly failed its own test

Witmer and Singer (1998) built the instrument most later work uses, the Presence Questionnaire, paired with an Immersive Tendencies Questionnaire for the differences a participant brings. They treat presence as a normal awareness phenomenon requiring directed attention. Across four experiments they reported high internal consistency, a weak but consistent positive relation between presence and task performance, and less presence among participants reporting more simulator sickness. Usoh, Catena, Arman and Slater (2000) then ran the test any measure of being in a place ought to pass. Ten participants searched for a box in a real office, ten in a virtual replica, and both groups completed two presence questionnaires in randomised order. Only the Slater, Usoh and Steed questionnaire produced a marginally higher mean for the real room. The Witmer and Singer questionnaire showed no significant difference at all. The authors concluded that such instruments may be useful within one environment and are doubtful for comparing across environments.

What the pooled evidence says, and what it cannot

Cummings and Bailenson (2015) aggregated 115 effect sizes from 83 studies of immersive technology and reported presence. Technological immersion had a medium-sized effect, with user tracking, stereoscopic visuals and wider fields of view more impactful than most other features. The effect belongs to particular manipulations rather than to immersion in general.

A pooled effect is not a licence to interpret any single study, whose presence difference may reflect the manipulation, the content or the instrument. Steed and colleagues (2016) collected presence and embodiment data from consumers on their own headsets, each seeing one of eight variants of a bar scenario. In-app responses gave tentative evidence that a self-avatar raised self-reported presence and embodiment, and that being invited to tap along lowered reported embodiment, and the authors' own word for the findings is tentative.

When the body answers a different question

Meehan, Insko, Whitton and Brooks (2002) looked for a measure that does not depend on the participant's account. They compared reactions to a non-threatening virtual room with reactions to a stressful virtual height across three experiments. Heart rate met their criteria for a presence measure, skin conductance less so and skin temperature not at all. Two further results are awkward for the argument the field wants. Adding a passive haptic element significantly increased presence, and presence rose with frame rate, from 15 to 20 to 30 frames per second.

Slater (2009) explains why the response rather than the rating became the preferred evidence. Drawing on a long series of experiments, he separated two orthogonal components. Place illusion is the sensation of being in the rendered place, constrained by the sensorimotor contingencies the system affords. Plausibility illusion is the sensation that depicted events are really occurring, set by whether events relate to the participant and stay credible. Participants know they are not in the place, yet what they do is a response to a place.

The objective measure has to be interpreted too

Skarbez, Brooks and Whitton (2021) tested the instruments against a manipulation that separates the system from the content. In a stressful virtual visual cliff they crossed high and low immersion with high and low coherence. Coherence is how far the experience matches the participant's expectations. They collected questionnaire and physiological data. Heart rate responded markedly differently across coherence conditions, while showing no difference across immersion conditions.

The measurement problem survives the move to physiology. A bodily response answers the experience as a whole, and it may track the credibility of the scenario rather than the capability of the display. Bahavan and colleagues (2023) recorded electroencephalography from twenty participants experiencing scenarios at varying levels of immersion and coherence, validated against questionnaires presented inside the virtual environment. Raising place illusion or plausibility illusion raised presence, and envelope amplitude correlation and spectral coherence discriminated presence, place illusion and plausibility illusion. It is a direction rather than a settled measure, since its validation runs through the self-report it is meant to replace.

The contested question: are presence scores commensurable?

The literature does not agree, and the disagreement has three positions rather than two. Witmer and Singer (1998) hold that presence is measurable with reliable subscales, and Usoh and colleagues (2000) leave that standing within a single environment, where a questionnaire can rank experiences of the same place.

Usoh and colleagues (2000) also supply the opposite position, which Skarbez, Brooks and Whitton (2017) generalise. Surveying the definitions, measures and models of presence and its neighbours, they report no widespread agreement on how to define or operationalise it after decades of research. Skarbez, Brooks and Whitton (2021) then demonstrate a concrete failure, questionnaires unable to separate immersion from coherence. On this reading a presence figure is local to the instrument, the environment and the task. Tran, Langlotz and Regenbrecht (2024) surveyed more than 320 mixed reality studies and found unvalidated or custom questionnaires commonplace, which they warn could bring a replication crisis.

A third position holds that objective measures should corroborate self-report rather than replace it. Meehan and colleagues (2002) built one such measure. Bahavan and colleagues (2023) extend the approach with a biosignal.

Evidence that would settle the dispute can be specified. An instrument would have to score a real environment above a faithful virtual replica, the test Usoh and colleagues attempted and only partly passed. It would have to separate the effect of a system manipulation from the effect of content and expectation, which Skarbez, Brooks and Whitton (2021) found several instruments failing. It would also have to predict a behavioural or physiological response in the same participants, which Bahavan and colleagues (2023) begin to attempt. Until then, a presence difference between two systems is evidence about responses, and the experience it is taken to reveal remains an inference.

The unit's own title carries a counter-argument, because it treats the questionnaire as the direct measure of presence; on the behavioural account an unasked response is measured directly and a rating is an inference about one's own state. The commensurability claim meets a second counter, because Cummings and Bailenson (2015) pooled 115 effect sizes from different instruments, though an average direction can be stable while the scales are not comparable.

What a designer should do while the dispute stands. Report the instrument, the environment and the task with any presence figure, and treat a presence difference as a ranking within one environment, not a cross-system metric. Name who answered it, since a questionnaire assumes a reader of the language it is written in and a second-language reader needs a spoken or translated version. Corroborate it with a behavioural or physiological measure from the same session, and otherwise design for the plausible and the coherent.

Where this is heading

Skarbez, Brooks and Whitton (2021) propose coherence as the construct that completes the picture, grouping the field's open questions under immersion, coherence, place illusion, plausibility illusion and presence. Work is moving towards instruments that predict a person's response before it is reported, and towards designs in which credibility and coherence, rather than raw fidelity, are the variables a designer manipulates.

Sources

• Slater, M. and Wilbur, S. (1997). A Framework for Immersive Virtual Environments (FIVE): Speculations on the Role of Presence in Virtual Environments. Presence: Teleoperators and Virtual Environments. doi:10.1162/pres.1997.6.6.603 Proposed that immersion can be assessed objectively as a characteristic of a technology, in terms of how inclusive, extensive, surrounding and vivid its illusion is, while presence is a state of consciousness related to being in a place. It is the founding statement of the asymmetry this unit rests on.
• Witmer, B. G. and Singer, M. J. (1998). Measuring Presence in Virtual Environments: A Presence Questionnaire. Presence: Teleoperators and Virtual Environments. doi:10.1162/105474698565686 Reported the Presence Questionnaire and an Immersive Tendencies Questionnaire, finding high internal consistency across four experiments but only a weak positive relation between presence and task performance, and less presence among those reporting more simulator sickness. It is the instrument later critiques attack.
• Usoh, M., Catena, E., Arman, S. and Slater, M. (2000). Using Presence Questionnaires in Reality. Presence: Teleoperators and Virtual Environments. doi:10.1162/105474600566989 Had ten participants search for a box in a real office and ten in a virtual replica: only the Slater, Usoh and Steed instrument scored the real room marginally higher, the Witmer and Singer instrument showed no significant difference, and the authors judged both doubtful for comparing across environments. That is the comparison a study of immersion wants to make.
• Meehan, M., Insko, B., Whitton, M. and Brooks, F. P. (2002). Physiological measures of presence in stressful virtual environments. ACM Transactions on Graphics. doi:10.1145/566654.566630 Compared physiological reactions to a non-threatening virtual room with reactions to a stressful virtual height, finding heart rate the most usable presence measure, skin conductance less so, skin temperature not at all, and passive haptics raising presence. It is the founding case for objective measurement.
• Slater, M. (2009). Place illusion and plausibility can lead to realistic behaviour in immersive virtual environments. Philosophical Transactions of the Royal Society B. doi:10.1098/rstb.2009.0138 Argued from a long series of experiments that realistic behaviour in an immersive system depends on two orthogonal components: place illusion, constrained by the sensorimotor contingencies the system affords, and plausibility illusion, set by whether events relate to the participant and stay credible. It relocates the evidence for presence from the rating to the response.
• Cummings, J. J. and Bailenson, J. N. (2015). How Immersive Is Enough? A Meta-Analysis of the Effect of Immersive Technology on User Presence. Media Psychology. doi:10.1080/15213269.2015.1015740 Aggregated 115 effect sizes from 83 studies and found a medium-sized effect of technological immersion on presence, with user tracking, stereoscopic visuals and wider fields of view more impactful than most other features, including content quality. It is the strongest quantitative support for immersion as a cause of presence.
• Steed, A., Friston, S., Lopez, M. M., Drummond, J., Pan, Y. and Swapp, D. (2016). An 'In the Wild' Experiment on Presence and Embodiment using Consumer Virtual Reality Equipment. IEEE Transactions on Visualization and Computer Graphics. doi:10.1109/tvcg.2016.2518135 Collected presence and embodiment data from consumers using their own headsets in uncontrolled settings across eight variants of a bar scenario, finding tentative evidence that a self-avatar raised self-reported presence and that being invited to tap along lowered self-reported embodiment. It is the strongest answer to the charge that laboratory presence results do not travel.
• Skarbez, R., Brooks, Jr., F. P. and Whitton, M. C. (2017). A Survey of Presence and Related Concepts. ACM Computing Surveys. doi:10.1145/3134301 Reviewed the definitions, measures and models of presence and of the neighbouring constructs, reporting that decades of research have produced no widespread agreement on how to define or operationalise the construct. It is the systematic statement of the measurement problem.
• Skarbez, R., Brooks, F. P. and Whitton, M. C. (2021). Immersion and Coherence: Research Agenda and Early Results. IEEE Transactions on Visualization and Computer Graphics. doi:10.1109/tvcg.2020.2983701 Crossed high and low immersion with high and low coherence in a stressful virtual visual cliff and collected questionnaire and physiological data: several existing presence questionnaires could not reliably separate the effects of immersion from those of coherence, and heart rate differed markedly across coherence conditions but not across immersion conditions. It shows a measure tracking the content of an experience rather than the capability of a system.
• Bahavan, T. T. N., Navaratnarajah, S., Owinda, D., Akalanka, I., Peiris, R. and Silva, A. D. (2023). Towards an objective measurement of presence, place illusion, and plausibility illusion in virtual reality using electroencephalography. Virtual Reality. doi:10.1007/s10055-023-00815-x Recorded electroencephalography from twenty participants experiencing scenarios at varying levels of immersion and coherence, validated against questionnaires administered inside the virtual environment, and found envelope amplitude correlation and spectral coherence discriminating presence, place illusion and plausibility illusion. It is the most recent attempt at an objective measure, on a sample small enough to read as a direction rather than an instrument.
• Tran, T. Q., Langlotz, T. and Regenbrecht, H. (2024). A Survey On Measuring Presence in Mixed Reality. Proceedings of the CHI Conference on Human Factors in Computing Systems. doi:10.1145/3613904.3642383 Analysed more than 320 mixed reality studies reporting a presence measurement, finding unvalidated and custom instruments commonplace and warning that the loss of comparability could contribute to a looming replication crisis.