Exploratory VR study links studio layouts to comfort and openness

An exploratory immersive virtual reality experiment found that 50 young adults perceived four studio apartment layouts differently, with significant variation in spatial comfort and spatial openness despite comparable floor area, furnishings, materials, colors, lighting and window openings. The within-participant study also identified integrated spatial comfort, combining impressions of familiarity, order, pleasantness and spaciousness, as the only perceptual dimension significantly associated with broader ratings of housing comfort and housing preference. The results describe associations within the experiment and do not show that any single architectural feature causes a particular response.

Researchers initially recruited 60 adults in their twenties, then retained data from 50 after excluding six participants whose movement or technical interruptions prevented stable exposure to the virtual environments and four who withdrew because of headset discomfort. The final sample included 25 men and 25 women, with a mean age of 25 years and a standard deviation of 2.1 years. Each participant viewed and evaluated all four layouts, producing 200 layout-level ratings while the participant-level sample remained 50. No a priori power analysis was conducted before recruitment, and the authors characterized the investigation as exploratory rather than confirmatory.

The virtual apartments were presented through an Oculus Rift head-mounted display with a resolution of 1,080 by 1,200 pixels per eye, a 90-hertz refresh rate and an approximate 110-degree field of view. The environments were developed in Unreal Engine 4 and rendered on an Alienware M51 laptop. Each exposure lasted about one minute, and participants navigated with a joystick along a predefined path rather than physically walking, providing more consistent viewing conditions. The four layouts appeared in randomized order to reduce carryover effects. Immediately after each virtual visit, participants completed a semantic-differential questionnaire, and after all four exposures they rated overall housing comfort and housing preference. Each person received $10 compensation.

The four plans varied in entrance orientation, whether the kitchen and bathroom shared a wall, their relative alignment and the sequence in which the two service spaces were encountered from the entrance. Plan 1 combined a short-side entry with adjacent, horizontally arranged service spaces and a bathroom-then-kitchen sequence. Plan 2 used a short-side entry with separated, vertically arranged spaces and the same sequence. Plan 3 had a long-side entry, adjacent horizontal spaces and a bathroom-then-kitchen sequence. Plan 4 had a long-side entry, separated vertical spaces and a kitchen-then-bathroom sequence. Because these properties were tested as bundled combinations within complete layouts, the experiment cannot isolate the independent effect of any single characteristic.

To direct attention toward spatial organization, the researchers kept several non-configurational features constant across the virtual environments. These included total floor area, room dimensions, furniture type and arrangement, material finishes, color scheme and lighting. Window openings were identical in size, number and type, although their feasible positions were constrained by each plan. Participants rated the environments on seven-point scales defined by bipolar adjective pairs, including comfortable–uncomfortable, spacious–narrow, open–close, pleasant–unpleasant, familiar–unfamiliar, modern–classical and dynamic–static. The instrument initially contained 24 pairs. Item screening removed a duplicated light–heavy pair and the strong–weak pair, which had a negative item-total correlation, leaving 22 pairs for the principal analyses.

The retained items had an overall Cronbach’s alpha of 0.899, indicating high internal consistency for the questionnaire as a whole. That statistic does not establish that the heterogeneous descriptors measure a single underlying construct. The researchers therefore used exploratory factor analysis with principal axis factoring and oblimin rotation to examine potentially correlated perceptual dimensions. Sampling adequacy was high: the Kaiser–Meyer–Olkin value was 0.893, and Bartlett’s test of sphericity was significant, with chi-square equal to 2,233.956, 231 degrees of freedom and p less than 0.001. Four factors accounted for 50.9% of the total variance. Their reliability ranged from modest to high, with Cronbach’s alpha values from 0.594 to 0.899.

The largest factor, which explained 33.9% of the variance, was interpreted as spatial comfort. It included impressions such as good, familiar, comfortable, orderly, pleasant and spacious. The other factors were labeled spatial change, spatial aesthetics and spatial openness. Spatial change accounted for 7.8% of the variance and included variety, distinctiveness and dynamism. Spatial aesthetics represented 6.5% and drew on modern, elegant, sophisticated and luxurious impressions. Spatial openness explained 2.9% and included brightness, lightness, openness, freshness, warmth and softness. Because the factor solution was exploratory and depended on a relatively small sample, the item set and analytical choices, the authors described these as preliminary evaluative groupings rather than established universal psychological constructs.

A one-way analysis of variance showed that the layouts differed significantly in spatial comfort, with F equal to 8.253 and p less than 0.001, and in spatial openness, with F equal to 9.687 and p less than 0.001. Spatial change did not differ significantly, with F equal to 1.353 and p equal to 0.259, while spatial aesthetics showed no significant difference, with F equal to 0.418 and p equal to 0.740. Tukey comparisons distinguished Plan 2 from each of the other three layouts on comfort, with Plan 2 receiving the lower comfort ratings. Plan 3 was distinguished from all three alternatives on openness and received the lower openness ratings. These are relative responses under standardized virtual conditions, not estimates of how the plans would be judged during long-term occupation.

Exploratory multiple regressions examined whether the four perceptual dimensions were associated with the participants’ broader housing evaluations. The housing-comfort model was statistically significant, with F equal to 12.113, 195 residual degrees of freedom and p less than 0.001. It explained 19.9% of the variance in housing comfort, with an adjusted R-squared of 0.183. Spatial comfort was the only significant factor, with a standardized coefficient of 0.351 and p less than 0.001. The housing-preference model was also significant, with F equal to 17.157, 195 residual degrees of freedom and p less than 0.001, and explained 26.0% of the variance, with an adjusted R-squared of 0.245. Spatial comfort was again the sole significant factor, with a standardized coefficient of 0.441 and p less than 0.001. Spatial change, spatial aesthetics and spatial openness were not independently associated with either outcome when all four dimensions were included together.

The results separate the ability of a perceptual dimension to distinguish layouts from its relationship to broader housing evaluation. Spatial openness varied across the plans, but it did not independently track housing comfort or preference in the adjusted-for-other-factors sense used by the regression models. Spatial comfort incorporated positive emotion together with familiarity, order, legibility and perceived spatial adequacy, and it was more closely connected to the overall housing judgments. The regressions nevertheless remain unadjusted exploratory associations: the models did not include dwelling type or prior residential history, and they did not explicitly account for dependence among the repeated ratings supplied by each participant.

The experiment offers a controlled way to compare complete compact-housing alternatives while holding many non-layout features constant, an approach that may be useful during early-stage design evaluation. It also complements separate physiological work on the same four configurations, which examined EEG and ECG responses and reported that the layouts were physiologically distinguishable. Such physiological measures can capture embodied activity but do not directly state whether someone regarded a space as familiar, pleasant, spacious or open; the semantic ratings address that interpretive dimension. VR remains an imperfect substitute for physical occupation, and the study is limited by its sample of young adults, its South Korean studio-housing context, the bundled architectural variables and the short exposure period. Larger independent samples, fully crossed experiments and comparisons with occupied dwellings would be needed before feature-level design rules or broader generalizations could be developed.

Subject of Research: Perceptual responses to compact studio apartment layouts in immersive virtual reality

Article Title: Immersive VR-based sensibility evaluation: capturing emotional responses to studio apartment layouts

Article References: Myung, J., & Gero, J. (2026). Immersive VR-based sensibility evaluation: capturing emotional responses to studio apartment layouts. Frontiers in Virtual Reality, 7, Article 1795411. https://doi.org/10.3389/frvir.2026.1795411

Image Credits: AI Generated

DOI: 10.3389/frvir.2026.1795411

Keywords: immersive virtual reality, studio apartments, compact housing, spatial comfort, spatial openness, semantic differential, architectural evaluation, housing preference, environmental psychology, virtual environments, Immersive, VR-based