Strachan-Regan and Baumann put 35 people in two virtual rooms in counterbalanced order, one curved and one rectangular, and measured mood, heart rate and creative output. Positive affect was 30.60 in the curved room against 23.94 in the rectangular one, a large difference at a Cohen’s d of 0.82. Negative affect ran the other way, 12.62 in the rectangular room against 11.51 in the curved, at a small d of 0.35. Heart rate was lower in the curved room, and scores on the Guilford Alternate Uses Task were higher there.
Tawil set out to test that same hypothesis properly and found against it. Forty-two participants, aged 18 to 40, explored four well-matched virtual living rooms freely, with room style included as a second variable. Thirty-three outcome variables were measured, covering affect, behaviour and cognition. After correcting for false discoveries, two differed, and both favoured the angular rooms rather than the curved ones. On the cognitive task, a mental arithmetic test taken from the Trier Social Stress Test, there was no effect of contour at all: Z was −0.431 at p = 0.666, with a Bayes factor of 5.12 in favour of the null, which counts as moderate evidence that there is nothing there. A frequentist and a Bayesian analysis agreed.
Saulton had participants judge whether a rectangular room was larger or smaller than a square room of reference, with the depth-to-width ratio and the viewing position varied, and 18 people per viewpoint in each culture. For the German participants, the same room read as larger from one position and smaller from the other, at r = .43 for both the 1:2 and the 1:3 rooms. For the South Korean participants the viewpoint made no difference, at p = .69, and two of them were excluded per viewpoint for not completing the task. Saulton reads that as East Asian spatial processing being more context-dependent. For a floor plan it means the same dimensions read as two different sizes depending on where the viewer stands, and for some populations not at all.
Moscoso ran the same virtual experiment in Norway, Switzerland and Greece, with 150, 118 and 138 participants, varying window size, room size, the use of the space and the sky outside. Regional differences showed up in how pleasant and how calm the space felt, between Greece and Norway at every window size and between Greece and Switzerland at the medium and large ones. On spaciousness, the paper reports that participants in Greece rated the small space as more spacious than participants in Switzerland did for the small and medium windows, with an increase of up to 7 per cent in their ratings.
Lee changed the room while people worked in it. Thirty people with telecommuting experience did four tasks covering spatial memory, attention, execution and working memory in a mixed-reality room whose design was altered in real time, with EEG and heart rate variability recorded. The design of the space did not affect spatial memory. The parasympathetic nervous system was more active in the spaces participants did not prefer, which Lee reads through the Yerkes-Dodson law as lower arousal and efficient short-term work, with the preferred space better suited to sustained work over time.
None of these five studies took place in a room anyone works in. Moscoso ran 406 participants across Norway, Switzerland and Greece. Tawil ran 42 and applied the strictest correction of the five. Strachan-Regan and Baumann ran 35, Lee 30, and Saulton 18 per viewpoint in each of two cultures.