How closed a room is

The architecture researcher Melina Forooraghi surveyed three activity-based office layouts at a Swedish public service organisation, with about 72% of 395 employees in one building and of 842 and 562 in another answering. Aesthetics scored highest at 84% satisfied. More than 60% were satisfied with functionality and task support. The lowest ratings went to storage at 48%, acoustic privacy at 56 and visual privacy at 47. Around 1,700 open comments came with the questionnaires, and the complaints about storage were specific: the lockers did not hold work materials, outdoor clothing and personal belongings at the same time.

The architecture researcher Lars Brorson Fich built a version of the Trier Social Stress Test in which the room itself could be varied, and put participants through it either in a closed room or in a room with openings. Saliva cortisol and heart rate variability were measured throughout. The closed room produced a more pronounced cortisol response to the stress induction, and the level stayed higher through recovery. No part of the autonomic nervous system differed. The paper is behind a paywall and only the abstract is available here, so the sample size cannot be given.

The computer scientist Crescent Jicol built shared and single offices in virtual reality and varied four things: how many people were present, how much floor area there was, whether there was noise and what the window showed. 20 participants estimated their personal space and rated their satisfaction with it. Satisfaction was 24.30 points higher in the small office than the medium one and 29.58 points higher than in the large one, both at p below 0.001, with no significant difference between medium and large. Any view out of the office helped. Noise made no difference, and neither did adding people while the floor area stayed constant. What moved perceived personal space was the floor area itself and whether the office was private.

The indoor environment researcher Monika Frontczak ranked visual privacy third of 15 environmental parameters for its contribution to overall workspace satisfaction, at an odds ratio of 1.26 across 52,980 occupants of 351 office buildings. Only the amount of space, at 1.57, and the noise level, at 1.27, did more. Satisfaction with all 15 contributed significantly, so what this study adds is the order and not the existence of an effect.

The medical researcher Zheng Zhang pooled 12 studies covering 12,719 patients on single against multi-patient intensive care rooms. The odds of an infection acquired in hospital were 0.68 in single rooms, with a confidence interval of 0.59 to 0.79. Colonisation and infection together was 0.44 and the incidence density 0.64. The outcome here is infection acquired in hospital, not satisfaction or stress, and no office study on this sheet measures anything comparable.

Frontczak covers 52,980 occupants and Zhang 12,719 patients, and both are pooled from other people’s data. Forooraghi surveyed three real offices. Fich and Jicol worked in virtual environments, Jicol with 20 participants and Fich with a sample the abstract does not give.