How salutogenic workplace characteristics influence psychological and cognitive responses in a virtual environment

Abstract

Until today, most research focussed on the effects of pathogenic workplace demands on employee illness instead of on salutogenic resources on health. Using a stated-choice experiment in a virtual open-plan office, this study identifies key design aspects that enhance psychological and cognitive responses, ultimately improving health outcomes. The study systematically varied six workplace attributes: screens between workstations, occupancy rate, presence of plants, views outside, window-to-wall ratio (WWR), and colour palette. Each attribute predicted perceptions of at least one psychological or cognitive state. Plants had the highest relative importance for all expected responses but views outside with ample daylight, red/warm wall colours, and a low occupancy rate without screens between desks were also important. Low-cost interventions like adding plants, removing screens, and using warm wall colours can contribute to a healthier open-plan office environment. These insights can guide workplace managers to design environments that support employees’ mental states and health.

Practitioner summary: Salutogenic workplace resources that promote health have been understudied. This study aimed to show which workplace characteristics caused positive psychological and cognitive responses to improve health, using a stated-choice experiment in a virtual office environment. Plants in the office were the most important attribute for employees’ psychological and cognitive responses.

Keywords:

• Salutogenic workplace design
• virtual office
• stated-choice experiment
• cognitive workplace response
• psychological workplace response

1. Introduction

The Salutogenic Theory, as introduced by Antonovsky (1979), explains that individuals need motivation (i.e. meaningfulness), courage, and resources (i.e. manageability) to devote themselves to the understanding of (i.e. comprehensibility) and coping with stressful events. Salutogenesis originates from the Latin word salus (meaning a positive state) (Keyes 2014) and explains how goal-oriented behaviour could contribute to a positive, healthy state, thereby strengthening the sense of coherence (SOC). According to the SOC construct, some individuals remain healthy under stressful events while others do not. Those with a strong SOC are confident about the predictability and comprehensibility of their environment, and about the probability that a certain situation will be as expected. On the other hand, pathogenesis originates from the Greek word pathos (meaning a suffering state) (Keyes 2014), and clarifies a deteriorating SOC. As Antonovsky (1979) argues, pathogenesis is inadequate, because it pushes towards a dichotomous distinction between a state of disease or non-disease. He therefore argues that the Salutogenic Theory should be used, because it asks individuals how they can maintain healthy.

In the workplace context, the interaction between environmental opportunities and demands, and individuals’ needs and abilities also causes either a pathogenic or a salutogenic reaction (Kalimo, Ei-Batawi, and Cooper 1987). In a salutogenic work environment, the integration of multiple spatial qualities promotes employee needs and performance (Mangone et al. 2017). While having sufficient space is a fundamental human need (Stamps 2009), biophilic design features, such as natural views outside, daylight entrance, environmental attractiveness through plants and freshly painted walls, as well as social features including social interactions and feelings of belonging, promote positive psychological (i.e. mood and stress) and cognitive responses (i.e. concentration and productivity). These responses could increase individuals’ health and enable them to recover from workplace stressors more effectively (Heerwagen et al. 1995; Roskams and Haynes 2019). According to Lei, Yuan, and Lau (2021), biophilic design contributes to both psychological health, productivity, and physiological health. Exposure to biophilic design triggers unconscious, emotional, or psychological reactions (Thatcher et al. 2020), such as a positive mood (Lei, Yuan, and Lau 2021) (i.e. ‘a set of feelings, ephemeral in nature, varying in intensity and duration, and usually involving more than one emotion’ (Lane and Terry 2000, 17).

Two theories, namely the Attention Restoration Theory (ART) and the Stress Reduction Theory (SRT) have been applied to biophilic design. Restoration or recovery from stress can be described as the ‘process of recovering physiological, psychological and social resources that have become diminished in efforts to meet the demands of everyday life’ (Hartig 2007, 164). The physiological component considers the reaction of the autonomic nervous system to cope or deal with a certain situation. The psychological component includes cognitive appraisal of the situation (e.g. emotion and coping responses) (Ulrich et al. 1991). ART, as introduced by Kaplan (1993), explains what kind of (cognitive) experiences stimulate recovery from fatigue or stress. A natural environment could help to mitigate stress and stimulate recovery from fatigue (Kaplan 1995). Around the same time, SRT was introduced by Ulrich et al. (1991) and explains that recovery from stress involves positive changes in both psychological states, physiological systems, and behaviours. SRT thus puts more emphasis on affective and physiological reactions to certain settings than ART. Natural settings might enable physiological and emotional recovery due to lower levels of complexity and fewer arousing components than urban settings. Therefore, natural settings can cause positive changes in emotional states, behaviours, or can sustain attention.

Roskams and Haynes (2021) indicate that it is interesting to study which environmental resources can reduce employees’ feelings of strain, and how these promote positive psychological responses that could eventually improve health. Until today, only a few studies focussed on salutogenic workplace aspects (e.g. colours, plants, and views outside), while most studies focussed on indoor environmental quality (IEQ) aspects, such as noise, air quality, and temperature (Bergefurt et al. 2022). The salutogenic aspects have most frequently been related to productivity, concentration, stress, and mood. But, as Bergefurt et al. (2022) argued in their systematic review, research about the effects of salutogenic workplace characteristics on these employee outcomes could be expanded. Therefore, the current study aims to gain insight into how the preference for salutogenic design aspects, including screens between workstations, occupancy rate, daylight entrance measured through window-to-wall ratio (WWR), having a view outside, plants in the office, and colour palette, influence employees’ perception of the ability to be productive and to concentrate, and their perception of expected stress levels and mood in an open-plan office environment.

The novelty of this study is the use of a stated-choice experiment, which allows the identification of trade-offs that employees make between the mentioned salutogenic design aspects. It makes clear which aspects are perceived more important than others to retain different positive psychological and cognitive responses that could stimulate health. Another novelty of this study is the use of a stated-choice experiment with non-immersive virtual reality (VR), existing of short videos to simulate an open-plan office environment. One of the advantages of this approach is that it overcomes the issues of a living lab experiment, where some factors cannot be controlled (Jo, Kim, and Jeon 2019). Some studies have already introduced a virtual office environment (Ahmaniemi et al. 2018; Yin et al. 2020), but, to the best of the authors’ knowledge, none of those studies so far focussed on the salutogenic open-plan office environment design specifically. Furthermore, until today, no research combined the use of a stated-choice experiment with non-immersive VR in the office context.

2. Salutogenic workplace characteristics

In this section, literature about the relationships between six salutogenic workplace characteristics and productivity, concentration, stress, and mood will be discussed. These characteristics are plants, views outside, colour palette, daylight entrance, screens between desks, and occupancy rate, respectively.

2.1. Plants

An individual can either be directly (e.g. plants) or indirectly (e.g. window views) exposed to nature or may be exposed to a representation (e.g. photos) (An et al. 2016). Direct nature exposure through plants helps to create a pleasant physical and social environment (Bjørnstad, Patil, and Raanaas 2015), which may result in a positive mood, increased productivity (Smith, Tucker, and Pitt 2011; Smith and Pitt 2009), and reduced stress levels (Ayuso Sanchez, Ikaga, and Vega Sanchez 2018; Elsadek and Liu 2021). If natural views are absent, employees will compensate for this by bringing plants into their workplace. Employees without window views were five times more likely to bring plants into their workspace than employees with window views (Bringslimark, Hartig, and Patil 2011). Some studies also considered the effect of vertical plants or green walls and found that they have a more positive influence on people’s mood than potted plants (Gunn, Vahdati, and Shahrestani 2022). Although these studies show significant effects of the presence of plants on mood, stress, concentration, and productivity, Bringslimark, Hartig, and Patil (2007) note that the effects are rather small and depend on the amount, position, and type of plants. They argue that too many plants in the office could even reduce employees’ productivity. In addition, Qin et al. (2014) show that small (compared to medium and large) and green-coloured (compared to tint and multi-coloured) plants were most preferred.

2.2. Views outside

Indirect exposure to nature can be provided by natural views outside (An et al. 2016). A natural view outside is one of the primary requirements for employees to feel satisfied in their office (Al Horr et al. 2016). As Kaplan (1993) argues, a window view functions as a micro-restorative experience and may recover individual’s attention. For instance, viewing a green roof for only 40 s already has several cognitive benefits and can restore individuals’ attention (Lee et al. 2015). Similarly, Beute and de Kort (2014) indicate that viewing natural scenes for only three minutes improves individuals’ hedonic tone (i.e. pleasantness or happiness) compared to viewing urban scenes. These findings show that views on natural scenes, such as green areas, clouds, and the sky (Jamrozik et al. 2019) or forest views (Shin 2007), provide a restorative (Smith, Fsadni, and Holt 2017) and calming effect (An et al. 2016), and decrease brain activity and stress (Bjørnstad, Patil, and Raanaas 2015; Jo, Kim, and Jeon 2019). As Sadick and Kamardeen (2020) argue, reducing employees’ stress through nature may have several psychological and cognitive consequences that could improve health, such as better cognitive performance through improved working memory and inhibition of automatic responses (Jamrozik et al. 2019).

2.3. Daylight entrance

Office windows are important for the entrance of daylight, views outside, and natural ventilation (Dogrusoy and Tureyen 2007). The entrance of daylight provides a pleasant environment which may increase people’s productivity (Fasi and Budaiwi 2015). Employees perceive their workspace more positive overall when windows are large. They rate a workspace with larger windows as more pleasant, interesting, complex, spacious, and bright (Moscoso et al. 2021). Furthermore, larger windows and proximity to a window positively influence productivity (Vischer 2007). While too limited daylight reduces employees’ mood, too much daylight may increase the risk of glare, thereby reducing concentration levels (Jamrozik et al. 2018). The window-to-wall ratio (WWR) can be used to determine the optimal daylight access, depending on the size and the number of windows. According to Hong et al. (2019), the WWR should range between 15 and 60%.

2.4. Colour palette

Küller et al. (2006) argue that good colour design in the office might positively contribute to employees’ mood. In general, cool colours are associated with concentration and attention, and warm colours with arousal. Kwallek, Lewis, and Robbins (1988) show that using warm colours (e.g. red) in the office increases anxiety and stress levels, and that cool colours (e.g. blue) increase depressive symptoms. While high-chroma colours do not effectively reduce stress, high-brightness colours do so (Hsiao, Hsiao, and Wang 2013). White or neutral colours are observed as most spacious and appropriate for the office, while orange and purple are the least preferred (Kwallek et al. 1996). Mahnke (1996) argues, furthermore, that pale colours, such as pale gold, orange, green, and sandstone are also appropriate for office environments.

2.5. Screens between desks

Open-plan offices (i.e. shared rooms with groups of workstations) (Bodin Danielsson et al. 2008) have gained attention because they were believed to increase communication and collaboration between employees (Yunus and Ernawati 2018). However, research shows that open-plan offices often lack a sense of visual and acoustic privacy, which may result in several health-related responses, including reduced productivity (Haynes, Suckley, and Nunnington 2017) and increased stress levels (Jahncke et al. 2011). In general, the impact of open-plan offices can be divided into two approaches. First, the socio-relations approach explains that the removal of partitions between workstations results in more social relations between employees, increasing their satisfaction and engagement (Yunus and Ernawati 2018). Second, according to the socio-technical approach, partitions are necessary to provide employees with a sense of privacy and to reduce the experience of distractions. As a result, employees feel more concentrated and productive (Yunus and Ernawati 2018).

The findings of Oldham and Brass (1979) seem to confirm the socio-technical approach. They suggest that the absence of walls or partitions between workstations influences the quality of employees’ work experience, including less feedback from supervisors, reduced task identity, and fewer friendships with colleagues. Yildirim, Akalin-Baskaya, and Celebi (2007) indicate that those with an enclosed workstation are most satisfied, because of improved visual and acoustic privacy and reduced distractions. However, Kim and de Dear (2013) found that satisfaction with both sound privacy and noise level is higher in offices without partitions. They argue that the unpredictability of sound due to the visual disconnection in offices with partitions reduces employees’ satisfaction with noise and might even increase the experience of noise distraction. Although these studies seem to provide stronger support for the socio-technical approach, to the best of the authors’ knowledge no studies so far studied the relation between desk-partitions and psychological responses (i.e. mood and stress).

2.6. Occupancy rate

The proximity to colleagues depends on the spatial density (i.e. area available per occupant) and social density (i.e. number of occupants per office) of the office. A higher spatial density means that the available space per workstation is larger, which might lead to less distractions and a higher concentration rate (Veitch 2018). In contrast, a higher social density might increase the frequency of social interactions but should be limited to prevent concentration-and productivity-issues (Hua et al. 2010; Veitch 2018). High-density offices are associated with increased psychosocial work stressors and reduced environmental satisfaction, causing more cognitive issues, such as lower work ability and more concentration difficulties (Aries, Veitch, and Newsham 2010; Herbig, Schneider, and Nowak 2016). Those who share the workplace with more colleagues (Rasheed, Khoshbakht, and Baird 2019) and those who are annoyed by high crowdedness in the office also rate their productivity more negative (Węziak-Białowolska, Dong, and McNeely 2018). However, Pejtersen et al. (2006) do not confirm these findings. They argue that psychosocial risk factors, such as quality of leadership, social support, motivation, and feedback, are independent of the number of occupants in the office. They do advocate for adapting the office occupancy according to the degree of concentration that is needed for a specific job type. As Roskams and Haynes (2019) argue, the occupancy rate might be adapted to reduce distractions and feelings of crowdedness. To the best of the authors’ knowledge, the effects of occupancy rate on psychological outcomes, such as mood and stress, have so far not been studied. Table 1 shows an overview of the literature that studied the relationships between six salutogenic workplace characteristics and productivity, concentration, stress, and mood.

Table 1. Literature review.

Author	Approach	Findings	Stress	Productivity	Concentration	Mood
Plants
Ayuso Sanchez, Ikaga, and Vega Sanchez (2018)	Experiment with subjective and objective observations	The presence of greenery (small plants) decreases stress.	x
Smith and Pitt (2009)	Experiment, one office with and one without plants	Employees felt slightly more productive in the office with plants.		x
Bringslimark, Hartig, and Patil (2007)	Cross-sectional survey	More plants (>3) within 1 m from desk decreases productivity.		x
		More plants in view increase productivity.		x
Smith, Tucker, and Pitt (2011)	Experiment, one office with and one without plants	Employees in the office with plants felt more productive.		x
Hähn, Essah, and Blanusa (2021)	Experiment, measures pre- and post-introduction of plants	Removal of plants from individual workstations, and only plants in break-out spaces increases perceived stress.	x	x
		Removal of plants from individual workstations, and only plants in break-out spaces decreases perceived productivity.
Toyoda et al. (2020)	Experiment, measures pre- and post-introduction of plants	Small plants at the desk are associated with lower psychological stress.	x
Nieuwenhuis et al. (2014)	Longitudinal experiment	Employees in the office with plants indicated increased ability to concentrate.			x
Bjørnstad, Patil, and Raanaas (2015)	Cross-sectional survey	Indoor nature contact is negatively associated with job stress.	x
Gray and Birrell (2014)	Longitudinal experiment	The introduction of plants mitigates against stress.	x
Elsadek and Liu (2021)	Living lab experiment	Flowering plants reduce stress levels.	x
Gunn, Vahdati, and Shahrestani (2022)	Scoping review	Vertical plants have more positive influence on mood than potted plants.				x
Views outside
Shin (2007)	Cross-sectional survey	Forest views are associated with lower job stress.	x
Jamrozik et al. (2018)	Living lab experiment	Not having a view outside negatively affects mood.				x
Bjørnstad, Patil, and Raanaas (2015)	Cross-sectional survey	Indoor nature contact is negatively associated with job stress.	x
Jo, Kim, et al. (2019)	Scoping review	Nature views decrease brain activity and stress.	x
Jamrozik et al. (2019)	Living lab experiment	Nature views improve cognitive performance.		x
Beute and de Kort (2014)	Laboratory experiment	Viewing nature improves hedonic tone.				X
Daylight (WWR)
Butler and Biner (1989)	Experiment	Medium-sized windows were preferred for employees’ mood.				x
Moscoso et al. (2021)	VR experiment	Large-sized windows lead to more positive evaluations of how pleasant, exciting, interesting, bright, complex, and spacious the office space was.				x
Fasi and Budaiwi (2015)	Building energy simulation	The entrance of daylight may increase productivity.		x
Vischer (2007)	Literature review	Larger windows and proximity to a window positively influence productivity.		x
Jamrozik et al. (2018)	Living lab experiment	Too limited daylight reduces mood.			x	X
		Too much daylight causes risk of glare, reducing concentration levels.
Colours
Hsiao, Hsiao, and Wang (2013)	Cross-sectional survey	High-chroma colours do not effectively reduce stress.	x
		High-brightness colours effectively reduce stress.
Küller et al. (2006)	Cross-sectional survey	Mood was rated higher in colour work environments.				x
Kwallek, Lewis, and Robbins (1988)	Laboratory experiment	Employees who switched between different-coloured (from blue-red or red-blue) workspaces were more aroused.				x
Kwallek, Soon, and Lewis (2007)	Longitudinal experiment	Employees who can screen against arousal perform better in red coloured offices than employees with low screening ability.				X
Moscoso et al. (2021)	VR experiment	Large-sized windows lead to more positive evaluations of how pleasant, exciting, interesting, bright, complex, and spacious the office space was.				x
Fasi and Budaiwi (2015)	Building energy simulation	The entrance of daylight may increase productivity.		x
Vischer (2007)	Literature review	Larger windows and proximity to a window positively influence productivity.		x
Jamrozik et al. (2018)	Living lab experiment	Too limited daylight reduces mood.			x	X
		Too much daylight causes risk of glare, reducing concentration levels.
Screens between desks
Haynes, Suckley, and Nunnington (2017)	Cross-sectional survey	Employees in open-plan offices without partitions rated productivity most negative.		x
Jahncke et al. (2011)	Laboratory experiment	No visual and acoustic privacy in open-plan offices increases stress levels.	x
Yunus and Ernawati (2018)	Mixed-method study	Partitions increase productivity and concentration.		x	X
Occupancy rate
Aries, Veitch, and Newsham (2010)	Cross-sectional survey	A higher social density increases concentration difficulties.			x
Rasheed, Khoshbakht, and Baird (2019)	Cross-sectional survey	Employees who shared their office with fewer people rated their productivity higher.		x
Veitch (2018)	Literature review	High spatial density leads to higher concentration.		x	x
		High social density leads to concentration-and productivity-issues.
Hua et al. (2010)	Field study	High social density leads to concentration-and productivity-issues.		x
Węziak-Białowolska, Dong, and McNeely (2018)	Cross-sectional survey	High crowdedness reduces productivity.		x

3. Methods

3.1. Stated choice experiment

To collect data on employees’ workplace preferences, a survey-based stated-choice experiment was used. In a stated-choice experiment, a respondent is provided with hypothetical alternatives that are described by several characteristics (i.e. attributes) to determine their choice behaviour (Hensher, Rose, and Greene 2015). The advantages of this method are that different attributes (the salutogenic workplace resources) can be controlled and varied, and that a robust model is obtained with a wide range of design attributes (Adamowicz, Louviere, and Swait 1998). In this study, six attributes were considered, namely screens between workstations, occupancy rate, WWR, views outside, colour palette, and plants. According to existing literature, these could all influence employees’ workplace preferences and their health. For each of the attributes three corresponding levels were designed and varied, which are shown in Table 2.

Table 2. Attributes and levels.

	Attributes	Attribute levels
1	Plants	No plants Horizontal plants Vertical plants (base)
2	Views outside	City view (base) Industrial view Natural view
3	Window-to-wall ratio (WWR)	20% 40% (base) 60%
4	Colour palette	Red/warm Blue-green/cool White/neutral (base)
5	Screens between desks	Front and side screens Front screens No screens (base)
6	Occupancy rate	25% 70% (base) 100%

A full factorial design consists of all possible combinations of the attribute levels, resulting in 3⁶ = 729 possible alternatives. Since this full factorial design would result in too many alternatives, an orthogonal fraction of the experimental design was used, leading to 27 alternative office designs (i.e. profiles). Two hypothetical alternatives were randomly combined into choice sets. Respondents were asked to choose between these profiles by selecting in which office (i.e. profile A or B) they would be able to work most productively, concentrated, feel least stressed, most relaxed, and most happy, respectively. They were also asked to indicate which office (A or B) they preferred the most in general. If a respondent had no preference for one of the two office designs, the ‘no preference’ option could be chosen. In total, respondents were shown four different choice sets, resulting in 24 (5 + 1 health states × 4 choice sets) choice observations per respondent (see Figure 1 for an example).

Figure 1. Example of stated choice experiment with choice task.

Figure shows an example of a choice task, in which people should indicate their preferred office scenario (A or B). If they do not have a preference, they can select the ‘no preference’ option.

Read the detailed description of this figure

The figure shows an example of a choice task, in which people should choose in which office scenario (A or B) they expect to work most productively and concentrated, feel least stressed, and most relaxed and happy. If they cannot make a choice, they can choose the ‘no preference’ button.

3.2. Modelling approach

A mixed-multinomial logit model (MMNL) was run to determine the probability that a particular alternative is chosen by an individual. The utility of the null-alternative ‘no preference’ was represented in a constant. A random component was estimated for each attribute, to capture the unobserved heterogeneity in the model. To estimate the parameters of the model, 150 Halton draws were used. The use of Halton sequences increases the accuracy of the model and reduces the computation time. The influence of personal- and physical workplace characteristics on respondents’ preferences were also analysed, by stepwise including control variables that were also questioned in the survey (e.g. age, gender). The control variables were entered into the model as non-random parameters. The package ‘GMNL’ in R was used to run the MMNL.

Dummy coding was used to code each attribute level that appeared in the scenario with one, and zero otherwise. The base level for each attribute received a value of zero. The parameters were estimated in a single model, handling all choice observations simultaneously and using health-state specific parameters for the attributes and a single constant parameter. In this way, the constant represents a general inclination of individuals for being indifferent between presented options.

3.3. Research design

The virtual office design was first drawn in SketchUp Pro 2021. The 3D Warehouse Repository was used to import tables, chairs, and cabinets. Twinmotion 2020.2 was used to add detailed surface materials and animated people. For each of the 27 office design scenarios, a 20-s video was then made in Twinmotion, in which the respondent was virtually guided through the office to a free workspace. Figures 2(a–d) show this walkthrough. The videos were exported as MP4 files and uploaded on YouTube.

Figure 2. (a) Screenshot of video at start of alternative 23; (b) at 5 s; (c) at 10 s; (d) at 20 s.

Screenshots a to d show the virtual office design of alternative 23, with white walls, several desks with desktops and people sitting behind the desk, and four windows with a city view.

Read the detailed description of this figure

(a) Shows the start position of the virtual office design of alternative 23. It includes white walls, white cabins at the left, and a glass door at the right side, (b) shows a screenshot of the virtual office design at 5 s, with several desks with people, and two windows with a city view, (c) shows a screenshot at 10 s, with desks with desktops and people sitting behind the desks, and four windows with a city view, and (d) shows a screenshot at 20 s (the end of the movie), including several desks with desktops and people sitting behind the desks.

3.3. Data collection

The stated-choice experiment was designed in the server-based software LimeSurvey. The software was used to make an online survey, consisting of the choice tasks and several other questions. These questions are related to employees’ gender, age, and personality. Personality was measured by the 10-item Big Five Inventory (BFI) (Rammstedt and John 2007). Furthermore, respondents were asked to describe their current office workspace (before the COVID-19 pandemic) and to indicate their current concentration-, productivity-, and stress-levels and how happy and relaxed they were on a 10-point scale. This survey was first distributed via e-mail to companies that financially support a larger overarching research project. As this resulted in a too small sample, it was decided to distribute the survey further via the LinkedIn networks of the first two authors. The sample is therefore non-random, which might have reduced the reliability of the results. Data were collected between September and December 2021, resulting in a total sample of 214 office workers. The study protocol was reviewed and accepted by the Ethic Review Board of Eindhoven University of Technology, the Netherlands, according to the Law on Medical Scientific Research involving Human Beings.

3.4. Sample descriptives

Table 3 shows that the sample size equals 214, with 856 observations in total (4 observations per respondent). Pedhazur and Schmelkin (1991) indicated that the number of observations should equal at least 30 times the number of predictors. This rule of thumb is satisfied, meaning that the sample size is acceptable. The sample consisted of somewhat more female (53.3%) than male respondents with a mean age of 39. Most respondents indicated agreeable personality traits, followed by neurotic traits. The inter-item correlation values for every personality trait were above 0.2. Most employees worked in an open-plan office before the COVID-19 pandemic (79.9%), characterised by front screens between desks (44.9%) or no screens (35.5%), a middle occupancy rate (43.9%), with high daylight entrance (61.5%), and a city view (55.6%). In most offices, white/neutral colours were used (54.2%), and horizontal plants were placed (67.3%). On average, employees rated their hedonic tone most positive and their stress level most negative. This indicates that employees rate their stress levels on average fairly high, while they also rate their hedonic tone rather positive.

Table 3. Sample description.

	Sample (N)	Sample (%)	Mean	SD
Personal characteristics
Gender
Male	100	46.7
Female	114	53.3
Personality
Agreeableness			3.85	.64
Neuroticism			3.51	.85
Openness			2.42	.73
Extraversion			2.37	.86
Conscientiousness			1.99	.63
Age			39.62	11.30
20–30 years	62	29.0
31–40 years	62	29.0
41–50 years	45	21.0
Older than 51 years	45	21.0
Physical workplace characteristics
Office type
Open-plan office	171	79.9
Shared office (with 1 or 2 colleagues)	20	9.3
Private office	23	10.7
Screens between workstations
Front screens	96	44.9
Front and side screens	35	16.4
No screens	76	35.5
Other	7	3.3
Occupancy rate
Low occupancy rate (25%)	55	25.7
Middle occupancy rate (70%)	94	43.9
High occupancy rate (100%)	63	29.4
Other	2	0.9
Daylight entrance
Low daylight entrance (20% WWR)	8	3.6
Middle daylight entrance (40% WWR)	76	34.4
High daylight entrance (60% WWR)	136	61.5
Other	1	0.5
Views outside
City view	119	55.6
Natural view	36	16.8
Industrial view	37	17.3
Other	22	10.3
Colour palette
White/neutral colours	116	54.2
Red/warm colours	37	17.3
Blue-green/cool colours	43	20.1
Other	18	8.4
Plants
Horizontal plants (potted plants)	144	67.3
Vertical plants	16	7.5
No plants	44	20.6
Other	10	4.7
Psychological and cognitive responses (1. Negative–10. Positive)
Hedonic tone (happy)			7.22	1.64
Concentration			7.03	1.66
Productivity			6.94	1.58
Tense arousal (relaxed)			6.62	1.98
Stress			4.27	2.40

4. Results

4.1. Goodness of fit statistics

Table 4 shows the goodness of fit statistics for the mixed multinomial logit model. For a good model fit, the adjusted ρ² value should range between 0.2 and 0.4 (Louviere, Hensher, and Swait 2000). An adjusted ρ² value of 0.13 indicates a modest fit. The log-likelihood of the estimated MMNL model (LL(β)) should show a statistical improvement over the log-likelihood of the base model (LL(0)), which is the case for this model.

Table 4. Goodness of fit statistics.

Goodness of fit statistics
Log likelihood of estimated parameters (LL(β))	−4499.2
Log likelihood of the null model (LL(0))	−5169.71
Adjusted McFadden’s Rho-square (ρ^2)	.130
Akaike Information Criterion (AIC)	9187.09

4.2. Relative importance of workspace attributes

Figure 3 shows the relative importance of the workspace attributes for the perceptions of each psychological or cognitive response to the workspace. These utility impacts are computed using the range between the lowest and the highest part-worth utility. The part-worth utility indicates the relative importance of each attribute for the employees’ perception of their own psychological or cognitive response. Plants have the highest relative importance for all perceived responses, followed by the occupancy rate. Furthermore, it can be observed that the relative importance of the colour palette for employees’ perception of the ability to concentrate is rather low. This can also be observed for screens between workstations for their general preference, and for the window-to-wall ratio for perceived stress levels.

Figure 3. Relative importance of workspace attributes per psychological or cognitive response.

Bars of different heights, showing the relative importance of each workspace attribute on a scale from 0 to 50%, indicated per psychological or cognitive response. For each response, a different grey colour has been used. A higher bar indicates a higher relative importance.

4.3. Workplace preferences per health state

Table 5 shows the results of the MMNL model (see also Figure 4). All attributes have a significant influence on the perception of at least one psychological or cognitive response. First, compared to vertical plants, the absence of plants has a negative effect on all expected responses, while horizontal plants provide a significant positive effect on all these responses. The effects of horizontal plants on psychological and cognitive responses are all relatively large. In addition, relative to city views (base), natural views have a positive effect on expectations regarding stress (i.e. reducing stress), feeling happy, feeling relaxed, and the general preference. Surprisingly, industrial views have the same positive effect. This shows that employees prefer both a natural and an industrial view, relative to city views, for these responses.

Figure 4. Utility values (β) for each attribute per psychological and cognitive response, relative to baseline level.

Bars with different heights, indicating the utility value for each attribute per psychological and cognitive response relative to a baseline level, on a scale from −3 to 3. A bar between 0 and −3 indicates a negative effect on the response relative to the base, and a bar between 0 and 3 indicates a positive effect on the response. Note: a indicates insignificant effect, negative effect on stress indicates expectations of increased stress.

Table 5. Mixed multinomial logit model.

Attribute	Level	β (z-value)	p	SD (z-value)	p	β (z-value)	p	SD (z-value)	p	β (z-value)	p	SD (z-value)	p
Constant	Constant	Productivity
		4.15 (.0003)		8.00 (5.49)	***	Concentration				Stress
Screens between workstations	Front and side screen	−.14 (-.50)				.22 (.57)				−.65 (−2.40)	*
	Front screen	−.60 (−1.86)	.	1.36 (1.10)		−2.35 (−3.12)	**	5.45 (3.77)	***	−1.03 (−3.52)	***	.84 (.77)
	No screens (base)
Occupancy rate	25%	1.08 (3.91)	***	1.08 (1.55)		2.13 (3.08)	**	4.30 (2.45)	*	.79 (3.07)	**	.38 (.35)
	70% (base)
	100%	−1.06 (−4.39)	***			−1.68 (−4.09)	***			−1.55 (−5.80)	***
Window-to-wall ratio	20%	−.10 (-.50)				−.39 (−1.16)				−.40 (−2.14)	*
	40% (base)
	60%	.74 (2.72)	**	1.68 (1.92)	.	.80 (2.06)	*	1.43 (2.06)	*	.24 (1.00)		1.87 (2.69)	**
Views outside	City view (base)
	Industrial view	−.27 (−1.07)				−.67 (−1.82)	.			.63 (2.45)	*
	Natural view	.34 (1.50)				.80 (1.94)	.			1.07 (3.92)	***
Colour palette	Red/warm	.21 (1.03)				−.23 (-.61)				.84 (3.79)	***	.43 (.35)
	Blue-green/cool	−.22 (−1.01)		.092 (.14)		−.121 (−2.57)	*	3.53 (3.11)	**	.045 (.25)
	White/neutral (base)
Plants	No plants	−2.07 (−4.29)	***	1.95 (2.49)	*	−2.78 (−4.02)	***	1.73 (1.38)		−2.56 (−5.65)	***	.78 (1.24)
	Horizontal plants	1.03 (3.47)	***			1.60 (3.09)	**			1.39 (4.14)	***
	Vertical plants (base)
Interaction parameters
Views outside	Age 31–40 * Industrial view	1.07 (2.41)	*			1.37 (2.30)	*
		Feeling happy				Feeling relaxed				General preference
Screens between workstations	Front and side screen	−.74 (−2.74)	**			−.88 (3.00)	**			−.58 (−2.00)	*
	Front screen	−.81 (−3.23)	**	.87 (1.06)		−1.00 (−3.08)	**	1.27 (1.07)		−.41 (1.45)		1.24 (1.65)	.
	No screens (base)
Occupancy rate	25%	.73 (3.33)	***	.24 (.25)		.81 (3.04)	**	.38 (.46)		.99 (3.80)	***	.0018 (.0037)
	70% (base)
	100%	−1.82 (−5.59)	***			−1.97 (−3.84)	***			−1.40 (−4.93)	***
Window-to-wall ratio	20%	−.53 (−3.01)	**			−.32 (1.54)				−.35 (−1.68)	.
	40% (base)
	60%	.33 (1.31)		2.03 (2.76)	**	.53 (1.75)	.	2.39 (2.30)	*	.85 (2.52)	*	2.26 (2.82)	**
Views outside	City view (base)
	Industrial view	1.20 (2.70)	**			.54 (2.02)	*			.60 (2.30)	*	2.17 (.043)
	Natural view	.49 (2.22)	*	.11 (.17)		.85 *(2.67)	*	.032 (.042)		.98 (3.28)	**
Colour palette	Red/warm	.93 (4.31)	***	.045 (.067)		.91 (3.86)	***	.13 (.12)		.88 (3.37)	***	1.41 (2.61)	**
	Blue-green/cool	.52 (1.47)				.040 (.20)				.17 (.84)
	White/neutral (base)
Plants	No plants	−2.34 (−7.04)	***	.018 (.029)		−2.48 (−3.36)	***	.97 (.66)		−2.58 (−4.76)	***	2.06 (2.62)	**
	Horizontal plants	1.20 (4.57)	***			1.32 (3.21)	**			1.19 (3.23)	**
	Vertical plants (base)
Interaction parameters
Occupancy rate	Female * 100% occupancy	.89 (2.69)	**			0.80 (1.96)	*
Plants	Female * Horizontal plants									.69 (1.85)	.
Colour palette	Open-plan office * blue-green/cool colours	−.70 (−1.75)	.
Views outside	Private office * natural view	1.40 (2.13)	*
	Open-plan office * industrial view	−1.19 (−2.42)	*

Note: Negative effect on stress indicates increased stress.

***p < 0.001; **p < 0.01; *p < 0.05; ^.p < 0.1.

Furthermore, a WWR of 60% positively affects expectations of productivity, concentration, and general preference, relative to a WWR of 40% (base). A 20% WWR has a significant negative effect on perceived stress and feelings of happiness. This shows that employees prefer a high WWR, while a lower WWR might cause stress and reduced happiness. In addition, red/warm colours positively affect expected stress (i.e. reducing stress), feelings of happiness, feeling relaxed, and general preference, relative to white/neutral colours (base). Blue-green/cool colours have a negative effect on the expectation to be able to concentrate compared to the baseline. This means that employees prefer red/warm wall colours for psychological responses.

Front screens between desks have a significant negative effect on the perception of all responses except productivity and general preference, relative to no screens (base). Front and side screens also have a significant negative effect on the expectation to experience stress, feel happy and relaxed, and on the general preference. These findings show that employees expect to prefer the absence of screens between workstations when considering their own psychological and cognitive responses. Last, an occupancy rate of 25% rather than 70% (base) has a positive effect on the perception of all psychological and cognitive responses, while a rate of 100% has a negative effect. Employees thus prefer a low occupancy rate when interpreting potential effects on their responses.

4.4. Heterogeneity in the model

Table 5 shows several random parameters, which indicate that there is some heterogeneity in employees’ preferences regarding several attributes. Significant standard deviations (at a 10% α level) are found for all attributes, except for views outside, for the perception of all psychological and cognitive responses. This means that these parameters vary from one individual to another. For instance, a relatively high significant standard deviation of front screens on concentration is found, which suggests that employees’ preference regarding front screens between workstations for their expected level of concentration differs significantly.

4.5. Interaction parameters

Table 5 reports several interaction parameters. First, females attach more importance to an occupancy rate of 100% to feel relaxed and happy. This suggests that females find a high occupancy rate important to feel relaxed and happy. Females also find horizontal plants important for their general preference. Another interaction effect shows that employees who already work in an open-plan office (instead of a private or shared office) attach more importance to industrial views to feel happy. For those who work in a private office (instead of an open-plan or shared office), the expected effect of natural views outside on feeling happy is larger. Last, among employees aged between 31 and 40 years old (instead of 20–30, 41–50, or 51>), the expected positive effect of industrial views on productivity and concentration is larger. This suggests that these employees find industrial views important for their productivity and concentration.

5. Discussion

5.1. Discussion and implications

While previous studies mainly focussed on the indoor environmental quality (IEQ) aspects of the work environment (Bergefurt et al. 2022), this study showed which salutogenic design aspects are perceived as important to retain positive psychological and cognitive responses to an open-plan office. The main contribution of this study is therefore that it provides evidence of the relative importance of several salutogenic workplace characteristics for employees’ expected psychological and cognitive responses in the open-plan office environment. From an academic perspective, this study is the first to use a stated-choice experiment with non-immersive virtual reality in the open-plan office context. It offers many opportunities for future research in the workplace design research field.

The results indicate that all design aspects, namely the presence of plants, views outside, colour palette, daylight entrance (WWR), screens between desks, and occupancy rate, were perceived to influence at least one psychological or cognitive response. First, the presence of plants in the workplace had the highest relative importance for all the expected responses that were surveyed. The absence of plants was perceived to have a negative effect on all psychological and cognitive responses, while a positive effect was perceived from the presence of horizontal plants. Remarkable is that 20% of the respondents indicated that plants were absent in their current workplace. Although the current study only showed the perception of the effects of plants, Hähn, Essah, and Blanusa (2021) found that the removal of plants from individuals’ workstations increased their stress levels and reduced their productivity. They argued that the introduction of plants in the workplace is a low-cost investment that can positively affect workers’ well-being and restore their attention. This means that even when workplace managers have a small budget, these interventions might, to some extent, contribute to employees’ health.

Furthermore, respondents perceived the indirect exposure to nature through natural views outside to have a positive effect on several of their psychological and cognitive states. This study found that industrial views outside were also perceived to have a positive effect on these responses, relative to city views. This positive expected effect of both natural and industrial views relative to city views may be attributed to the relatively green character of the modelled industrial views in the VR environment. As Al Horr et al. (2016) indicated, having an outside view is one of the primary requirements for employees to feel satisfied in their office environment. Further research could determine what attributes of an office view exactly attract and satisfy employees.

Another prerequisite of the office is the entrance of sufficient daylight. This study showed that a WWR of 60% positively affected expectations of productivity, concentration, and general preference relative to the baseline. In line with this finding, Moscoso et al. (2021) found that offices with sufficient daylight were regarded as more pleasant and spacious and that employees were more satisfied with the views outside. Especially after the COVID-19 lockdowns, employees’ preference to have access to a window, daylight, and a view outside, or to feel connected to the outside, might have become stronger. However, in the current study, only the WWR was considered as an attribute, leaving variations in daylight entrance unstudied. It also remains unknown what the daylight entrance would be at each of the WWR-levels in a real office and how it would affect employees’ psychological or cognitive response.

Furthermore, indirect exposure to nature can be arranged using natural colours and materials in the office. Only a few previous studies included colours, especially in relation to other design aspects. This study showed that red/warm colours were perceived to have a positive influence on all psychological states, but not on the cognitive states (productivity and concentration). This is not surprising, since red causes both excitement and arousal (Nag 2019; van der Voordt, Bakker, and de Boon 2017). This study also showed that blue-green/cool colours were expected to have a negative effect on concentration. While white and neutral colours might be considered most appropriate for the office environment (Kwallek et al. 1996), this study shows that employees do prefer the use of warm colours to feel happy and relaxed, and to reduce their feelings of stress. More research is needed to understand why both neutral and warm colours have a positive influence on employees’ psychological responses.

In the current research, both office scenarios with front screens between workstations and front and side screens were perceived by respondents to have an expected negative effect on their responses, in comparison to no screens. Kim and de Dear (2013) explain that enclosing workstations with screens to support concentrative activities at the office improves employees’ satisfaction with visual privacy, but not with sound privacy. They argued that employees with enclosed workstations are partly disconnected from their surroundings. Employees cannot predict or control ambient sounds and might determine these as more disturbing. Therefore, Appel-Meulenbroek et al. (2022) argue that open-plan offices should consist of sufficient breakout areas and individual rooms to support both concentrative and communicative activities instead of enclosing workstations.

Last, a high occupancy rate might act as an environmental workplace demand. Although respondents generally preferred a low occupancy rate, the study also showed that among female workers a high occupancy rate was expected to have a positive effect on feeling happy and relaxed. During the COVID-19 lockdown, when employees were obliged to work from home, the balance between work-and family-responsibilities of dual-family careers might have been disturbed. Females more frequently took care of their children when schools and childcare were closed (Feng and Savani 2020), which might have reduced their health. Females felt more stressed, depressed, and anxious, and reported lower well-being and sleep quality during the pandemic (Cellini et al. 2021; Pieh, Budimir, and Probst 2020). Now that employees are invited to return to the office, females might feel more relaxed and happier than males when being surrounded by colleagues at the workplace instead of their children.

Although the method of this study did not allow to measure actual health effects, the study did gain insights into the expected influence of workplace design on psychological and cognitive responses. Preferred workplace design options may positively influence individuals’ primary appraisal (i.e. primary psychological and cognitive responses), which is an individual’s first evaluation of a situation. This evaluation (i.e. primary response) can consequently influence a person’s secondary appraisal. In secondary appraisal, an individual evaluates what resources are needed to improve the situation. As Folkman et al. (1986) argue, primary and secondary appraisals determine if the fit between the individual and the environment is beneficial for individuals’ health. Positive psychological and cognitive responses due to preferred workplace design solutions could thus have a health-promoting role. In future studies, other methodologies (e.g. living lab studies) can be used to objectify the effects of preferred salutogenic workplace characteristics on psychological and cognitive responses and eventually on (mental) health.

A practical implication of this study is that workplace managers and designers can use the insights to optimise workplace design according to general preferences of employees and their expectations of psychological and cognitive consequences. Some low-cost interventions, such as placing plants in the office, painting walls in warm colours, or removing screens between workstations, might have a positive influence on individuals’ psychological or cognitive responses or at least be perceived as such and also appear to influence employees’ general preference. These interventions should first be studied in living lab experiments, before implementing them in real office environments, since one-size-fits-all solutions to workplace design do not exist.

5.2. Limitations and future research

Although this study gained valuable insights, some limitations remain. First, respondents were asked to indicate how realistic they found the virtual office. Only 18% of the respondents indicated the office-scenarios to be very realistic, and 70% somewhat realistic. Some respondents commented that the render quality was not optimal and that plants were rendered too large. The high relative importance of plants might, to some extent, be explained by the render-size of the plants. Therefore, future research should focus on increasing the reality of the renders. Another limitation is the absence of several important control variables. Although several personal characteristics were included, employees were not asked to indicate the kind of task they imagined themselves doing in the VR scenarios while filling in the survey. In future studies, a hypothetical situation could be included about the individual’s imaginary work task in the VR environment.

Furthermore, in this study, only the influence of the relative preference of a minimal selection of attributes on psychological and cognitive responses was investigated. This means that potentially important attributes, such as environmental sounds, varying heights of screens between desks, or plants that are placed at individuals’ desks as well as on the floors, could not be studied. Furthermore, the results were based on subjective evaluations of employees’ preferences regarding these workspace attributes, meaning that the actual effects of these features on their responses and eventually on their health were not investigated. Especially the evaluation of the preferred VR office scenario for individual’s productivity might be difficult and thus differ from observations in a real office. Future studies could use longitudinal living lab approaches to elaborate on the actual effect of such workspace features on health. For instance, physiological indicators to measure body responses, heart rate, pulse rate, and blood pressure, could be used to study the effects of salutogenic workplace aspects on health (Jo, Song, and Miyazaki 2019). Another interesting technology is eye tracking, which allows one to verify what attracts employees’ attention in a physical or virtual environment (e.g. a biophilic design intervention) (Lei, Yuan, and Lau 2021). This technology can be combined with a stated-choice experiment, to verify the observed preferences of salutogenic attributes.

Next to drawbacks in the virtual design of the office, the fit of the model was a bit low, as the adjusted McFadden’s Rho-Square value indicated. A low fit indicates that there is a relatively high randomness in the choice behaviour, which could be caused by the difficulty of the choice task. It could also be caused by variation between individuals in assumptions regarding attributes that were not controlled in the experiment. Therefore, in future studies, a different set of design attributes or reducing the difficulty of the choice task might result in a higher Rho-Square value.

6. Conclusion

While previous research mainly focussed on pathogenic workplace demands, the objective of this study was to investigate which salutogenic design aspects are perceived as important for positive psychological and cognitive responses of employees to a virtual open-plan office. This study provides evidence of the relative importance of several salutogenic workplace characteristics for expected psychological and cognitive responses. Results demonstrated that each attribute predicted at least one psychological or cognitive state. Employees anticipated a preference for the presence of plants, windows with sufficient daylight and a view outside, the use of warm colours, the absence of screens between workstations, and a low occupancy rate for positive psychological and cognitive responses. These findings can be used by workplace managers to optimise the workplace based on employees’ preferences, particularly in light of the COVID-19 pandemic and the subsequent return to the office. Implementing these design attributes should be considered to improve employees’ health.

Acknowledgements

Bergefurt, Appel-Meulenbroek, and Arentze (2022).

Link: https://www.cfpb.nl/media/uploads/file/TWR-III-Proceedings_compressed.pdf

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by AMred, ABB, Ahrend, Arcadis, ASML, ASR, Cushman & Wakefield, Draaijer + Partners, EDGE Technologies, Leesman, Mansveld, PWC, Royal HaskoningDHV, Rijkswaterstaat, and SoftdB Expert.