Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The work environment pilot: An experiment to determine the optimal office design for a technology company


24 Jun 2020: Pitchforth J, Nelson EC, van den Helder M, Oosting W (2020) Correction: The work environment pilot: An experiment to determine the optimal office design for a technology company. PLOS ONE 15(6): e0235428. View correction


Over the past few decades many corporate organisations have moved to open-plan office designs, mostly due to financial and logistical benefits. However, recent studies have found significant drawbacks to open plan offices and it is unclear how office designs can facilitate the best work output and company culture. Current design practice aims to optimise efficiency of space, but no previous research has tested the effect of office design experimentally in a working office. This paper describes an experiment comparing four different office designs (Open-plan, Zoned open-plan, Activity based, and Team offices) against a suite of wellbeing and productivity metrics in a real world technology company. Results suggest that two very different designs (Zoned open-plan and Team offices) perform well compared to Open-plan office designs. Zoned open-plan and Team office designs improved employee satisfaction, enjoyment, flow, and productivity, while Activity based and Open-plan designs performed poorly by comparison. The Open-plan office design was rated more poorly by employees, had higher levels of unsafe noise, and once employees no longer had to be in the Open-plan office design of the experiment, they spent more time at their desks.


Office design is an important factor in many variables of interest to managers, but the impact of open office design has rarely been explored experimentally. Since open offices have become mainstream, research has questioned their efficiency and general likability. Recent research has even gone so far as to wholly condemn open office designs, but designers are working with creative ideas to improve the open workplace [1]. To date, there has been no randomized, controlled trial in a working organisation capable of establishing such causal relationships. This paper presents the first randomized, controlled experiment in a working international technology company with the aim of finding an optimal open office design for employees.

Office design challenges

Office design can be a source of satisfaction [2, 3], engagement [4, 5], productivity [6, 7], and employee health [8, 9], but how companies design an office to meet the needs of their organisation is an open question. In the past, this question has frequently been answered through corporate strategy and budget restrictions [10]. However, as organisations become more data-driven and office space costs increase, many companies are beginning to consider other types of costs such as loss of productivity and decreased employee engagement [11]. These considerations have led to new questions regarding the impact of office design choices on employees.

A number of different office designs have been used over time to facilitate a wide range of work styles and goals. In this study we examine four designs common to modern companies, here described as Activity based, Open-plan (our control design), Zoned open-plan and Team office.

Activity based offices are flexible zone-based environments with unassigned seating that provide a range of spaces intended for different usages in an open setting. They have become popular due to aspects such as space reduction and cost savings, with employees making more efficient use of the office floorspace. Despite the financial benefits of the approach, activity based designs have been shown to decrease comfort, privacy and productivity [12] and contribute to emotional exhaustion [13]. Further, satisfaction with activity based environments is limited to one subset of workers [14]. However, recent work [15] has suggested that some of these issues can be overcome by including employees in the process of changing from their existing environment to an activity based environment. These mixed findings suggest that there is more that can be learned about this office design approach through experimental observation.

Open-plan offices are designed with minimal separation of spaces, such that the office floor is without internal walls or doors. They are currently very popular in large corporations [11], but they are associated with a range of issues including increased disturbances and lack of privacy [1619]. Past research has highlighted the tendency for open-plan office designs to drive negative behaviors and attitudes of employees through loss of space and increased contact with coworkers [17, 1923]. It has been established that environmental variables such as noise and visual disturbances [24, 25], poor air quality [26], temperature [27], and lighting [28] have an impact on satisfaction, engagement, and productivity in open plan environments, suggesting that examining open-plan office design in an experimental context adds value to the literature.

In contrast to Open-plan offices, cubicles are a way of breaking up open office floors with partitions between desks, providing an enclosed desk space for each employee. The design was once the most common type of office design, but has become less popular in recent years [29]. Some benefits of this type of office design, such as reduced visual distraction leading to increased perseverance [30] have been identified, but these are now seen as being offset by negative cultural effects and reduced collaboration [31, 32]. Most work on this design has focussed on individual cubicles, with no known work on designs with a team of employees in a single large cubicle.

While many researchers have chosen to frame their work in terms of different office design styles, others have remained more general in their approach and have instead focussed on relationships between specific environmental and commercially important variables. Here we review some variables that are proposed to be affected by office design.

Environmental variables

Sound disturbances have been documented to reduce cognitive performance, decrease motivation, and increase stress levels [24, 3336]. The prevalence of both audio and visual disturbances have flagged the need for improving employee privacy in the workplace [19].

Other environmental aspects of workplaces such as temperature, air quality and light have shown both negative mental and physical reactions to sub-optimal environments [37, 38]. The term ‘sick building syndrome’ was coined over 30 years ago when it was discovered that many of the offices we work in had poor air quality and were making people ill [37]. The 1984 World Health Organization (WHO) report suggested that up to 30% of new and remodeled buildings worldwide may be subject to complaints related to poor indoor air quality [39].

Office temperature has been established to be linked to both productivity [27, 40] and satisfaction [41, 42]. Recent studies have found that around 22 degrees appears to be an optimal temperature [27, 43, 44]. While there are many academic, government, and industry organisations specifying optimal office temperatures [4547], there is still debate around the specifics of how temperature affects employees. For example, [48] found that the relationship between temperature and performance was different for different types of work, and [49] found that women perform better on some types of work in higher temperatures than men. It may be that there is so much variation in reactions to office temperature that finding one optimal temperature is not possible [50]. However, given the ability to control office temperature and relative ease of measurement, it is of great value to understand more about how this variable affects employees.

Lighting is another important aspect of office design that can have significant effects on employees. The effects of different intensity, colour, and positioning of lighting on employees have been established in a number of studies [18, 51, 52], although there at present is no conclusive method of designing an optimal office lighting plan. Much of the previous work has focussed on how light can affect productivity by affecting circadian rhythms and making employees drowsy [52], and results suggest lighting states similar to natural light at the end of the day should be avoided. Recently [53] proposed a more technologically based solution involving estimating employee drowsiness, then adjusting lighting and air conditioning accordingly. With such a strong link established but little work on experimental observation outside a laboratory environment, there is still much to be learned about the effect of this variable on employee behaviour.

Occupancy and personal variables

Occupancy refers to the number of people using a space, and is an important factor in office operations planning [54, 55] as it affects a wide range of decisions such as required cleaning staff and opening hours. In academic literature, the primary focus of collecting occupancy data through sensors has been for further modelling purposes [56, 57] but has not been examined in relation to environmental variables such as temperature, noise, or air quality.

Productivity is the primary variable of interest to most organisations, as it is the source of profit and innovation. Given that productivity can mean different things in different job roles, it is easy to see why most studies of productivity have focussed on heavily operational roles with clear productivity metrics such as sales [58, 59], manufacturing [60, 61], or nursing [6264]. A notable exception is the work of [65], who examined the relationship between work habits and productivity in software developers. They note that software developer productivity is very difficult to measure, and that in such a situation it is more useful to measure perceived productivity. While this method may be less accurate than observing metrics of productivity, it is likely to be more relevant for modern knowledge industry companies. A measure for perceived productivity based on role and identity theory was introduced by [66], and has since been used in a variety of work areas such as public service [67] and nursing [68].

One tool commonly proposed as a proxy for productivity in technology driven companies is Git, a software first introduced in 2012 [69] which assists developers in coding productivity at a group level. The framework allows developers to ‘commit’ their code to collectively managed code bases in a systematic and well-recorded fashion. Scholtes in 2016 [70], demonstrated that logs of git activity can usefully be examined in relation to productivity. [71] raised the concern that git activity could be misleading given the software’s potential weakness to bad actors and issues with transporting repositories. These issues are not a concern in the current study however, as the corporate environment serves as a protection from both bad actors and infrastructure problems.

Satisfaction is another important concept for organisational and office design research, and is used as a primary measurement for the sentiment of employees. A common measurement of satisfaction with the working environment is the Leesman satisfaction index [72], which is seen as useful for benchmarking purposes. While satisfaction is clearly related to organisational and operational factors such as management styles [7375] and job requirements [76, 77], there is a large body of work that explores satisfaction with reference to office design. [78] found that satisfaction was related to view of nature, and [14] suggested that offices requiring more place-switching provided a higher level of satisfaction. In addition to the field-studies on satisfaction in office designs [79, 80], there have been a number of quasi-experiments examining employee satisfaction before and after office design changes [15, 80]. These studies show that satisfaction is of high interest to both researchers and corporations, but the construct is yet to be explored experimentally in the context of working technology companies.

Since its introduction in the literature, managers have been interested in how to help their employees reach psychological flow states, as these are seen to be related to productivity [81, 82]. The concept of flow stems from the positive psychology literature [83, 84], which uses the term to describe an optimal mental state for accomplishing a task. [85] outlines the various methods for measuring flow in different environments, such as questionnaire-based scales such as the WOLF scale [86], or the Experience Sampling Methodology [87] which requires frequent sampling at an individual level [88]. Hypothesised drivers of flow include task challenge [89], daily recovery [90], job characteristics [91], and organisational and personal resources [92], but no solid theory has yet been established on this topic. There is very little work on the relationship between office design and flow states, with recent work by [93] a notable exception, who invoke flow as one of the benefits of their proposed Emotional Design approach. [90] explored both energy at work and flow, focusing on the role of ‘detaching’ from work tasks outside of working hours.

Engagement (or its opposite, burnout), and enjoyment are closely related concepts that are also proposed to be affected by office design. Burnout is often described as the antipode of engagement [94], and has been noted to cause significant costs to businesses in the form of lost productivity [95]. Recently, it has become more convenient to measure burnout with the introduction of Maslach’s Burnout Inventory [96] a tool which is used to diagnose burnout in countries like The Netherlands, where this study was conducted. While previous research has focussed largely on effects of more abstract burnout drivers such as job role [97, 98] or social environment [99, 100], relatively little attention has been paid to the effects of the physical work environment on burnout [101, 102].

Enjoyment of a space is also very useful for organisations to measure, as it is related to other staff behaviours such as productivity and wellbeing [103]. Enjoyable office spaces can improve activity of sedentary workers [104], help guide corporate culture [105], and improve creativity [106]. While studies have demonstrated a clear relationship between enjoyment of office spaces and benefits to organisations, there is still a need to examine how different office designs vary in enjoyment from an experimental paradigm.

Studies of open office designs have considered these variables observationally, but there is still a notable gap in the evidence from systematic experimentation in real corporate environments. Previous research has explored aspects of employee sentiment and behaviour in relation to various aspects of the work environment, but none have employed an experimental design, tending to prefer quasi-experimental (non-randomized) and case study approaches such as work by Cisco [107] and Hewlett-Packard [108]. The present paper describes the first experimental analysis of these four open office designs in direct comparison. It also takes a multivariate approach to understanding employee behaviour, with analysis based on automated data collection in addition to more commonly used research tools such as questionnaires and interviews. By systematically observing the performance of four common open office designs in a modern technology focussed company, we demonstrate that an experimental approach to office design can help corporations find optimal office designs and design elements for their employee population, as well as providing evidence of the relationships between office design and commonly studied constructs with strong ecological validity.

Given the current state of office design research and the noted gap in the literature, the aim of this research is to use an experimental approach to compare four open office designs, and determine if there is an optimal open office design which fosters productive work but also reaches high levels of satisfaction for employees of a large corporate tech company.

Materials and methods

This study involved human subjects, and was formally approved with written confirmation by the B.V. Works Council in the Netherlands, which is a legally designated body charged with upholding ethical and professional standards for the company. All data were analyzed anonymously in accordance with European GDPR regulations. is a large corporate tech company founded 23 years ago in the Netherlands with an employee population of 17,500 people globally, 5,580 of which are based in the global headquarters in Amsterdam. Their employee population includes software and system developers, travel market specialists, marketing, finance, and a range of support roles facilitating the entire customer experience.


Participants were recruited as whole teams from a sample of all departments and job roles in, excluding Customer Service agents who work in call centers rather than an office environment. The final sample included 288 participants from 22 teams in a similar mix of functions to that found in the wider company including code development, project management, administration of HR processes, and strategy formulation. Each team ranged in size from four members to 20 members.

All participants were informed of the intent to run an experiment and were shown outlines of the environment they would be seated in. They discussed their participation in the experiment with their team, who decided to participate or not. Communications to participants stressed that there was no requirement to participate in the experiment, and that there would be no consequences for non-participation. The experimental area was clearly marked with signage indicating the boundaries of the experiment, and that participants could leave the area if they did not want to participate. All experiment plans were presented to the Works Council, a legally mandated body with oversight of all changes regarding employees, who provided written consent on behalf of the employee population.

Participants in the sample ranged in age from 22–59 years old with a mean age of 32. The gender split loosely reflected the company population with 59.84% identifying as Male. 51 nationalities were represented, including Asia, Europe, North and South America, Africa, and Oceania (Fig 1).

Sample selection

Sampling was achieved using a stratified convenience sampling approach wherein teams were asked to volunteer for the experiment, with a preference for complete organisational sets of teams (known as ‘tracks’). Teams were selected from those that volunteered to participate, working closely with team and track managers to ensure full coverage of job roles. The final choice of participants ensured representation was proportional to the makeup of departments in the company.

Experiment design

A uniform cross-over design [109112] was employed to systematically observe the effect of office design on the constructs of interest. The specifics of the design and consequent analysis method is described below.

Blocking and randomized exposure.

Teams of participants were assigned to one of four groups. An R script was used to randomly assign teams to groups such that each group had equal numbers of individual participants.

Each group was then exposed to each office design for two weeks at a time before moving to the next design. The sequence of exposure was randomized such that no group was exposed to the designs in the same order as any other group. See Table 1 for the exposure schedule adopted.

Table 1. Latin Square [110] exposure schedule starting with random assignment of groups to designs, then arranging assignment such that each group was exposed to each design in a different order.

Letters A through D represent the group assigned to that design for that wave.

Scheduling the exposure pattern for groups in this way ensured that sequence effects and time effects were both controlled for when the experiment was analysed over all four waves. This assumption is explicitly validated in the analysis phase.

Office designs.

Office designs were designed in workshops with a working group of 10 employees led by’s Real Estate team and CBRE, a commercial workplace design consultancy. The overall aim of the process was stated to be ‘to design an office space that is healthier, happier, and more productive’. Using a heuristic needs gathering process, three office designs were developed for testing that were proposed to suit the working style of the employee population. With the addition of the Open-plan office design as a control design, four designs were chosen for testing in the final experiment.

All office designs were implemented on the same floor of a single building in Amsterdam such that approximately a quarter of the available floor space was occupied by each design (see Fig 2 for a floor plan of the experimental area). Participants were informed through a variety of communication channels that this area was being monitored through sensors, and that they were able to opt out by leaving the space.

  1. Open-plan (Control)
    The Open-plan design is very similar to many large technology based companies around the world, and serves as a control design for this study (Fig 3). In the present study it is an open-plan design with groups of six desks, allowing three people to sit on either side.
  2. Zoned open-plan
    The Zoned open-plan design was proposed initially by project architects as their first answer to the needs gathering process. Zoned open-plan is similar to the open-plan design for working spaces, but added soundproof doors between working and collaboration spaces (Fig 4). Each Zoned open-plan zone had no more than 40 people in a room (compared to a maximum of 72 occupants), and each room included at least two ‘phone booth’ style units with soundproof doors. Plants were integrated into the space using hanging planter boxes above each set of desks.
  3. Activity based
    The Activity based design is an open-plan design in which desks are not officially assigned to a specific employee and includes activity-centered zones (Fig 5). Other spaces are provided such as small, one person rooms with desks and screens (known as ‘focus rooms’), phone booths, and a variety of collaboration spaces of different sizes and levels of privacy.
  4. Team office
    The Team office design is the closest to the traditional design of the cubicle, with each cubicle large enough to sit six or four people(Fig 6). Team offices each contained six of four desks, and included a whiteboard and large screen for sharing content. Each space was delineated by walls of sound absorbing panels.

Fig 2. Floorplan of the experimental floor.

Blacked out areas represent empty spaces (e.g. atriums) and non-experimental areas. Coloured areas represent experimental zones Control (red), Limited open plan (purple), Zoned open plan (yellow), and Team offices (blue).

Fig 3. Open-plan (Control).

Digital render of the Open-plan Control area.

Fig 4. Zoned open plan.

Digital render of the Zoned Open-plan design.

Fig 5. Activity-based design.

Digital render of the Activity-based design.

Fig 6. Team office design.

Digital render of the Team office design.


The experiment measured outcomes over 9 constructs. Operational definitions are defined in Table 2.

Table 2. Definitions and measurements of constructs considered in the experiment.

Healthy ranges for environmental variables were chosen based on suggestions from the literature [4547]. Table 3 contains the final ranges that were chosen for this experiment.

Table 3. Ranges of environmental variables considered to be healthy in this experiment.

A generalized linear model is adopted for the analysis of each of the outcome measures, stating the effect of time, group membership, carryover effect, and treatment effect explicitly [109, 110]. The model is formally defined as: (1) where:

g = The appropriate link function for the response distribution.

Yij = The value of outcome variable Y for time i in group j.

μ = The overall response mean.

πi = The effect of time i.

sj = The effect of the jth group.

τd(i, j) = The effect of the design d at period i to group j. Design d corresponds to the designs referred to in Table 1.

ρd(i−1,j) = The carryover effect applied at period i − 1 to the group j, under the design d.

βm = The effect of covariate m on the outcome Y.

For perception based outcomes, the covariates introduced to the model were based on demographic features that were believed to be of importance to determining the experience of the office space, such as age, gender, workgroup (i.e. company department), nationality, introversion, preference for morning or evening work, and the level of collaboration required to work.

Most response variables were normally distributed, and thus most models are linear models (identity link). Git commits were Poisson distributed, so a log link function is chosen. For each outcome the overall effect of each term is examined using an ANOVA. In cases where office design has a significant effect on the outcome variable, we examine the coefficients of the generalised linear model for all terms that were shown to have a significant effect on improving the fit of the model to data.



Two brands of sensor, ERS (82 sensors) and PointGrab CogniPoint (282 sensors), were used to measure environmental and behavioural variables. Each sensor type was administered by a third party sensor specialist who collected and collated the data before sending it for analysis by Data Scientists. Sensor locations were recorded by the installation company as a map, then these points were manually input to the QGIS software [113] to ensure each set of readings was associated with specific coordinates.

ERS sensors [114] measure a set of environmental variables (light, temperature, noise, movement, CO2 levels) simultaneously, and send these measurements to a receiver which in turn transmits data to servers. This process occurs regularly at 15 minute intervals.

PointGrab CogniPoint [115] sensors are used to sense desk occupancy. These sensors contain infrared cameras and have functionality to set a zone of interest within the video frame. They produce a count of the number of people seen within the zone of interest once every 5 minutes.


A questionnaire was administered to all participants at the end of each wave of exposure. All responses were required to be anonymous, so respondents provided the design to which they were assigned for that wave and responses were analysed at the group level. Differences between groups were controlled by the repeated-measures experiment design.

The questionnaire was designed based on past studies measuring both individual outcomes as well as group outcomes of environmental change. Satisfaction and perceived productivity were measured with 1 item, engagement with 3 items, enjoyment with 3 items, energy with 3 items, and flow with 4 items (see Table 4).

Git commits.

Git commits were collected as a loose proxy for technical output. This was achieved using a bash script that was run once at the end of the project using:

git shortlog -sne

with some extra parameters to filter counts for only experimental participants.

Data analysis and preparation

Data were analysed using R based on CSV files from sensor providers (for environmental measurements) and Qualtrics [116] (for survey responses). Data from each wave were engineered into a database in Hive to allow for modelling. Sensors were hung in the space for 2 weeks during a beta test to see both functionality of the devices, results of the space and behavior. It was established during this time that sensor measurements could be limited to 8 AM to 6 PM on weekdays, as no employees were in the office outside those times. This was confirmed in a check of the data during the experiment.

Occupancy data were aggregated by hour to calculate the average proportion of occupancy per design for that period. This decision was made to normalise the number of measurements taken per period, as each sensor provided readings asynchronously (such that there was a different number of readings per sensor within each hour window) and due to the Activity based design having a lower number of occupancy sensors in the design (due to the desk-sharing ratio introduced). In addition, a binomial treatment of the data would have considered each measurement to be independent, which is not the case for occupancy data taken from sensors, where each reading is both time and sensor dependent. Following an assessment of the data choices, the hourly aggregation method was considered most actionable by stakeholders. Analyses of the data in both raw and aggregated forms are provided in the supporting data for the reader’s convenience.



The questionnaire had a declining response rate over the course of the experiment (Fig 7). As responses were to be analysed at the group level, this decrease in responses over time was not considered to be a major problem as time is controlled through the repeated measures experiment design.

Fig 7. Response rates.

The response rate for each survey wave by design.

For each outcome a GLM was used to determine the effect of the experimental design on that outcome while also controlling for covariates. The effect of experimental design was found to be significant for all outcomes except for energy and commits, which are subsequently excluded from reporting in results. Covariates with significant effects are reported here, and non-significant effects are reported in the supplementary information.


A linear model revealed a significant effect of office design on participant satisfaction with the workplace (F = 39.958, p < 0.001) (Table 5). Participants reported 12% higher satisfaction in the Zoned open-plan design (est. = 1.169, se = 0.295, p < 0.001) and 8% higher satisfaction in the Team office design (est. = 0.781, se = 0.287, p = 0.007) than in the Open-plan design (Table 6). Participants reported 14% lower satisfaction in the Activity based design (est. = -1.391, se = 0.307, p < 0.001) than in the Open-plan design.

Of the covariates, only the number of days in the concept had a significant effect on satisfaction (F = 6.275, p = 0.012). Further exploration of the covariates revealed that each additional day participants had spent in the concept was associated with 3% higher satisfaction (est. = .303, se = 0.121, p = 0.012).


A linear model revealed a significant effect of office design on employee engagement in the workspace (F = 3.279, p = 0.0208) (Table 7). Participants reported 7% higher engagement in the Team office design than in the Open-plan design (est. = 0.701, se = 0.293, p = 0.017). The Activity based and Zoned open-plan designs did not receive significantly different engagement scores to the Open-plan design (est. = -0.383, se = 0.312, p = 0.221, and est. = 0.461, se = 0.301, p = 0.126 respectively).

Of the covariates, age (F = 7.425, p = 0.007), workgroup (F = 2.726, p = 0.013), introversion (F = 9.763, p < 0.001), preferred time of day to work (F = 3.71, p < 0.001), and the level of collaboration required for participants’ tasks (F = 4.591, p = 0.011) all had significant effects on engagement (Table 8).

Further exploration of the covariates revealed that ratings of engagement increased by.4% for each year of participant age (est. = 0.048, se = 0.019 p = 0.011). Participants in the finance department reported 8% higher average engagement (est. = 0.791814 se = 0.393079 p = 0.045). Participants who identified themselves as extroverts reported 5% higher engagement than those who identified as introverts or halfway between (est. = 0.455, se = 0.160, p = 0.014). Participants who identified themselves as morning people reported 3% higher engagement than those who identified as afternoon or evening people (est. = -0.298, se = 0.121, p = 0.014). Participants whose tasks required high levels of collaboration reported 9% higher engagement than those whose tasks required medium or low collaboration (est. = 0.994, se = 0.333, p = 0.003).


A linear model revealed a significant effect of office design on participant enjoyment of the workspace (F = 15.742, p < 0.001) (Table 9). Participants reported 11% more enjoyment in the Zoned open-plan design than in the Open-plan design (est. = 1.071, se = 0.364, p = 0.003). Conversely, participants reported 9% lower enjoyment in the Activity based design than in the Open-plan design (est. = -0.896, se = 0.378, p = 0.018). The Team office design did not receive significantly different enjoyment responses to the Open-plan design (est. = 0.649, se = 0.355, p = 0.068) (Table 10).

Of the covariates, the effect of the previous wave (F = 3.584, p < 0.001), gender (F = 15.742, p = 0.007), nationality (F = 2.222, p = 0.016), workgroup (F = 2.3033, p = 0.033), preferred time of day to work (F = 5.173, p = 0.023), and days spent in the concept (F = 7.096, p = 0.008) were all significant in the model.

Further exploration of the covariates revealed that the carryover effect of the previous office design was only significant in the first wave (i.e. when there was no previous design), such that participants in their first wave of the experiment reported 10% higher enjoyment of the workspace, regardless of office design (est. = 0.995, se = 0.388, p = 0.010). Participants who identified as female reported 3% higher enjoyment than participants who identified as male (est. = 0.305, se = 0.269, textitp = 0.258). Participants who worked in the HR department reported 10% higher enjoyment than those who worked in other departments (est. = 0.968, se = 0.459, p = 0.035). Participants who identified themselves as morning people reported 4% higher enjoyment of the workspace than those who identified as afternoon or evening people (est. = -0.363, se = 0.148, p = 0.0143). Each additional day participants had spent in the concept was associated with 4% higher enjoyment (est. = 0.395, se = 0.148, p = 0.008).


A series of linear models revealed no significant effects of office design on aggregated environmental variables (F = 20.629, df = 3) (Table 11). However, the Open-plan design recorded sound peaks outside of healthy ranges 20% more often than the Team Office design (est. = 0.195, se = 0.029, p = < 0.001), 29% more often than the Zoned open-plan design (est. = 0.291, se = 0.030, p = < 0.001), and 27% more often than the Activity based office design (est. = 0.269, se = 0.030, p = < 0.001) (Tables 12 and 13).

The Activity based design recorded an average temperature that was 0.4% lower than the Open-plan design, which was statistically significant (est. = -0.041, se = 0.014, p = 0.025) (Table 14). No significant differences were found for light, humidity, or air quality between the office space designs.


Office design did not have a significant effect on self reported energy levels at work (F = 1.439, p = 0.231) (Tables 15 and 16).


A linear model revealed a significant effect of office design on flow (F = 20.529, p < 0.001) (Table 17). Participants reported 12% higher flow in the Team office design (est. = 1.247, se = 0.281, p < 0.001) and 15% higher flow in the Zoned open-plan design (est. = 1.531, se = 0.288, p < 0.001) than in the Open-plan design. The Activity based office design did not receive significantly different flow responses to the Open-plan design (est. = -0.081, se = 0.299, p = 0.785).

Of the covariates, the effect of the previous wave (F = 4.823, p < 0.001) and the number of days spent in the concept (F = 5.871, p = 0.016) contributed significantly to the explanatory power of the model (Table 18).

Further exploration of the covariates revealed that the carryover effect of the previous office design was only significant in the first wave (i.e. when there was no previous design), such that participants in their first wave of the experiment reported 6% higher flow, regardless of office design (est. = 0.559, se = 0.307, p = 0.069). Each additional day participants had spent in the concept was associated with 3% higher flow (est. = 0.284, se = 0.117, p = 0.016).


A linear model revealed a significant effect of office design on productivity (F = 31.570, p < 0.001) (Table 19). Participants reported reported 10% higher productivity in the Team office (est. = 1.032, se = 0.394, p = 0.009) and 17% higher productivity in the Zoned open-plan (est. = 1.715, se = 0.404, p < 0.001) designs than in the Open-plan design. Conversely, participants reported 14% lower productivity in the Activity based design than in the Open-plan design (est. = -1.424, se = 0.420, p < 0.001).

Of the covariates, the effect of having no previous design (F = 3.618, p = 0.006), and the number of days spent in the concept (F = 4.473, p = 0.035) had a significant effect on the explanatory power of the model (Table 20). Further exploration of the covariates revealed that participants reported 9.4% higher productivity in their first wave of the experiment regardless of office design (est. = 0.943, se = 0.431, p = 0.029). Each additional day participants had spent in the concept was associated with 3.5% higher productivity (est. = 0.349, se = 0.165, p = 0.035).


Office design was not found to have an effect on the number of git commits made by participants (χ2 = 0.372, df = 259) (Tables 21 and 22).


A linear model revealed a significant effect of office design on hourly occupancy (F = 20.842, p = < 0.001) (Table 23). Sensors in the Activity based design recorded a 5% higher hourly occupancy than the Open-plan design (est. = 0.052, se = 0.008, p = < 0.001) (Table 24). There were no significant differences in hourly occupancy between the Open-plan design and the Zoned open-plan or Team office designs.

Of the covariates, both the wave of the study (F = 18.871, p = < 0.001) and the previous design (F = 6.981, p = < 0.001) had a significant effect on hourly occupancy.

Further exploration of the covariates revealed that the carryover effect of the previous office design was significant in the first wave, such that sensors recorded 6% higher hourly occupancy in the first wave of the experiment, across all designs (est. = 0.062, se = 0.009, p = < 0.001). Additionally, moving out of the Open-plan design was associated with a 4% increase in hourly occupancy (est. = 0.036, se = 0.009, p = < 0.001). Sensors recorded 2% lower hourly occupancy for each wave of the experiment (est. = -0.018, se = 0.004, p = < 0.001), indicating that hourly occupancy overall decreased as the experiment progressed.


While this research yielded many results, three specific findings should be of interest to researchers and practitioners.

First, our results demonstrated that office designs can affect employees’ satisfaction, engagement, enjoyment, flow, and productivity. The Zoned open-plan design was highest rated in terms of employees’ satisfaction, enjoyment, flow, and self-reported productivity. The Team Office design was highest rated for engagement and additionally was higher rated for satisfaction, flow, and self-reported productivity than the Open-plan or Activity based designs. The Open-plan office did not perform better than any other office designs, on any of these outcomes. Additionally, moving out of the Open-plan office was associated with an increase in desk occupancy, suggesting that employees’ preference for other designs was matched by their actual behaviour in showing up more when they no longer had to participate in the Open-plan design of the experiment. Finally, the sensors in the Activity based design reported a 5% increase in the proportion of desks that were occupied compared to other designs. However, given that the Activity based design included 27% fewer workstations than other designs, we interpret this finding to mean that desks were harder to find in this design, rather than that more people were occupying the area. The Activity based design did not perform better than other designs on any other outcomes and was worse than the Open-plan office for employee satisfaction, enjoyment, and productivity.

Second, noise was the main environmental variable to differentiate office designs. The noise levels in the Open-plan office were within healthy ranges 20% and 30% less often than in the other three designs, indicating a large difference between Open-plan and other designs in terms of noise. Thus, part of the reason for employees’ more positive ratings of other office designs may be that these are less distracting, and more comfortable, compared to Open-plan. While it is well established that quieter work spaces facilitate productivity, this research demonstrates the critical role of office design in managing noise.

Third, the consistency of results demonstrated that, in general, office design can be experimentally tested in a working office without disrupting business continuity. The cross-over design method proved to be a practical and informative method for office experimentation. However, we did observe higher enjoyment, flow, and occupancy when participants had not been in any other experimental design previously. This observation suggests that simply running an experiment on office design with employees may have a positive effect on enjoyment of the space, engagement, and occupancy. While this experimental design is noted to be be not as efficient as other designs [117], the advantages in overcoming commercial and logistic challenges make it a valuable tool in the practitioners toolkit.

We additionally observed some interesting findings from our exploratory analysis of the covariates. First, employees differed in their reported engagement with their work. Specifically, employees whose jobs required high levels of collaboration also reported high levels of engagement with their work, and this effect was slightly larger than the effect for office design. Additionally, employees identifying as extroverts and employees identifying as morning people reported higher engagement with their work. Finally, some employees reported higher enjoyment of their work than others, specifically, women, people working in HR, and people whose work required high levels of collaboration. However, in general, the effect of office design was larger than that of any employee covariates, suggesting that office design is of equal or greater importance to employee satisfaction, engagement, flow, and productivity than any personal variables.

We encountered some challenges in the process of conducting this experiment. First, the real-world setting prevented fully randomized selection of participants, as certain teams were unable or unwilling to participate, and teams could not be separated for the experiment and so allocation was randomized at the team level. We controlled for this experimentally through a randomized exposure pattern, and statistically by explicitly including a term for group effects in the statistical models. Second, while the study included a diverse group in terms of nationality, the sample was on average younger than the general population, as they were drawn from a specific corporate population. Results may therefore not generalise to companies with an older population. We recommend repeating the experiment in other companies with different age ranges. Third, the exposure of participants to designs could not be strictly controlled. For example, if a meeting room could not be booked within one experimental area, sometimes a team might book a meeting room in a different experimental area. This is unavoidable in a real world setting, but did not happen commonly, and participants were discouraged from the practice. Fourth, as a corporate research project, anonymity was especially important and this prevented us from tracking individual respondents from one design to the next. Finally, the survey was conducted in English, and while English is the language spoken at the office, this may have introduced some bias in a population that speaks many languages natively.


This paper presents a cross-over experimental evaluation of open office designs in a working technology company. It builds on previous studies of employee experience and behaviour in the office by combining many different factors to determine an optimal office design. We found that Zoned open-plan and Team office designs improved employee satisfaction, enjoyment, flow, and productivity, while Activity based and Open-plan designs performed poorly by comparison. The Open-plan office design was rated more poorly by employees, had higher levels of unsafe noise, and once employees no longer had to be in the Open-plan office design of the experiment, they spent more time at their desks. Given that the Open-plan design is used in many major technology companies, these findings should be noted by the wider industry as it suggests that many companies could benefit from redesigning their offices. The results suggest that office design can have a significant impact on employee productivity, health, and wellbeing across a range of factors. This effect should be considered and studied by companies faced with the decision of how to design their office spaces, and factored into budgeting and design stages.


The authors acknowledge the reviewers for taking the time to provide their valuable comments and guidance, which have greatly helped them to improve the manuscript.


  1. 1. Bernstein ES, Turban S. The impact of the ‘open’ workspace on human collaboration. Philosophical Transactions of the Royal Society B: Biological Sciences. 2018;1753(373).
  2. 2. Shafaghat A, Keyvanfar A, Lamit H, Mousavi SA, Abd Majid MZ. Open Plan Office Design Features Affecting Staff’s Health and Well-being Status. Jurnal Teknologi. 2014;70(7).
  3. 3. Villanueva ME, Caroche ML, Ani A, Ramirez M. Impact of Office Design on Satisfaction of Government Employees in Laguna Provincial Capitol, Philippines. Advances in Social and Occupational Ergonomics. 2019; p. 398–408.
  4. 4. Feige A, Wallbaum H, Janser M, Windlinger L. Impact of sustainable office buildings on occupant’s comfort and productivity. Journal of Corporate Real Estate. 2013;15(1):7–34.
  5. 5. Nieuwenhuis M, Knight C, Postmes T, Haslam SA. The relative benefits of green versus lean office space: Three field experiments. Journal of Experimental Psychology: Applied. 2014;20(3):199–214. pmid:25068481
  6. 6. Brennan A, Chugh JS, Kline T. Traditional versus Open Office Design. Environment and Behavior. 2002;34(3):279–299.
  7. 7. Hatch MJ. The Symbolics of Office Design: An Empirical Exploration*. In: Symbols and Artifacts. Routledge; 2017. p. 129–146.
  8. 8. Sharif HR, Zafarmand S, Naeini HS, Etemadi F. The Effects of Physical Agents on Occupants’ Satisfaction in Office Environment. International Journal of Occupational Hygiene. 2016;8(3):136–144.
  9. 9. Richardson A, Potter J, Paterson M, Harding T, Tyler-Merrick G, Kirk R, et al. Office design and health: a systematic review. New Zealand Medical Journal. 2017;130(1467):39. pmid:29240739
  10. 10. Rolfö LV. Relocation to an activity-based flexible office—Design processes and outcomes. Applied Ergonomics. 2018;73:141–150. pmid:30098629
  11. 11. CRBE. 2019 EMEA Occupier Survey; 2019. Available from:
  12. 12. Kim J, Candido C, Thomas L, de Dear R. Desk ownership in the workplace: The effect of non-territorial working on employee workplace satisfaction, perceived productivity and health. Building and Environment. 2016;103:203–214.
  13. 13. Laurence GA, Fried Y, Slowik LH. “My space”: A moderated mediation model of the effect of architectural and experienced privacy and workspace personalization on emotional exhaustion at work. Journal of environmental Psychology. 2013;36:144–152.
  14. 14. Hoendervanger JG, De Been I, Van Yperen NW, Mobach MP, Albers CJ. Flexibility in use: Switching behaviour and satisfaction in activity-based work environments. Journal of Corporate Real Estate. 2016;18(1):48–62.
  15. 15. Rolfö L, Babapour Chafi M. Policies for sharing workspaces in activity-based flex offices. In: 48th Annual Conference of the Association of Canadian Ergonomists. Banff; 2017. p. 339–344. Available from:
  16. 16. Brookes MJ, Kaplan A. The office environment: Space planning and affective behavior. Human factors. 1972;14(5):373–391.
  17. 17. Sundstrom E, Burt RE, Kamp D. Privacy at work: Architectural correlates of job satisfaction and job performance. Academy of Management Journal. 1980;23(1):101–117.
  18. 18. Hedge A. The open-plan office: A systematic investigation of employee reactions to their work environment. Environment and Behavior. 1982;14(5):519–542.
  19. 19. Maher A, von Hippel C. Individual differences in employee reactions to open-plan offices. Journal of environmental psychology. 2005;25(2):219–229.
  20. 20. Desor JA. Toward a psychological theory of crowding. Journal of Personality and Social Psychology. 1972;21(1):79.
  21. 21. Hundert AT, Greenfield N. Physical space and organizational behavior: A study of an office landscape. In: Proceedings of the 77th annual convention of the American Psychological Association. vol. 1; 1969. p. 601–602.
  22. 22. Oldham GR, Rotchford NL. Relationships between office characteristics and employee reactions: A study of the physical environment. Administrative Science Quarterly. 1983;28(4):542–556.
  23. 23. Lashina T, Chraibi S, Despenic M, Shrubsole P, Rosemann A, van Loenen E. Sharing lighting control in an open office: Doing one’s best to avoid conflict. Building and Environment. 2019;148:1–10.
  24. 24. Evans GW, Johnson D. Stress and open-office noise. Journal of applied psychology. 2000;85(5):779. pmid:11055149
  25. 25. de Korte E, Kuijt-Evers L, Vink P. Effects of the office environment on health and productivity 1: auditory and visual distraction. In: International Conference on Ergonomics and Health Aspects of Work with Computers; 2007. p. 26–33.
  26. 26. Wargocki P, Wyon DP, Fanger O. Productivity is affected by the air quality in offices; 2000. Available from:
  27. 27. Seppanen O, Fisk WJ, Lei QH. Effect of temperature on task performance in an environment. Ernest Orlando Lawrence Berkeley National Laboratory, Berkeley, CA (US); 2006.
  28. 28. Katzev R. The impact of energy-efficient office lighting strategies on employee satisfaction and productivity. Environment and Behavior. 1992;24(6):759–778.
  29. 29. Roderick I. The politics of office design. Journal of Language and Politics. 2016;15(3):274–287.
  30. 30. Roberts AC, Yap HS, Kwok KW, Car J, Soh CK, Christopoulos GI. The cubicle deconstructed: Simple visual enclosure improves perseverance. Journal of Environmental Psychology. 2019;63:60–73.
  31. 31. Chen S. Harvesting a Culture of Design: A Review of Organizational Design Research. In: The Design Imperative. Cham: Springer International Publishing; 2019. p. 13–35. Available from:
  32. 32. Walker JP. Design, Learn, Repeat: Architecture to Promote Learning Organization Behavior. University of Cincinnati; 2017. Available from:
  33. 33. Kjellberg A, LandstrÖM ULF, Tesarz M, Söderberg L, Akerlund E. The effects of nonphysical noise characteristics, ongoing task and noise sensitivity on annoyance and distraction due to noise at work. Journal of Environmental Psychology. 1996;16(2):123–136.
  34. 34. Jahncke H, Hygge S, Halin N, Green AM, Dimberg K. Open-plan office noise: Cognitive performance and restoration. Journal of Environmental Psychology. 2011;4(31):373–382.
  35. 35. Sailer U, Hassenzahl M. Assessing noise annoyance: an improvement-oriented approach. Ergonomics. 2000;43(11):1920–1938. pmid:11105980
  36. 36. Sundstrom E, Town JP, Rice RW, Osborn DP, Brill M. Office noise, satisfaction, and performance. Environment and behavior. 1994;26(2):195–222.
  37. 37. Burge S, Hedge A, Wilson S, Bass JH, ROBERTSON A. Sick building syndrome: a study of 4373 office workers. The Annals of Occupational Hygiene. 1987;31(4A):493–504. pmid:3439759
  38. 38. Bergs J. The effect of healthy workplaces on the well-being and productivity of office workers. In: Proceedings of International Plants for People Symposium. Amsterdam; 2002. p. 12.
  39. 39. Environmental Protection Agency. Sick Building Syndrome. Washington: United States Environmental Protection Agency; 1991.
  40. 40. Lan L, Lian Z, Pan L. The effects of air temperature on office workers’ well-being, workload and productivity-evaluated with subjective ratings. Applied Ergonomics. 2010;42(1):29–36. pmid:20478555
  41. 41. Wagner A, Gossauer E, Moosmann C, Gropp T, Leonhart R. Thermal comfort and workplace occupant satisfaction—Results of field studies in German low energy office buildings. Energy and Buildings. 2007;39(7):758–769.
  42. 42. Choi J, Aziz A, Loftness V. Investigation on the impacts of different genders and ages on satisfaction with thermal environments in office buildings. Building and Environment. 2010;45(6):1529–1535.
  43. 43. Fang L, Wyon DP, Clausen G, Fanger PO. Impact of indoor air temperature and humidity in an office on perceived air quality, SBS symptoms and performance. Indoor air. 2004;14(7):74–81. pmid:15330775
  44. 44. Gunay HB, Shen W, Newsham G, Ashouri A. Modelling and analysis of unsolicited temperature setpoint change requests in office buildings. Building and Environment. 2018;133:203–212.
  45. 45. Occupational Health and Safety Australia. Offices: Temperature and humidity—what are the’rules’?; 2019. Available from:
  46. 46. Canadian Center for Health and Safety. Thermal Comfort for Office Work: OSH Answers; 2019. Available from:
  47. 47. Society for Human Resource Management. Are we required to keep the workplace a certain temperature?; 2019. Available from:
  48. 48. Tanabe Si, Nishihara N, Haneda M. Indoor Temperature, Productivity, and Fatigue in Office Tasks. HVAC&R Research. 2007;13(4):623–633.
  49. 49. Chang TY, Kajackaite A. Battle for the thermostat: Gender and the effect of temperature on cognitive performance. PLOS ONE. 2019;14(5):e0216362. pmid:31116745
  50. 50. Akimoto T, Tanabe Si, Yanai T, Sasaki M. Thermal comfort and productivity—Evaluation of workplace environment in a task conditioned office. Building and Environment. 2010;45(1):45–50.
  51. 51. Ishii H, Kanagawa H, Shimamura Y, Uchiyama K, Miyagi K, Obayashi F, et al. Intellectual productivity under task ambient lighting. Lighting Research & Technology. 2018;50(2):237–252.
  52. 52. Mills PR, Tomkins SC, Schlangen LJ. The effect of high correlated colour temperature office lighting on employee wellbeing and work performance. Journal of Circadian Rhythms. 2007;5(0):2. pmid:17217543
  53. 53. Kogo T, Tsujikawa M, Kiuchi Y, Nishino A, Hashimoto S. Model Predictive Control of Shallow Drowsiness: Improving Productivity of Office Workers. In: International Conference of the IEEE Engineering in Medicine and Biology Society; 2019. Available from:
  54. 54. Fuerst F, McAllister P. An Investigation of the Effect of Eco-Labeling on Office Occupancy Rates. Journal of Sustainable Real Estate. 2009;1(1):49–64.
  55. 55. Erickson VL, Lin Y, Kamthe A, Brahme R, Surana A, Cerpa AE, et al. Energy efficient building environment control strategies using real-time occupancy measurements. In: Proceedings of the First ACM Workshop on Embedded Sensing Systems for Energy-Efficiency in Buildings—BuildSys’09. New York, New York, USA: ACM Press; 2009. p. 19. Available from:
  56. 56. Duarte C, Wymelenberg KVD, Rieger C. Revealing occupancy patterns in an office building through the use of occupancy sensor data. Energy and Buildings. 2013;67:587–595.
  57. 57. Azar E, Menassa CC. A comprehensive analysis of the impact of occupancy parameters in energy simulation of office buildings. Energy and Buildings. 2012;55:841–853.
  58. 58. Sana SS, Panda S. Optimal sales team’s initiatives and pricing of pharmaceutical products. International Journal of Systems Science: Operations & Logistics. 2015;2(3):168–176.
  59. 59. Lyubovnikova J, Legood A, Turner N, Mamakouka A. How Authentic Leadership Influences Team Performance: The Mediating Role of Team Reflexivity. Journal of Business Ethics. 2017;141(1):59–70.
  60. 60. Taylor AA. Assessing the Effect of Stress on Employee’s Productivity at the Work Place and how It can be Managed. Presbyterian University College; 2018.
  61. 61. Esteve-Pérez S, Pieri F, Rodriguez D. Age and productivity as determinants of firm survival over the industry life cycle. Industry and Innovation. 2018;25(2):167–198.
  62. 62. Chan MF. Factors associated with perceived sleep quality of nurses working on rotating shifts. Journal of Clinical Nursing. 2009;18(2):285–293. pmid:19120755
  63. 63. Park E, Lee HY, Park CSY. Association between sleep quality and nurse productivity among Korean clinical nurses. Journal of Nursing Management. 2018;26(8):1051–1058. pmid:29855101
  64. 64. Ribeiro JP, Gomes GC, Mota MS, Silva CD, Fuculo PRB Junior, Ribeiro JP, et al. Productivity of subjectivity and autonomy of nursing professionals in Pediatrics. Revista Brasileira de Enfermagem. 2019;72(suppl 1):41–48. pmid:30942343
  65. 65. Meyer AN, Barton LE, Murphy GC, Zimmermann T, Fritz T. The Work Life of Developers: Activities, Switches and Perceived Productivity. IEEE Transactions on Software Engineering. 2017;43(12):1178–1193.
  66. 66. Welbourne TM, Johnson DE, Erez A. The role-based performance scale: Validity analysis of a theory-based measure. Academy of management journal. 1998;41(5):540–555.
  67. 67. Kardinasari R, Iskandar TZ, Nugraha Y, Jatnika R. Social Sensitivity Effect to Public Service Competence and Its Impact on the Head of Sub-district Performance in West Java Province. Journal of Psychology Research. 2019;9(1).
  68. 68. Avby G, Kjellström S. LearnOvation: an intervention to foster exploration and exploitation behaviour in health care management in daily practice. BMC Health Services Research. 2019;19(1):319. pmid:31113453
  69. 69. Spinellis D. Git. IEEE Software. 2012;29(3):100–101.
  70. 70. Scholtes I, Mavrodiev P, Schweitzer F. From Aristotle to Ringelmann: a large-scale analysis of team productivity and coordination in Open Source Software projects. Empirical Software Engineering. 2016;2(21):642–683.
  71. 71. Bird C, Rigby PC, Barr ET, Hamilton DJ, German DM, Devanbu P. The Promises and Perils of Mining Git. In: Proceedings of the 2009 6th IEEE International Working Conference on Mining Software Repositories. MSR’09. Washington, DC, USA: IEEE Computer Society; 2009. p. 1–10. Available from:
  72. 72. Leesman Consulting Group. Benchmark employee experience; 2019. Available from:
  73. 73. Platis C, Zoulias E. Organization Style and Its Effect on Employee Satisfaction and Personal Performance. In: Strategic Innovative Marketing. Springer, Cham; 2017. p. 151–158.
  74. 74. Girma S. The relationship between leadership style, job satisfaction and culture of the organization. International Journal of Applied Research. 2016;4(2):542–556.
  75. 75. Ohunakin F, Adeniji AA, Akintayo ID. Transactional Leadership Style and Employee Job Satisfaction among Universities’ Guest Houses in South-West Nigeria. In: 3rd International Conference On African Development Issues. Ota, Nigeria; 2016. p. 368–371.
  76. 76. Bhave DP, Glomb TM. The Role of Occupational Emotional Labor Requirements on the Surface Acting–Job Satisfaction Relationship. Journal of Management. 2016;42(3):722–741.
  77. 77. Patton D. Predictive Relationships Between School Counselor Role Ambiguity, Role Diffusion, and Job Satisfaction. Walden Dissertations and Doctoral Studies. 2019.
  78. 78. Lottrup L, Stigsdotter UK, Meilby H, Claudi AG. The Workplace Window View: A Determinant of Office Workers’ Work Ability and Job Satisfaction. Landscape Research. 2015;40(1):57–75.
  79. 79. Leder S, Newsham GR, Veitch JA, Mancini S, Charles KE. Effects of office environment on employee satisfaction: a new analysis. Building Research & Information. 2016;44(1):34–50.
  80. 80. Hongisto V, Haapakangas A, Varjo J, Helenius R, Koskela H. Refurbishment of an open-plan office—Environmental and job satisfaction. Journal of Environmental Psychology. 2016;45:176–191.
  81. 81. Demerouti E, Mäkikangas A. What predicts flow at work? In: Fullagar C, Delle Fave A, editors. Flow at Work: Measurement and Implications. 1st ed. Oxon: Routledge; 2017. p. 66–80. Available from:
  82. 82. Duke É, Montag C. Smartphone addiction, daily interruptions and self-reported productivity. Addictive Behaviors Reports. 2017;6:90–95. pmid:29450241
  83. 83. Privette G. Peak experience, peak performance, and flow: A comparative analysis of positive human experiences. Journal of Personality and Social Psychology. 1983;45(6):1361–1368.
  84. 84. Csikszentmihalyi M. Flow: the psychology of optimal experience. 6th ed. New York: Harper & Row; 1990. Available from:
  85. 85. Moneta GB. On the Measurement and Conceptualization of Flow. In: Advances in Flow Research. New York, NY: Springer New York; 2012. p. 23–50. Available from:
  86. 86. Bakker AB. The work-related flow inventory: Construction and initial validation of the WOLF. Journal of Vocational Behavior. 2008;72(3):400–414.
  87. 87. Fullagar CJ, Kelloway EK. Flow at work: An experience sampling approach. Journal of Occupational and Organizational Psychology. 2009;82(3):595–615.
  88. 88. Dimotakis N, Ilies R, Judge TA. Experience sampling methodology. In: Modern research methods for the study of behavior in organization. Routledge; 2013. p. 319–348.
  89. 89. Engeser S, Rheinberg F. Flow, performance and moderators of challenge-skill balance. Motivation and Emotion. 2008;32(3):158–172.
  90. 90. Demerouti E, Bakker AB, Sonnentag S, Fullagar CJ. Work-related flow and energy at work and at home: A study on the role of daily recovery. Journal of Organizational Behavior. 2012;33(2):276–295.
  91. 91. Nielsen K, Cleal B. Predicting flow at work: Investigating the activities and job characteristics that predict flow states at work. Journal of Occupational Health Psychology. 2010;15(2):180–190. pmid:20364915
  92. 92. Salanova M, Bakker AB, Llorens S. Flow at Work: Evidence for an Upward Spiral of Personal and Organizational Resources*. Journal of Happiness Studies. 2006;7(1):1–22.
  93. 93. Vischer JC, Wifi M. The Effect of Workplace Design on Quality of Life at Work. In: Fleury-Bahi G, Pol E, Navarro O, editors. Handbook of Environmental Psychology and Quality of Life Research. Cham: Springer International Publishing; 2017. p. 387–400.
  94. 94. González-Romá V, Schaufeli WB, Bakker AB, Lloret S. Burnout and work engagement: Independent factors or opposite poles? Journal of Vocational Behavior. 2006;68(1):165–174.
  95. 95. Parry T, Sherman B. A Pragmatic Approach for Employers to Improve Measurement in Workforce Health and Productivity. Population Health Management. 2012;2(15):61–64.
  96. 96. Maslach C, Jackson SE, Leiter MP. Manual—Maslach Burnout Inventory Fourth Edition; 2019. Available from:
  97. 97. Demerouti E, Bakker AB, de Jonge J, Janssen PP, Schaufeli WB. Burnout and engagement at work as a function of demands and control. Scandinavian journal of work, environment & health. 2001;27(4):279–86.
  98. 98. Xanthopoulou D, Bakker AB, Dollard MF, Demerouti E, Schaufeli WB, Taris TW, et al. When do job demands particularly predict burnout? Journal of Managerial Psychology. 2007;22(8):766–786.
  99. 99. Escribà-Agüir V, Martín-Baena D, Pérez-Hoyos S. Psychosocial work environment and burnout among emergency medical and nursing staff. International Archives of Occupational and Environmental Health. 2006;80(2):127–133. pmid:16710712
  100. 100. Charoensukmongkol P, Moqbel M, Gutierrez-Wirsching S. The role of co-worker and supervisor support on job burnout and job satisfaction. Journal of Advances in Management Research. 2016;13(1):4–22.
  101. 101. Papierska A. An office design and employee stress: an office design that reduces employee’s stress and increases employee productivity. Universidade de Lisboa; 2018.
  102. 102. Jennings BM. Work Stress and Burnout Among Nurses: Role of the Work Environment and Working Conditions. Agency for Healthcare Research and Quality (US); 2008. Available from:
  103. 103. Heerwagen JH. Design, Productivity and Well Being: What are the Links? In: American Institute of Architects Conference on Highly Effective Facilities. Cincinatti: American Institute of Architects; 1998. p. 23. Available from:
  104. 104. Foley B, Engelen L, Gale J, Bauman A, Mackey M. Sedentary Behavior and Musculoskeletal Discomfort Are Reduced When Office Workers Trial an Activity-Based Work Environment. Journal of Occupational and Environmental Medicine. 2016;58(9):924–931. pmid:27454397
  105. 105. Alexandersson A, Kalonaityte V. Playing to Dissent: The Aesthetics and Politics of Playful Office Design. Organization Studies. 2018;39(2-3):297–317.
  106. 106. Miller AM. Fun in the workplace: Toward an environment behavior framework relating office design, employee creativity, and job satisfaction. University of Florida; 2005.
  107. 107. Cisco Systems. Office Design Case Study: How Cisco Designed the Collaborative Connected Workplace Environment—Cisco on Cisco—Cisco; 2004. Available from:
  108. 108. Worsham AP. Effects of office design on personal control and its correlates; 1983. Available from:
  109. 109. Bose M, Dey A. Optimal Crossover Designs. World Scientific; 2009.
  110. 110. Bose M, Dey A. Crossover Designs. In: Handbook of Design and Analysis of Experiments. CRC Press; 2015. p. 159–196.
  111. 111. Kenward MG, Jones B. Design and Analysis of Cross-Over Trials. 3rd ed. CRC Press; 2009.
  112. 112. Wu CF, Hamada M. Experiments: Planning, Analysis, and Optimization. 2nd ed. Wiley; 2009.
  113. 113. QGIS Development Team. QGIS Geographic Information System; 2019. Available from:
  114. 114. Elsys se. home site; 2019. Available from:
  115. 115. PointGrab. Cognipoint sales and specifications; 2019. Available from:
  116. 116. Qualtrics. Qualtrics; 2018. Available from:
  117. 117. Jankar J, Mandal A, Yang J. Optimal Crossover Designs for Generalized Linear Models. Journal of Statistical Theory and Practice. 2020.