ArticlePDF Available

The impact of the ‘open’ workspace on human collaboration



Organizations’ pursuit of increased workplace collaboration has led managers to transform traditional office spaces into ‘open’, transparency-enhancing architectures with fewer walls, doors and other spatial boundaries, yet there is scant direct empirical research on how human interaction patterns change as a result of these architectural changes. In two intervention-based field studies of corporate headquarters transitioning to more open office spaces, we empirically examined—using digital data from advanced wearable devices and from electronic communication servers—the effect of open office architectures on employees' face-to-face, email and instant messaging (IM) interaction patterns. Contrary to common belief, the volume of face-to-face interaction decreased significantly (approx. 70%) in both cases, with an associated increase in electronic interaction. In short, rather than prompting increasingly vibrant face-to-face collaboration, open architecture appeared to trigger a natural human response to socially withdraw from officemates and interact instead over email and IM. This is the first study to empirically measure both face-to-face and electronic interaction before and after the adoption of open office architecture. The results inform our understanding of the impact on human behaviour of workspaces that trend towards fewer spatial boundaries. This article is part of the theme issue ‘Interdisciplinary approaches for uncovering the impacts of architecture on collective behaviour’.
Cite this article: Bernstein ES, Turban S. 2018
The impact of the ‘open’ workspace on human
collaboration. Phil.Trans.R.Soc.B373: 20170239.
Accepted: 3 May 2018
One contribution of 11 to a theme issue
‘Interdisciplinary approaches for uncovering the
impacts of architecture on collective behaviour’.
Subject Areas:
behaviour, ecology
interaction, transparency, collaboration,
communication, spatial boundaries,
collective intelligence
Author for correspondence:
Ethan S. Bernstein
The impact of the ‘open’ workspace on
human collaboration
Ethan S. Bernstein1and Stephen Turban2
Harvard Business School, Boston, MA, USA
Harvard University, Cambridge MA, USA
ESB, 0000-0001-9819-0639
Organizations’ pursuit of increased workplace collaboration has led managers
to transform traditional office spaces into ‘open’, transparency-enhancing
architectures with fewer walls, doors and other spatial boundaries, yet there
is scant direct empirical research on how human interaction patterns change
as a result of these architectural changes. In two intervention-based field
studies of corporate headquarters transitioning to more open office spaces,
we empirically examined—using digital data from advanced wearable devices
and from electronic communication servers—the effect of open office
architectures on employees’ face-to-face, email and instant messaging (IM)
interaction patterns. Contrary to common belief, the volume of face-to-face
interaction decreased significantly (approx. 70%) in both cases, with an
associated increase in electronic interaction. In short, rather than prompting
increasingly vibrant face-to-face collaboration, open architecture appeared
to trigger a natural human response to socially withdraw from officemates
and interact instead over email and IM. This is the first study to empirically
measure both face-to-face and electronic interaction before and after the
adoption of open office architecture. The results inform our understanding
of the impact on human behaviour of workspaces that trend towards
fewer spatial boundaries.
This article is part of the theme issue ‘Interdisciplinary approaches for
uncovering the impacts of architecture on collective behaviour’.
1. Introduction
Boundaries between ‘us’ and ‘them’ have long captured human interest. Yet
even as social scientists continue to study the value of a vast array of bound-
aries [1], in an era in which the nature of work is changing [2– 4], managers
and organizational scholars have increasingly framed boundaries as barriers
to interaction that ought to be spanned [5–8], permeated [9] or blurred [10]
to increase collaboration. In the most physically salient and concrete example,
‘spatial boundaries’ [11] at work—such as office or cubicle walls—are being
removed to create open ‘unbounded’ offices in order to stimulate greater
collaboration and collective intelligence. Does it work?
Prior theory is divided—and empirical evidence mixed—on the effect that
removing spatial boundaries has on human behaviour in the space previously
within those boundaries (e.g. [12,13]). On the one hand, sociological theory pre-
sents a strong argument that removing spatial boundaries to bring more people
into contact should increase collaboration and collective intelligence. The notion
that propinquity, or proximity, predicts social interaction [14]—driving the for-
mation of social ties and therefore information exchange and collaboration—is
one of the most robust findings in sociology [15,16]. It has been observed in
contexts as diverse as the US Congress [17,18], nineteenth-century boarding
houses [19], college dormitories [14], laboratories [20], co-working spaces [21]
and corporate buildings [22]. When spatial boundaries—such as walls—are
removed, individuals feel more physically proximate, which, such theory
suggests, should lead to more interaction. Such interaction is a necessary
&2018 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution
License, which permits unrestricted use, provided the original
author and source are credited.
on July 2, 2018 from
foundation for collective intelligence—a form of distributed
intelligence that arises from the social interaction of individ-
uals [23] and that predicts, more so than the intelligence of
individual members, a group’s general ability to perform a
wide variety of tasks [24 26]. Much like the swarm intelli-
gence observed among cognitively simple agents such as
social insects and other animals [27 –29], collective intelli-
gence for groups of humans requires interaction [30]. If
greater propinquity drives greater interaction, it should
generate greater collaboration and collective intelligence.
On the other hand, some organizational scholars, espec-
ially social psychologists and environmental psychologists,
have shown that removing spatial boundaries can decrease
collaboration and collective intelligence. Spatial boundaries
have long served a functional role at multiple levels of
analysis, helping people make sense of their environment
by modularizing it [31], clarifying who is watching and
who is not, who has information and who does not, who
belongs and who does not, who controls what and
who does not, to whom one answers and to whom one
does not [32]. This school of thought, like theories of organ-
izational design and architecture [29], assumes that spatial
boundaries built into workspace architecture support collab-
oration and collective intelligence by mitigating the effects of
the cognitive constraints of the human beings working within
them. Like social insects which swarm within functionally-
determined zones ‘partitioned’ by spatial boundaries (e.g.
hives, nests or schools) [29], human beings—despite their
greater cognitive abilities—may also require boundaries to
constrain their interactions, thereby reducing the potential
for overload, distraction, bias, myopia and other symptoms
of bounded rationality. Research as far back as the founda-
tional Hawthorne Studies [33,34] shows that being walled
off can therefore increase interaction within the separated
group [33]. Similarly, subsequent workplace design research
(for reviews, see [35 –38])—though mixed in its findings—
suggests that open offices can reduce certain conditions
conducive to collaboration and collective intelligence,
including employee satisfaction [39,40], focus [41– 44],
psychological privacy [45,46] and other affective and
behavioural responses [40,41,43,47,48]. Such negative psycho-
logical effects of open offices conceivably may lead to less,
not more, interaction between those within them [49],
reducing collaboration and collective intelligence.
To our knowledge, no prior study has directly measured
the effect on actual interaction that results from removing spatial
boundaries to create an open office environment. Past work-
place design research, rather than directly and objectively
measuring behaviours, has relied heavily on survey-based
or activity-log methodologies, which provided self-reported
measures, or on social observation studies, which provided
an observer’s subjective interpretation of human interactions.
Several decades ago, when much of the workplace design
research was conducted, measuring actual interaction patterns
of individuals at work in both traditional and open office
environments would have been prohibitively difficult, but
new ‘people analytics’ technology has made it quite feasible.
Using two field studies of organizations transforming
their office architecture by removing spatial boundaries to
become more open, we empirically measure the effect on inter-
action, carefully tracking face-to-face (F2F) interaction before
and after the transition with wearable sociometric devices
[50,51] that avoid the imprecise and subjective survey-based
self-reported measures typical of previous office collaboration
studies [52,53]. We also measure two digital channels of inter-
action—email and instant messaging (IM) [54– 56]—using
information from the organizations’ own servers.
In the first study, we focus on the most basic set of empiri-
cal questions: what is the effect of transitioning from cubicles
to open workspaces on the overall volume and type of interac-
tion, with what implications for organizational performance
based on the company’s own performance management sys-
tem? In the second study, we replicate the first study’s results
and then consider two more-targeted empirical questions:
how does spatial distance between workstations moderate
the effect of transitioning from cubicles to open workspaces
and how do individual employee interaction networks, both
F2F and electronic, change differentially? While the first
study considers interactions involving individuals, the second
considers interactions for dyads (both sides of the inter-
action), allowing a more precise but limited investigation of
the effects.
2. Study 1
The first empirical study, a quasi-field experiment [57,58],
was conducted at the global headquarters of OpenCo1,
a Fortune 500 multinational. In a so-called war on walls,
OpenCo1 decided to use the latest open office workstation
products to completely transform the wall-bounded work-
spaces in its headquarters so that one entire floor was open,
transparent and boundaryless.
The redesign—which required people to move from
assigned seats on their original floor to similarly assigned
seats on a redesigned floor of the same size—affected employees
in functions including technology, sales and pricing, human
resources (HR), finance, and product development, as well as
the top leadership. Of those people, a cluster of 52 (roughly
40%) agreed to participate in the experiment. A comparison of
HR data for participants and nonparticipants provided no evi-
dence of nonresponse bias. Because of the nature of office
space,all employees moved fromthe old space to the redesigned
space at the same time, so the experiment wasstructured with an
interrupted time-series design [58].
To capture a full, data-rich picture of interaction patterns
both before and after the boundaries were removed, partici-
pants were asked to wear a sensor, known as a sociometric
badge [59], that recorded, in great detail, their F2F inter-
actions: an infrared (IR) sensor captured whom they were
facing (by making contact with the other person’s IR
sensor), microphones captured whether they were talking
or listening (but not what was said), an accelerometer cap-
tured body movement and posture, and a Bluetooth sensor
captured spatial location (figure 1). All sensors recorded
time-stamped data in 10 ms intervals. Based on prior research
using these sociometric badges [50], an F2F interaction was
recorded when three conditions were met: two or more
badges (i) were facing each other (with uninterrupted infra-
red line-of-sight), (ii) detected alternating speaking, and
(iii) were within 10 m of each other. The interaction ended
when any of the three criteria ceased to be true for more
than 5 s. While these criteria were based on precedent from
significant prior use of sociometric badges, sensitivity analy-
sis showed the results to be robust to reasonable alternative
assumptions (including shorter distances in 1 m increments, Phil. Trans. R. Soc. B 373: 20170239
on July 2, 2018 from
different lag times before concluding an interaction, and
different speaking patterns). This F2F data was combined
with email and IM data for the same time periods, collected
from the company’s servers, to create a full picture of these
professionals’ interactions before and after the redesign.
Data were collected in two phases: for 15 workdays (three
weeks) before the redesign and, roughly three months later,
for 15 workdays after the redesign. Three-week data collection
windows were chosen as a balance between the organiz-
ation’s desire to minimize the burden of the research study
on its employees and our need to control for the possibility
of idiosyncratic daily and weekly variations in employee
schedules. The three-month gap between phases was chosen
for two reasons. First, work at OpenCo1’s global head-
quarters followed quarterly cycles, so a three-month gap
allowed us to conduct the two data-collection phases at the
same point in the quarter. Second, it allowed just over two
months of adjustment after the move, enough for people to
have settled into their new environment but not so much
that the work they did could have changed much.
The dataset included 96 778 F2F interactions, 84 026 emails
(18 748 sent, 55012 received, 9755 received by cc and 511
received by bcc) and 25 691 IMs (consisting of 221 426 words).
The most straightforward and conservative empirical strategy
for analysing the intervention was to simply aggregate and
then compare pre-intervention and post-intervention volumes:
Yit ¼
1Postit ÞþXperson fixed effects þ1it :ð2:1Þ
, the dependent variable, is the amount of interaction—F2F
or electronic—where ‘i’ is the individual in question and ‘t’ is
the phase (pre- or post-redesign). Post
is an indicator variable
that equals 1 if the interaction occurred after the redesign. The
main estimation used ordinary least-squares (OLS) regressions
with person fixed effects, although all results were robust to
the exclusion of person fixed effects. Standard errors were cor-
rected for autocorrelation and clustered by individual [60]. If
the redesign increased F2F interaction, we should see a posi-
tive and significant
— the coefficient reported in the ‘Post’
column of table 1—when Y
is F2F interaction (the first row
of table 1). More generally, in table 1, the effect on a particular
kind of interaction due to the transition to more open architec-
ture is reported in the ‘post’ column, where a negative number
indicates reduced interaction and a positive number indicates
increased interaction.
(a) Study 1 results
(i) Volume of interaction
Although OpenCo1’s primary purpose in opening up the
space had been to increase F2F interactions, the 52 participants
now spent 72% less time interacting F2F. Prior to the redesign,
they accumulated 5266 min of interaction over 15 days, or
roughly 5.8 h of F2F interaction per person per day. After the
redesign, those same people accumulated only 1492 min of
interaction over 15 days, or roughly 1.7 h per person per day.
Even though everyone on the floor could see everyone else
all the time (or perhaps because they could), virtual interaction
replaced F2F interaction in the newly boundaryless space.
After the redesign, participants collectively sent 56% (66)
more emails to other participants over 15 days, received 20%
(78) more emails from other participants, and were cc’d
on 41% (27) more emails from other participants. (For the
received and cc’d volumes, emails sent are counted once for
each recipient.) Bcc: activity, which was low in volume and
limited to a small subset of individuals, did not significantly
change. IM message activity increased by 67% (99 more mess-
ages) and words sent by IM increased by 75% (850 more
words). Thus—to restate more precisely—in boundaryless
space, electronic interaction replaced F2F interaction.
(ii) Performance outcome
Should we be concerned about these effects? One indication
of the meaningfulness of this shift in behaviour was its
effect on performance. In an internal and confidential
management review, OpenCo1 executives reported to us
qualitatively that productivity, as defined by the metrics
used by their internal performance management system,
had declined after the redesign to eliminate spatial bound-
aries. Consistent with research on the impact of a decline in
media richness on productivity [54,55] and on the particular
challenges of email [61], it is not necessarily surprising that
productivity declined due to a substitution of email for F2F
interaction. What is surprising is that more open, transparent
architecture prompted such a substitution.
3. Study 2
Given the findings from Study 1, another organization was
recruited to further this research. Our goal was to conduct a
conceptual replication of the first study with a longer time
window. This second empirical study was also a quasi-field
experiment at a Fortune 500 multinational and was conducted
at the global headquarters of OpenCo2.
At the time of the
study, OpenCo2 was in the process of a multi-year head-
quarters redesign, which—as in Study 1—involved a
transformation from assigned seats in cubicles to similarly
assigned seats in an open office design, with large rooms of
desks and monitors and no dividers between people’s desks.
We again collected F2F data using sociometric badges
and email data from company servers, this time for 100
employees from a single floor, which was roughly 45% of
the employees on that floor. As in Study 1, data were
Figure 1. Sociometric badge. (Online version in colour.) Phil. Trans. R. Soc. B 373: 20170239
on July 2, 2018 from
collected in two phases: for eight weeks starting three months
prior to the redesign of this particular floor and for eight
weeks starting two months after the redesign. But for this
study, we also collected detailed data on the participants;
namely, three employee attributes—gender, team assignment
and role—and one architectural attribute—desk location. In
the first phase, desks were in cubicles, so seats were roughly
2 m apart and directly adjacent to one another. In the second
phase, seats still lay roughly 2 m apart and directly adjacent
to one another, but were grouped at undivided and unwalled
tables of six to eight. Seat location allowed us to calculate the
physical distance between dyads of employee workstations
before and after the redesign, such that we could include
physical distance, as well as the other employee attributes, as
control variables. The OpenCo2 dataset included 63 363 min
of F2F interaction and 25 553 emails, all generated by
1830 dyads—those with interaction—of the 100 employees
involved. Mindful of Study 1’s consistent results across
multiple forms of electronic communication, Study 2 only
collected email data to measure electronic interaction. The
empirical strategy was similar:
Yjt ¼
1PostjtÞþXdyad fixed effects þ1jt ð3:1Þ
Yjt ¼
2Physical DistancejtÞþ
In equation (3.1), as in equation (2.1), Y
, the dependent
variable, is the amount of interaction, F2F or electronic. How-
ever, because the physical-distance control variable was
dyadic, Y
must also be specific to a particular dyad ‘j’
(rather than to an individual ‘i’, as in Study 1). As in
Study 1, ‘t’ refers to the phase (pre- or post-redesign). Post
is an indicator variable that equals 1 if the dyadic interaction
occurred after the redesign. In equation (3.2), we investigate
specific control variables—characteristics of each dyad—
rather than just dyad fixed effects. Physical Distance
is the
distance between the dyad’s workstations, measured as the
shortest walking path (in metres). Gender
, Team
and Role
are indicator variables that equal 1 if the two individuals in
the dyad were of the same gender, on the same team, or in
the same role, and equal 0 otherwise. The main estimation
used OLS regressions with either dyad fixed effects (2) or
distance, gender, team and role controls (3). Standard errors
of the coefficients were corrected for autocorrelation and clus-
tered by dyad [60]. If the redesign increased F2F interaction,
we should see a positive and significant
—the coefficient
reported in the ‘post’ row of table 2—when Y
is F2F
interaction. More generally, in table 2, we report the effect of
the transition to open architecture on particular types of inter-
action in the ‘post’ row, where a negative number indicates
reduced interaction and a positive number indicates increased
interaction. For the control variables, we report the coefficient
for the entire sample without regard to whether the office
architecture involved cubicles or open spaces, as our purpose
in including those variables is to remove gender, team and
role effects from the variable of interest, Post. For example,
the significant and positive coefficient for Team means that
those on the same team communicated more than those on
different teams (for both cubicles and open spaces), and the
significant and positive coefficient for Role means that those
in the same role communicated more than those in different
roles (for both cubicles and open spaces).
Table 1. Impact of open offices on interaction at OpenCo1. Models are OLS with person fixed effects and with standard errors clustered by individual in
parentheses. Coefficients represent minutes of face-to-face (F2F) interaction, number of email messages or IM messages, or number of words in IM between a
member of the study and all others at work during the period of the study. *p,0.05; **p,0.01; ***p,0.001.
type of interaction post constant obs.
F2F interaction
minutes of F2F interaction time (indicated by proximity of individuals
combined with spoken words by at least one party)
email interaction (sent)
total number of emails sent by participants to other participants
email interaction (received: To)
total number of emails received by participants from other participants, where
the recipient appeared in the ‘To:’ field
email interaction (received: cc)
total number of emails received by participants from other participants, where
the recipient appeared in the ‘Cc:’ field
email interaction (received: bcc)
total number of emails received by participants from other participants, where
the recipient appeared in the ‘Bcc:’ field
IM interaction (number of messages)
total number of instant messages sent by participants to other participants
IM interaction (cumulative word count of messages)
total number of words sent in instant messages by participants to other participants
104 Phil. Trans. R. Soc. B 373: 20170239
on July 2, 2018 from
(a) Study 2 results
(i) Volume of interactions
As a result of the redesign, 643 dyads decreased their F2F
interaction and 141 dyads increased it. At the same time,
222 dyads decreased their email interaction and 374 dyads
increased it. Like OpenCo1, OpenCo2 had hoped, by opening
up the space, to increase F2F interactions, but the results did
not bear this out. The 100 employees—or 1830 dyads—we
tracked spent between 67% (Model 1, 12.79/17.99) and 71%
(Model 2, 9.81/14.63) less time interacting F2F. Instead, they
emailed each other between 22% (Model 3, 1.24/5.75) and
50% (Model 4, 1.54/3.07) more.
As one might suspect, dyads on the same team or with the
same role communicated more, both F2F and by email, relative
to dyads on different teams or in different roles. Gender, in con-
trast, had no significant effect on the volume of either form of
interaction. Physical distance did show a small inverse effect
on F2F interaction (Model 2): the nearer the two workstations,
the more F2F interaction. This effect was notable both for its
small size relative to the size of the effect of the open office
and for the fact that it was limited to F2F interaction (not
email). We investigate this in further detail next.
(ii) The effect of physical distance on F2F versus email
Model 2 of table 2 shows that the effect of physical distance on
F2F interaction is small—and the effect on email insignificant—
relative to that of openness. The relatively small effect of
distance on F2F interaction was surprising given that repeated
studies have shown that people talk more to those who are
physically closer to them [62,63]. When others are physically
proximate, it is easier to be aware of them [64], start conversa-
tions with them [64,65], unexpectedly encounter or overhear
them [66], and manage their impressions of our collaborative
work behaviour [67]. Nonetheless, our review of these prior
studies found none that directly measured interaction volumes,
and thus perhaps—while present—the effect of distance on F2F
interaction may be far more minimal than previously thought.
Table 2, however, does not allow us to compare the rela-
tive effects of physical distance on F2F interaction and on
email interaction. To do so, we used a latent space model
called the Latent Position Clustering Model [68] to take into
account clustering and to control for other covariates. We
find that physical distance affected F2F interaction twice
as much as it did email interaction. As a robustness check,
we used several machine learning algorithms, such as a
Random Forest, to see if changes in F2F networks prompted
by changes in physical distance predicted changes in email
networks. Across all models, we find that F2F networks
and email networks respond very differently to changes in
the built environment, with changes in one type of network
failing to predict changes in the other.
This variance between the adaptation of F2F and elec-
tronic networks in response to a change in physical space is
an important finding for future research on collaboration and
collective intelligence. In several notable cases, past research
has relied on email alone [69,70] to study topics ranging from
the Enron debacle to the relationship between office layout
and interaction, basing claims about F2F interaction on
findings from electronic interaction data. Our finding that
changes in workplace design affect electronic and F2F
interaction networks differently (and, on some measures, in
opposite directions) should make future researchers wary of
using one network as a proxy for the other.
4. Discussion
We began with a specific research question: does removing
spatial boundaries at work to create open, unbounded offices
Table 2. Impact of open offices on interaction at OpenCo2. Models are OLS with standard errors clustered by dyad in parentheses. Models 1 and 3 include dyad
fixed effects. In Models 1 and 2, coefficients represent minutes of F2F interaction between a particular dyad during the period of the study. In Models 3 and 4,
coefficients represent number of emails between a particular dyad during the period of the study. *p,0.05, **p,0.01, ***p,0.001.
type of interaction
123 4
F2F with fixed
F2F with
email with fixed
email with
change in volume:
0 if before redesign, 1 if after
physical distance
walking distance (in metres) between desks
0 if different genders, 1 if same
0 if different teams, 1 if same
0 if different roles, 1 if same
constant 17.99***
observations 3660 3660 3660 3660 Phil. Trans. R. Soc. B 373: 20170239
on July 2, 2018 from
increase interaction? Our two empirical field studies were
consistent in their answer: open, unbounded offices reduce
F2F interaction with a magnitude, in these contexts, of about
70%. Electronic interaction takes up at least some of the slack,
increasing by roughly 20% to 50% (as measured by ‘To:’
received email).
Many organizations, like our two field sites, transform their
office architectures into open spaces with the intention of
creating more F2F interaction and thus a more vibrant work
environment. What they often get—as captured by a steady
stream of news articles professing the death of the open office
[71– 73]—is an open expanse of proximal employees choosing
to isolate themselves as best they can (e.g. by wearing large
headphones [74]) while appearing to be as busy as possible
(since everyone can see them). Recent studies [75] and earlier
research [40,41,43,47,48] have investigated the self-reported dis-
satisfaction of employeesin openoffices, but to our knowledge,
we are the first to empirically study the direct behavioural
impact of open office space on the volume of F2F and electronic
interaction. Our results support three cautionary tales.
First, transitions to open office architecture do not
necessarily promote open interaction. Consistent with the
fundamental human desire for privacy [76] and prior evi-
dence that privacy may increase productivity [32,45], when
office architecture makes everyone more observable or ‘trans-
parent’, it can dampen F2F interaction, as employees find
other strategies to preserve their privacy; for example, by
choosing a different channel through which to communicate
[39]. Rather than have an F2F interaction in front of a large
audience of peers, an employee might look around, see that
a particular person is at his or her desk, and send an email.
The second caution relates to the impact of a transition to
open office architecture on collective intelligence. We still
have much to learn about how collective intelligence works
[77], as we borrow from and distinguish parallel work on
swarm intelligence among social insects and some other ani-
mals. While the earliest work assumed open spaces would
promote collective intelligence among humans, our findings
support more recent work that has begun to suggest otherwise.
Kao & Couzin, in modelling the presence of multiple cues and
the possibility of observing them, find that intermediate (rather
than maximal) levels of cues produce higher levels of collective
intelligence [78]. We see a close relationship between our find-
ing that open, ‘transparent’ offices may be overstimulating
and thus decrease organizational productivity and Kao &
Couzin’s demonstration that finitely bounded, and often
small, group size maximizes decision accuracy in complex,
realistic environments. Similarly, recent collective intelligence
work suggests that, like our open offices, too much information
from social data can be problematic, partly because of
challenges focusing attention [74,79], but also for reasons that
extend to more general functions of human cognition. For
example, by connecting human cognition and collective intelli-
gence with the behaviour of eusocial insects, Toyokawa et al.
found that richness in social information was detrimental to
collective intelligence outcomes, with performance being best
when social learning opportunities were constrained [80].
Similarly, in a study involving human subjects, Bernstein et al.
found that intermittent rather than constant social influence
produced the best performance among humans collectively
engaged in complex problem solving [81]. As we are reminded
in Hight & Perry’s article on collective intelligence and architec-
tural design, ‘collective intelligence is not simply technical, but
also explicitly social, political, and by extension, professional’
[2, p. 6]. Our findings empirically reinforce their caution that
the relationship between architectural design and collective
intelligence extends beyond technical considerations.
The third caution is that transitions to open office archi-
tecture can have different effects on different channels of
interaction. In our studies, openness decreased F2F inter-
action with an associated increase in email interaction. In
the digital age, employees can choose from multiple channels
of interaction [54] and a change in office architecture may
affect that choice.
Complementing prior research on media richness sug-
gesting that substituting email for F2F interaction can lower
productivity [53], our studies highlight two other conse-
quences. First, because fundamentally different mechanisms
drive F2F and email interaction, the physical propinquity that
redesigned offices seek to achieve has a direct effect only on
F2F interaction, not on email, yet drives interaction from F2F
to email. Adopting open offices, therefore, appears to have
the perverse outcome of reducing rather than increasing pro-
ductive interaction. Second, F2F and email networks differ.
Although prior studies have investigated one or the other
[56,82], none has empirically linked F2F and email network
interaction to discern how good a proxy one is for the other.
We find that they are poor proxies for each other. Therefore,
an intervention that redirects interaction from one network to
another, like the open office redesigns studied here, not only
changes the channel of interaction, but also skews whom a
person interacts with. That can have profound consequences
for how—and how productively—work gets done.
In summary, because the antecedents of human interaction
at work go beyond proximity and visibility, the effects of
open office architecture on collaboration are not as simple as
previously thought. While it is possible to bring chemical
substances together under specific conditions of temperature
and pressure to form the desired compound, more factors
seem to be at work in achieving a similar effect with humans.
Until we understand those factors, we may be surprised to
find a reduction in F2F collaboration at work even as we architect
transparent, open spaces intended to increase it.
Data accessibility. We are unable to provide open access to our data owing
to their sensitive nature and the nondisclosure and confidentiality
agreements that surround them. Please contact the corresponding
author for more information.
Authors’ contributions. E.S.B. carried out all work on Study 1 and drafted
the manuscript. S.T. carried out all work on Study 2 and helped draft
the manuscript. Both authors gave final approval for publication.
Competing interests. We declare we have no competing interests.
Funding. Funding for these studies was provided by the Division
of Research and Faculty Development at the Harvard Business School.
Acknowledgements. The authors thank Editor Steve Fiore and two anon-
ymous reviewers for developmental, insightful and encouraging
comments throughout the review process, as well as Senior Commis-
sioning Editor Helen Eaton for her guidance. We also thank Ben
Waber, Taemie Kim, Laura Freeman and the rest of the team at
Humanyze, without whom we would have been unable to collect
the unique datasets underlying these studies.
OpenCo1 is a pseudonym for the corporation’s real name, which has
been anonymized.
OpenCo2 is a pseudonym for the corporation’s real name, which has
been anonymized. Phil. Trans. R. Soc. B 373: 20170239
on July 2, 2018 from
1. Lamont M, Molna
´r V. 2002 The study of boundaries
in the social sciences. Ann. Rev. Sociol. 28, 167
195. (doi:10.1146/annurev.soc.28.110601.141107)
2. Hight C, Perry C. 2006 Collective intelligence in
design. Architect. Des. 76, 5– 9. (doi:10.1002/
3. Parker GG, Van Alstyne MW, Choudary SP. 2016
Platform revolution: how networked markets are
transforming the economy and how to make them
work for you. New York, NY: WW Norton.
4. Brynjolfsson E, McAfee A. 2014 The second machine
age: work, progress, and prosperity in a time of
brilliant technologies. New York, NY: WW Norton.
5. Rao H, Monin P, Durand R. 2005 Border crossing:
bricolage and the erosion of categorical boundaries
in French gastronomy. Am. Sociol. Rev. 70,
968–991. (doi:10.1177/000312240507000605)
6. Aldrich H, Herker D. 1977 Boundary spanning roles
and organization structure. Acad. Manag. Rev. 2,
217–230. (doi:10.5465/amr.1977.4409044)
7. Granovetter M. 1983 The strength of weak ties:
a network theory revisited. Sociol. Theory 1,
201–233. (doi:10.2307/202051)
8. Burt RS. 1992 Structural holes: the social structure of
competition. Cambridge, MA: Harvard University
9. Western M, Wright EO. 1994 The permeability of
class boundaries to intergenerational mobility
among men in the United States, Canada, Norway
and Sweden. Am. Sociol. Rev. 59, 606– 629.
10. Fox C, Guglielmo TA. 2012 Defining America’s racial
boundaries: blacks, Mexicans, and European
immigrants, 1890– 1945. Am. J. Sociol. 118,
327–379. (doi:10.1086/666383)
11. O’Leary MB, Cummings JN. 2007 The spatial,
temporal, and configurational characteristics of
geographic dispersion in teams. MIS Q. 31,
433–452. (doi:10.2307/25148802)
12. Fayard AL, Weeks J. 2011 Who moved my cube?
Harvard Business Rev. 89, 103110.
13. Ward R. 2017 If the walls could talk: at the
interface of design, materiality and culture. In
Academy of Management Symposium,Philadelphia,
PA,August 2017. no. 10012.
14. Festinger L, Schachter S, Back K. 1950 Social
pressures in informal groups: a study of human
factors in housing. Oxford, UK: Harper.
15. DiMaggio P, Garip F. 2012 Network effects and
social inequality. Ann. Rev. Sociol. 38, 93– 118.
16. Burt RS. 2004 Structural holes and good ideas.
Am. J. Sociol. 110, 349 399. (doi:10.1086/421787)
17. Chown JD, Liu CC. 2015 Geography and power in an
organizational forum: evidence from the US Senate
chamber. Strateg. Manag. J. 36, 177– 196. (doi:10.
18. Liu CC, Srivastava SB. 2015 Pulling closer
and moving apart: interaction, identity, and
influence in the US Senate, 1973 to 2009.
Am. Sociol. Rev. 80, 192217. (doi:10.1177/
19. Parigi P, Bergemann P. 2016 Strange bedfellows:
informal relationships and political preference
formation within boardinghouses, 18251841.
Am. J. Sociol. 122, 501 531. (doi:10.1086/688606)
20. Allen TJ, Cohen SI. 1969 Information flow in
research and development laboratories. Admin. Sci.
Q. 14, 12 19. (doi:10.2307/2391357)
21. Crosina E. 2018 Start me up: on becoming an
entrepreneur in a shared work space. Doctoral
dissertation, Boston College Carroll School of
22. Kleinbaum AM, Stuart TE, Tushman ML. 2013
Discretion within constraint: homophily and
structure in a formal organization. Organ. Sci. 24,
13161336. (doi:10.1287/orsc.1120.0804)
23. Le
´vy P. 1997 Collective intelligence. New York, NY:
Plenum/Harper Collins.
24. Woolley AW, Aggarwal I, Malone TW. 2015
Collective intelligence and group performance. Curr.
Dir. Psychol. Sci. 24, 420424. (doi:10.1177/
25. Malone TW, Bernstein MS. 2015 Handbook of
collective intelligence. Cambridge, MA: MIT Press.
26. Woolley AW, Chabris CF, Pentland A, Hashmi N,
Malone TW. 2010 Evidence for a collective
intelligence factor in the performance of human
groups. Science 330, 686– 688. (doi:10.1126/
27. Salminen J. 2012 Collective intelligence in humans:
a literature review. In Collective Intelligence Proc.,
Cambridge, MA,April 18 20 2012. Cambridge, MA:
28. Bonabeau E, Dorigo M, Theraulaz G. 1999 Swarm
intelligence: from natural to artificial systems.
New York, NY: Oxford University Press.
29. Ireland T, Garnier S. 2018 Architecture, space and
information in constructions built by humans and
social insects: a conceptual review. Phil. Trans. R.
Soc. B 373, 20170244. (doi:10.1098/rstb.2017.0244)
30. Kabo F. 2018 The architecture of network collective
intelligence: correlations between social network
structure, spatial layout and prestige outcomes in
an office. Phil. Trans. R. Soc. B 373, 20170238.
31. Baldwin CY, Clark KB. 2000 Design rules: the power
of modularity. Cambridge, MA: MIT Press.
32. Bernstein ES. 2012 The transparency paradox: a role
for privacy in organizational learning and
operational control. Admin. Sci. Q. 57, 181– 216.
33. Mayo E. 1933 Human problems of an industrial
civilization. New York, NY: Macmillan.
34. Roethlisberger FJ, Dickson WJ. 1939 Management
and the worker: an account of a research program
conducted by the Western Electric Company.
Cambridge, MA: Harvard University Press.
35. Brennan A, Chugh JS, Kline T. 2002 Traditional
versus open office design: a longitudinal field study.
Environ. Behav. 34, 279– 299. (doi:10.1177/
36. Becker F. 1991 Workplace planning, design, and
management. In Advances in environment, behavior,
and design (eds EH Zube, GT Moore), pp. 115 151.
New York, NY: Springer.
37. Sundstrom E, Sundstrom MG. 1986 Work places: the
psychology of the physical environment in offices and
factories. New York, NY: Cambridge University Press.
38. De Croon EM, Sluiter JK, Kuijer PPFM, Frings-Dresen
MHW. 2005 The effect of office concepts on worker
health and performance: a systematic review of the
literature. Ergonomics 48, 119– 134. (doi:10.1080/
39. Becker FD, Gield B, Gaylin K, Sayer S. 1983 Office
design in a community college: effect on work
and communication patterns. Environ. Behav. 15,
699726. (doi:10.1177/0013916583156002)
40. Oldham GR, Brass DJ. 1979 Employee reactions to
an open-plan office: a naturally occurring quasi-
experiment. Admin. Sci. Q. 24, 267284. (doi:10.
41. Brookes MJ, Kaplan A. 1972 The office environment:
space planning and affective behavior. Hum. Factors
14, 373 391. (doi:10.1177/001872087201400502)
42. Smith-Jackson TL, Klein KW. 2009 Open-plan offices:
task performance and mental workload. J. Environ.
Psychol. 29, 279– 289. (doi:10.1016/j.jenvp.2008.
43. Sundstrom E, Herbert RK, Brown DW. 1982 Privacy
and communication in an open-plan office: a case
study. Environ. Behav. 14, 379 392. (doi:10.1177/
44. Sundstrom E, Town JP, Brown DW, Forman A,
Mcgee C. 1982 Physical enclosure, type of job, and
privacy in the office. Environ. Behav. 14, 543– 559.
45. Sundstrom E, Burt RE, Kamp D. 1980 Privacy at
work: architectural correlates of job satisfaction and
job performance. Acad. Manag. J. 23, 101 117.
46. Hundert AT, Greenfield N. 1969 Physical space and
organizational behavior: a study of an office
landscape. In Proc. 77th Ann. Convention American
Psychological Association, vol. 1, pp. 601– 602.
Washington, DC: APA.
47. Oldham GR, Rotchford NL. 1983 Relationships
between office characteristics and employee
reactions: a study of the physical environment.
Admin. Sci. Q. 28, 542 556. (doi:10.2307/2393008)
48. Zalesny MD, Farace RV. 1987 Traditional versus open
offices: a comparison of sociotechnical, social
relations, and symbolic meaning perspectives. Acad.
Manag. J. 30, 240 259.
49. Hatch MJ. 1987 Physical barriers, task
characteristics, and interaction activity in research
and development firms. Admin. Sci. Q. 32,
387399. (doi:10.2307/2392911)
50. Kim T, McFee E, Olguin DO, Waber B, Pentland A.
2012 Sociometric badges: using sensor
technology to capture new forms of collaboration. Phil. Trans. R. Soc. B 373: 20170239
on July 2, 2018 from
J. Organ. Behav. 33, 412427. (doi:10.1002/
51. Chaffin D, Heidl R, Hollenbeck JR, Howe M, Yu A,
Voorhees C, Calantone R. 2017 The promise and
perils of wearable sensors in organizational
research. Organ. Res. Methods 20, 3– 31. (doi:10.
52. Marsden PV. 1990 Network data and measurement.
Ann. Rev. Sociol. 16, 435– 463. (doi:10.1146/
53. Marsden PV. 2011 Survey methods for network
data. In The SAGE handbook of social network
analysis (eds J Scott, PJ Carrington), pp. 370– 388.
Thousand Oaks, CA: Sage.
54. Daft RL, Lengel RH. 1984 Information richness: a new
approach to managerial behavior and organizational
design. Res. Organ. Behav. 6, 191– 233.
55. Daft RL, Lengel RH. 1986 Organizational information
requirements, media richness and structural
design. Manag. Sci. 32, 554 571. (doi:10.1287/
56. Yates J, Orlikowski WJ. 1992 Genres of
organizational communication: a structurational
approach to studying communication and media.
Acad. Manag. Rev. 17, 299326. (doi:10.5465/amr.
57. Campbell DT, Stanley JC. 1963 Experimental and
quasi-experimental designs for research. Chicago, IL:
Rand McNally.
58. Thyer BA. 2012 Quasi-experimental research designs.
Oxford, UK: Oxford University Press.
59. Choudhury T, Pentland A. 2003 Sensing and
modeling human networks using the sociometer.In
Proc. 7th IEEE Int. Symp. Wearable Computers, 21
23 October 2003, White Plains, NY (ISWC2003), pp.
216222. White Plains, NY: IEEE Computer Society.
60. Bertrand M, Duflo E, Mullainathan S. 2004 How
much should we trust differences-in-differences
estimates? Q. J. Econ. 119, 249– 275. (doi:10.1162/
61. Barley SR, Meyerson DE, Grodal S. 2011 E-mail as
a source and symbol of stress. Organ. Sci. 22,
887906. (doi:10.1287/orsc.1100.0573)
62. Allen TJ. 1977 Managing the flow of technology.
Cambridge, MA: MIT Press.
63. Kraut R, Edigo C, Galegher J. 1990 Patterns of
contact and communication in scientific research
collaborations. In Intellectual teamwork: social and
technological foundations of cooperative work (eds R
Galegher, R Kraut, C Egido), pp. 149 171. Hillsdale,
NJ: Lawrence Erlbaum Associates.
64. Bellotti V, Bly S. 1996 Walking away from the
desktop computer: distributed collaboration and
mobility in a product design team. In Proc. ACM
Conf. Computer Supported Cooperative Work,
Boston, MA, 620 November 1996 (CSCW ‘’96)
(ed. MS Ackerman), pp. 209218. New York, NY:
ACM Press.
65. Whittaker S, Frohlich D, Daly-Jones O. 1994
Informal workplace communication: what is it like
and how might we support it? In Proc. SIGCHI Conf.
Human Factors in Computing Systems, Boston, MA,
24 28 April 1994 (CHI ‘94) (eds B Adelson,
S Dumais, J Olson), pp. 131 137. New York, NY:
ACM Press.
66. Teasley S, Covi LA, Krishnan MS, Olson JS. 2002
Rapid software development through team
collocation. IEEE Trans. Software Eng. 28,671 683.
67. Birnholtz J, Dixon G, Hancock J. 2012 Distance,
ambiguity and appropriation: structures affording
impression management in a collocated
organization. Comp. Hum. Behav. 28, 1028– 1035.
68. Gormley IC, Murphy TB. 2010 A mixture of experts
latent position cluster model for social network
data. Stat. Methodol. 7, 385– 405. (doi:10.1016/j.
69. Michalski R, Palus S, Kazienko P. 2011 Matching
organizational structure and social network
extracted from email communication. In Lecture
notes in business information processing (ed.
W Abramowicz), pp. 197 – 206. New York, NY:
70. Zack MH. 2000 Researching organizational systems
using social network analysis. In Proc. 33rd Ann.
Hawaii Int. Conf. System Sciences (HICSS), Maui, HI,
4 7 January 2000, pp. 7 13. Hawaii: IEEE.
71. Faulhaber J. 2016 The post-cubicle office and its
discontents. New York Times, February 25, 2016. See
(accessed 9 October 2017).
72. Borzykowski B. 2017 Why open offices are bad for
us. BBC. See
(accessed 9 October 2017).
73. Giacobbe A. 2015 Open offices seem great– until
you work in one. Boston Globe, March 5, 2015. See
F2Zy3BqCfbMTm4Mn6gVBzH/story.html (accessed 9
October 2017). (doi:10.1098/rspb.2013.3305)
74. Birnholtz J, Gutwin C, Hawkey K. 2008 Privacy in
the open: how attention mediates awareness and
privacy in open-plan offices. Proc. Group ’07: Int.
Conf. Supporting Group Work, Sanibel Island, FL,
4 7 November 2007, pp. 51 60. New York, NY:
Association for Computing Machinery. (doi:10.1145/
75. Kim J, de Dear R. 2013 Workspace satisfaction: the
privacy-communication trade-off in open-plan
offices. J. Environ. Psychol. 36, 18– 26. (doi:10.
76. Bernstein E. 2017 Making transparency transparent:
the evolution of observation in management
theory. Acad. Manag. Ann. 11, 1– 50. (doi:10.5465/
77. Schut MC. 2010 On model design for simulation of
collective intelligence. Inf. Sci. 180, 132– 155.
78. Kao AB, Couzin ID. 2014 Decision accuracy in
complex environments is often maximized by small
group sizes. Phil. Trans. R. Soc. B 281, 20133305.
79. Bernstein E, Mortensen M, Leonardi P. 2018
Unbounded attention: the benefits of an attention-
based lens on work relationships. Harvard Business
School Work. Pap.
80. Toyokawa W, Kim H, Kameda T. 2014 Human
collective intelligence under dual exploration-
exploitation dilemmas. PLoS ONE 9, e95789. (doi:10.
81. Bernstein E, Shore J, Lazer D. 2018 How
intermittent breaks in interaction improve
collective intelligence. Harvard Business School
Work. Pap.
82. Quintane E, Kleinbaum AM. 2011 Matter over mind?
E-mail data and the measurement of social
networks. Connections 31, 22– 46. Phil. Trans. R. Soc. B 373: 20170239
on July 2, 2018 from
... The benefits of either design type appear to be mixed, and indeed a blending or semi-open type may yield the most positive effects. For example, Bernstein and Turban (2018) found that the presence of walls, partitions, or boundaries in offices can reduce interruptions from external sources but increase interaction within the separated group. Zamani (2018) reported that the lack of private spaces to have discussions with team members in the ED impacted team collaboration negatively. ...
... A high percentage of respondents in the low IC group perceived to have high collaboration with those in proximity, that is, with those seated within the physician workstation. Bernstein and Turban (2018) also reported similar findings within a colocated group of people that was separated from rest of the staff. However, perceived collaboration with nurses and with the physician involved in the handoff were not correlated with levels of IC. ...
Background: Frequent interruptions, inadequate privacy, and lack of collaboration are barriers to safe and efficient end-of-shift handoffs between emergency physicians. Varying levels of visibility to and from physicians can impact these outcomes. This study quantifies potential visual exposure of physicians in workstations with varying enclosure levels using isovist connectivity (IC) as a measure. Further, this study examines the association of IC with number of interruptions/hour, perceived collaboration, and privacy during handoffs. Methods: In-person observations were conducted during 60 handoffs to capture interruptions. Surveys were administered to the incoming and outgoing physicians to garner their perceptions of the extent of interruptions, collaboration, and privacy. Spatial analysis was conducted using DepthmapX. Results: Findings demonstrate significant differences in IC scores based on (a) physicians location within the workstation during; (b) handoff approach (individual or collaborative); (c) position during handoff (sitting or standing). Documented interruptions were highest in the high IC locations and lowest in the medium and low IC locations. Physicians in low IC locations perceived to have sufficient privacy to conduct handoffs. Limitations and conclusion: It should be noted that the three pods, each housing a physician workstation with different enclosure levels, varied in number of patient rooms, patient acuity, overall size, and the location of workstations. While contextual variables were considered to the extent possible, several other factors could have resulted in differences in number of interruptions and collaboration levels. This study provides design recommendations for handoff locations and a method to test emergency physician workstation designs prior to construction.
... The results of studies on the impact of spatial openness and transparency on communication and collaboration are mixed (see literature reviews by De Croon et al., 2005;Engelen et al., 2019); while some studies indicate increased communications , others suggest that spatial openness may be overstimulating and thus decrease concentration and productivity (Bernstein and Turban, 2018;Brennan et al., 2002). In other words, more openness and transparency do not always result in the intended positive outcomes. ...
Full-text available
This thesis explores the interrelations between the design characteristics of activity-based offices, users’ perceptions of them, and users’ sense of coherence. The goal is twofold: (i) contribute to conceptualizations of healthy activity-based offices and (ii) facilitate practical use of the sense of coherence theory for office designers. Most research into healthy offices has focused on harm-causing factors (pathogenic aspects) while overlooking the health-promoting design characteristics in activity-based offices (salutogenic aspects). This thesis is a response to the call for a paradigm shift and explores the particular design characteristics of activity-based offices that promote health, drawing on the salutogenic approach and sense of coherence theory. The thesis builds on a literature review and two mixed methods case studies on activity-based offices. Drawing on the sense of coherence framework, three types of design characteristics were identified: (i) those that promote a clear understanding of office environments, (ii) those that enhance users' access to relevant resources, and (iii) those that evoke meaning for users to cope with stressors. These characteristics and the perceptions of them are interrelated meaning that they can have multiple impacts on users’ sense of coherence. The findings also highlighted temporal changes in users’ perceptions, indicating that novelties of the new office wore off and the initial problems observed in the office environment worsened. Moreover, activity-based offices were not always perceived as intended because of suboptimal design solutions and contextual factors. In conclusion, there are no definitive answers to how to design healthy activity-based offices. Activity-based offices are complex environments and consist of many interacting aspects including the design characteristics, individuals’, and their work-related prerequisite as well as organization-related factors that influence users’ perceptions and their sense of coherence. The framework developed in this thesis may contribute to better-informed discussions about designing for sense of coherence. The thesis suggests that healthy activity-based offices should be viewed as a "moving project" that develops over time through experimentation and adaptation, with management’s involvement. Thus, a healthy activity-based office provides users resources and opportunities to codesign an environment that enables them (i) build meaningful social relationships, (ii) manage visual and acoustic distractions, (iii) read and understand workspaces, and (iv) receive support from management in their daily work.
... For example, some companies have transformed traditional office spaces such as separate offices or walled cubicles into open workspaces (i.e., no boundaries) with the expectation that it will spur more face-to-face interactions and increase collaboration and creativity. However, some research suggests that these open-space efforts have exactly the opposite of the intended effect, greatly reducing face-to-face interaction and ironically increasing email and instant messaging communications(Bernstein & Turban, 2018). ...
Loneliness is a complex set of aversive feelings that arises when people perceive that their belongingness needs are not being met. Usually, these feelings of loneliness are temporary because people successfully cope with their loneliness by connecting with others. However, for some people, their attempts to cope with loneliness are unsuccessful, and their loneliness becomes chronic, which can have severe consequences for their mental and physical health. Understanding the causes and consequences of loneliness is critical for developing interventions to reduce loneliness, a need made more urgent by the dramatic rise in reported loneliness over the last few decades. In this review, we provide a synthesis of the research on how people cope with loneliness through consumption situations and the extent to which these coping strategies are successful. We also provide a discussion of how the marketplace has responded to the rapidly increasing levels of chronic loneliness worldwide. We conclude with an agenda for future research to answer both basic and applied research questions regarding the causes, consequences, and underlying processes of loneliness.
... However, the actual effects of these new designs and technologies are limited. For example, the open office design that created a buzz in recent years has turned out to have unclear or negative impacts on office workers' productivity (Bernstein and Turban, 2018;Haynes, 2008;Brennan, Chugh and Kline, 2002). Agglomeration remains the main source of superior productivity in the built environment. ...
Full-text available
Concepts like 'creative city', 'world city', and 'eco city' arrive with loud celebration, but fade in just a few years. In recent years, 'smart city' has been a buzzword. As each fad emerges, urbanists debate its meaning and implications. However, why so many urban concepts circulate at all is rarely focused on. This study attempts to answer this question based on the Marxian view of the built environment as a fixed capital. We focus on the differences between the built environment and other types of fixed capital, and show how these differences render capital circulation in the built environment sector more fragile. We claim that such fragility cannot be fixed within the circuit of capital, so external intervention is necessary and deployment of catchy urban concepts is a resorted method of such intervention.
... The organisations hoped that this change would increase collaboration, but the users mostly experienced an increase in informal communication. In this regard, literature on open offices shows varying results; while some studies found that communication increased in spaces shared with more colleagues (e.g., de Been et al., 2015;Engelen et al., 2019), others found that face-to-face interaction decreased in favour of electronic interaction (Bernstein and Turban, 2018), or that the perception of communication worsened over time (Haapakangas et al., 2019). Gerdenitsch et al. (2018) indicate that the interaction across teams increased right after the implementation of a flexible office design and stabilised in the longer term. ...
Full-text available
Different office types provide the preconditions for distinct user experiences. However, research evidence on how users appraise flexible offices such as activity-based or combi offices is not as abundant as in the case of the more traditional open-plan and cell offices. Furthermore, the available literature shows discrepant results between flexible offices. The main difference between traditional and flexible offices is that the design of the latter is intended for users to switch between different shared spaces and workstations oriented to support different activities, needs and preferences. This office design may offer new opportunities and challenges for users as their experiences at work may be influenced by design qualities (or constellations of them) that are not present in traditional offices. I study the experiences of users with flexible offices because I want to understand the influences that the design qualities of office artefacts and spaces have on such experiences, as well as their design implications. In addition, I utilise the acquired knowledge to explore design opportunities for positive user experiences with flexible offices. In this regard, the research angle adopted builds on a UX theoretical background and a practical approach with multiple user studies in real office environments. The findings show that user experiences with flexible offices are influenced by interrelated design qualities of the spaces and artefacts in use, rather than isolated qualities. These (tangible and intangible) qualities define the nature of an artefact, a space, or constellations of them that users experience, for instance the qualities of an office chair vs. a meeting room. Experiences are subjective, but relate to both individual and collective experiences, for example using an ergonomic workstation vs. sharing such workstations. The findings also suggest that designing for user experiences with flexible offices is a highly complex endeavour, and that emphasis should be placed on designing for the experiences of pleasure, community, autonomy, purpose, and control over the environment. Utilising this knowledge to develop and test research prototypes allowed for a richer understanding of the experiential process and its relation to more systemic aspects such as the context of use or the temporality of experiences. Derived from these research activities and their findings, I present in this thesis the tentative SEEX (Stimuli-Evaluation-EXperiential outcome) model of how user experiences take place. This thesis contributes knowledge on theoretical and practical levels for academics and practitioners to continue studying office user experiences from a UX perspective, support informed decisions in the planning, operation, and evaluation of offices, and explore design opportunities for office environments.
... Modern technology offers a range of possibilities next to traditional video cameras for capturing motion, such as wearables with a built-in accelerometer and eye trackers (e.g., Pagnotta et al., 2020). Assessments of body movements, limb and head positioning, facial expressions, and eye movements can be used to assess movement complementarity (e.g., Kapcak et al., 2019), synchrony (e.g., Prochazkova et al., 2022), and social proximity between people (e.g., with sociometric batches; Bernstein & Turban, 2018). Other types of behavioral data, specifically data extracted from audio recordings, allow for analysis of communication patterns, such as measuring speech turn-taking and speech interruptions (Zhou et al., 2020), as well as the use of non-verbal cues and speech fillers (e.g., Cassell et al., 2007). ...
Wearable technology enables collecting continuous in situ data from multiple people in various modalities, which can enhance team research and support, as the dynamic coupling of signals between interacting individuals (i.e., team coordination dynamics) is believed to reflect underlying processes and states of team functioning and performance. We conducted a systematic review on existing literature to evaluate the prospective use of wearable technology in research and practice. Using the IMOI framework as an organizing tool, our review revealed considerable support linking team coordination dynamics in different modalities to team functioning and performance, but also explicated the field’s nascent status.
... The former review study has found that certain physical elements of the work environment can contribute to creativity, e.g., elements such as indoor plants and flowers, calm and inspiring colors, window view, and daylight ( Table 1, p. 14 [23]). However, openness can also have a negative impact, e.g., declined job satisfaction and motivation [25], and less focus, with negative impacts on collaboration and creativity [26,27]. ...
Full-text available
This explorative case study investigates health-promoting office design from an experience and meaning-making perspective in an activity-based flex-office (A-FO) in a headquarter building. This small case study (n = 11) builds using qualitative data (walk-through and focus group interviews). A reflexive thematic analysis (RTA) of the experience of design approach was performed on this from a health and sustainability perspective, including the physical, mental, and social dimensions of health defined by WHO. Results show a wide range in participants' experiences and meaning-making of the health-promoting office design of their office building. The control aspect plays a central role in participants' experiences, including factors such as surveillance and obeyance, related to status and power, in turn associated with experiences of pleasantness, symbolism , and inclusiveness. Three main themes are identified in participants´ experiences: (1) comfort -non-comfort, (2) outsider-insider, and (3) symbolism. The major finding of the study is the ambiguity among participants about the health-supportive office design of the office building per se and its various environments. There is a sense that it is chafing, due to dissonance between the intention of the office and the applied design.
Our research aims to study the phenomenon of Collective Intelligence (CI) in Instrumented Collective Learning Situations (ICLS). The theoretical framework used is linked to the factors of autonomy and learning strategies (Bandura, 2000; Morin, 1999; Ryan and Deci, 2020), to Rabardel's instrumental genesis (1995) and Engeström's activity theory (2015), as well as to the definitions of the contours of CI (Gréselle-Zaïbet, 2007; Lévy, 1994; Woolley et al., 2010). Our problematic revolves around the identification, support and assessment of the emergence of CI in ICLS. The research method, inspired by THEDRE (Mandran, 2017), proposes the iterative implementation of three studies to apprehend CI through mixed methods and Computer-Supported Collaborative Learning (CSCL). Our first study allows us to qualitatively examine the results of a computer-based learning setting for exploring psycho-cognitive variables into the production of mind maps. The second study experiments with a semi-controlled pedagogical setting for the observation of mutual aid and collaboration links in ICLS based on self-reported data and interaction logs. The third attempts to assess the emergence of a CI in ICLS by combining quantitative and qualitative data. The synthesis of our results leads us to propose some directions for discussion in order to continue the work on supporting the emergence of CI in a digital context.
Conference Paper
Full-text available
We propose WaddleWalls, a room-scale interactive partitioning system using a swarm of robotic partitions that allows occupants to interactively reconfigure workspace partitions to satisfy their privacy and interaction needs. The system can automatically arrange the partitions’ layout designed by the user on demand. The user specifies the target partition’s position, orientation, and height using the controller’s 3D manipulations. In this work, we discuss the design considerations of the interactive partition system and implement WaddleWalls’ proof-of-concept prototype assembled with off-the-shelf materials. We demonstrate the functionalities of WaddleWalls through several application scenarios in an open-planned office environment. We also conduct an initial user evaluation that compares WaddleWalls with conventional wheeled partitions, finding that WaddleWalls allows effective workspace partitioning and mitigates the physical and temporal efforts needed to fulfill ad hoc social and privacy requirements. Finally, we clarify the feasibility, potential, and future challenges of WaddleWalls through an interview with experts.
The contribution of the open-plan office to work and organisation has long been a matter of some debate. Aside from its economic advantages, it is argued that it provides an important opportunity for colleagues to share knowledge and help each other. It is recognised, however, that the presence and participation of others can undermine the ability of personnel to concentrate on individual tasks and subjects work to interruption. This paper seeks to show how these seemingly contradictory issues are matters participants themselves orient to on a daily basis. In particular, it explores the interactional practices in and through which participants address and, to some extent reconcile, these competing demands; initiating brief conversations while seeking to preserve the integrity of the ongoing tasks in which colleagues are engaged. This article focuses on participants progressively establishing momentary encounters that enable them to exchange information and resolve the inevitable difficulties colleagues face in organisations.
Full-text available
Significance Many human endeavors—from teams and organizations to crowds and democracies—rely on solving problems collectively. Prior research has shown that when people interact and influence each other while solving complex problems, the average problem-solving performance of the group increases, but the best solution of the group actually decreases in quality. We find that when such influence is intermittent it improves the average while maintaining a high maximum performance. We also show that storing solutions for quick recall is similar to constant social influence. Instead of supporting more transparency, the results imply that technologies and organizations should be redesigned to intermittently isolate people from each other’s work for best collective performance in solving complex problems.
Full-text available
The similarities between the structures built by social insects and by humans have led to a convergence of interests between biologists and architects. This new, de facto interdisciplinary community of scholars needs a common terminology and theoretical framework in which to ground its work. In this conceptually oriented review paper, we review the terms ‘information’, ‘space’ and ‘architecture’ to provide definitions that span biology and architecture. A framework is proposed on which interdisciplinary exchange may be better served, with the view that this will aid better cross-fertilization between disciplines, working in the areas of collective behaviour and analysis of the structures and edifices constructed by non-humans; and to facilitate how this area of study may better contribute to the field of architecture. We then use these definitions to discuss the informational content of constructions built by organisms and the influence these have on behaviour, and vice versa. We review how spatial constraints inform and influence interaction between an organism and its environment, and examine the reciprocity of space and information on construction and the behaviour of humans and social insects. This article is part of the theme issue ‘Interdisciplinary approaches for uncovering the impacts of architecture on collective behaviour’.
Full-text available
A social network represents interactions and knowledge that transcend the intelligence of any of its individual members. In this study, I examine the correlations between this network collective intelligence, spatial layout, and prestige or status outcomes at the individual and team levels in an organization. I propose that spatially influenced social cognition shapes which individuals become members of prestigious teams in organizations, and the prestige perception of teams by others in the organization. Prestige is a pathway to social rank, influence and upward mobility for individuals in organizations. For groups, perceived prestige of work teams is related to how team members identify with the group and with their collaborative behaviours. Prestige enhances a team's survivability and its access to resources. At the individual level, I ran two-stage Heckman sample selection models to examine the correlation between social network position and the number of prestigious projects a person is a member of, contingent on the association between physical space and social ties and networks. At the team level, I used linear regressions to examine the relationship among network structure, spatial proximity and the perceived prestige or innovativeness of a project team. In line with my hypotheses, for individuals there is a significant correlation between physical space and social networks, and contingent on that, between social network positions and the number of prestigious projects that a person is a member of. Also in accordance with my hypotheses, for teams there is a significant correlation between network structure and spatial proximity, and perceived prestige. While cross-sectional, the study findings illustrate the importance of considering the spatial domain in examinations of how network collective intelligence is related to organizational outcomes at the individual and team levels. This article is part of the theme issue ‘Interdisciplinary approaches for uncovering the impacts of architecture on collective behaviour’.
Observation is key to management scholarship and practice. Yet a holistic view of its role in management has been elusive, in part due to shifting terminology. The current popularity of the term “transparency” provides the occasion for a thorough review, which finds (a) a shift in the object of observation from organizational outcomes to the detailed individual activities within them; (b) a shift from people observing the technology to technology observing people; and (c) a split in the field, with managers viewing observation almost entirely from the observer’s perspective, leaving the perspective of the observed to the realm of scholarly methodology courses and philosophical debates on privacy. I suggest how the literature on transparency and related literatures might be improved with research designed in light of these trends.
This article presents a dynamic view of how institutions external to the state foster informal relationships that influence the voting preferences of elected officials. Institutions external to the state can activate different identities, potentially leading to outcomes at odds with predictions based on party identification and political ideology. This theoretical perspective is applied to a period of increasing partisanship and polarization—the birth of the second party system in America—when congressmen lived together in boardinghouses chosen independently of preexisting political preferences. Common residency is found to mute the effect of partisanship while reinforcing the North-South divide; this effect was strongest during the period of most acute polarization. This finding sheds light on the interaction between state and nonstate institutions and their differential impact on the political process. It also suggests a cause of current high levels of political polarization: the lack of time congressmen spend together informally.