E-Mail as a Source and Symbol of Stress

Abstract

The increasing volume of e-mail and other technologically enabled communications are widely regarded as a growing source of stress in people's lives. Yet research also suggests that new media afford people additional flexibility and control by enabling them to communicate from anywhere at any time. Using a combination of quantitative and qualitative data, this paper builds theory that unravels this apparent contradiction. As the literature would predict, we found that the more time people spent handling e-mail, the greater was their sense of being overloaded, and the more e-mail they processed, the greater their perceived ability to cope. Contrary to assumptions of prior studies, we found no evidence that time spent working mediates e-mail-related overload. Instead, e-mail's material properties entwined with social norms and interpretations in a way that led informants to single out e-mail as a cultural symbol of the overload they experience in their lives. Moreover, by serving as a symbol, e-mail distracted people from recognizing other sources of overload in their work lives. Our study deepens our understanding of the impact of communication technologies on people's lives and helps untangle those technologies' seemingly contradictory influences.

Full text

Organization Science
Vol. 22, No. 4, July–August 2011, pp. 887–906
issn 1047-7039 eissn 1526-5455 11 2204 0887 doi 10.1287/orsc.1100.0573
© 2011 INFORMS
E-mail as a Source and Symbol of Stress
Stephen R. Barley
Department of Management Science and Engineering, School of Engineering, Stanford University, Stanford, California 94305,
sbarley@stanford.edu
Debra E. Meyerson
School of Education, Stanford University, Stanford, California 94305, debram@stanford.edu
Stine Grodal
Department of Strategy and Innovation, School of Management, Boston University, Boston, Massachusetts 02215, grodal@bu.edu
The increasing volume of e-mail and other technologically enabled communications are widely regarded as a growing
source of stress in people’s lives. Yet research also suggests that new media afford people additional flexibility and
control by enabling them to communicate from anywhere at any time. Using a combination of quantitative and qualitative
data, this paper builds theory that unravels this apparent contradiction. As the literature would predict, we found that
the more time people spent handling e-mail, the greater was their sense of being overloaded, and the more e-mail they
processed, the greater their perceived ability to cope. Contrary to assumptions of prior studies, we found no evidence that
time spent working mediates e-mail-related overload. Instead, e-mail’s material properties entwined with social norms and
interpretations in a way that led informants to single out e-mail as a cultural symbol of the overload they experience in
their lives. Moreover, by serving as a symbol, e-mail distracted people from recognizing other sources of overload in their
work lives. Our study deepens our understanding of the impact of communication technologies on people’s lives and helps
untangle those technologies’ seemingly contradictory influences.
Key words: e-mail; stress; work–life; coping; communication technologies
History: Published online in Articles in Advance September 30, 2010.
Introduction
Concern about the increasing pace and volume of work
and the implications of both for the stress that peo-
ple face in their lives has been mounting for decades
(Schor 1993, Jacobs and Gerson 1998). Scholars have
identified numerous reasons for the growing tyranny of
work, including the rise of dual-career families (Nippert-
Eng 1996, Moen 2003, Jacobs and Gerson 2004), the
need for many Americans to hold multiple jobs to make
ends meet (Schor 1993), the compression of more activ-
ities into a given unit of time (Robinson and God-
bey 1997) and, most important for the purposes of this
paper, the advent of new communication technologies
(Sproull 2000, Kaufman-Scarborough 2006, Golden and
Geisler 2007).
Although e-mail, cell phones, and other communica-
tion technologies ostensibly offer greater flexibility and
control over work (Hill et al. 2001, Valcour and Hunter
2005), their material features and properties are also
frequently said to contribute to overload (e.g., Richtel
2003, Alvarez 2005, Stross 2008, Chesley 2005, Boswell
and Olson-Buchanan 2007). Commentators claim that
e-mail creates extra work, in part, because its asyn-
chrony allows people to send and receive work-related
messages at anytime. Similarly, the ubiquitous access
afforded by the mobility of wireless devices allows work
to invade times and places that were previously safe
from the workplace’s intrusion (Murray and Rostis 2007,
Middleton and Cukier 2006). Those who offer advice
on how to reduce communication overload also typically
advise people to regain control by deploying a technol-
ogy’s material features more effectively, for example,
using e-mail filters or the mute button on a cell phone
(Jarvenpaa and Lang 2005, Dabbish and Kraut 2006).
Although it makes little sense to dispute that the ubiq-
uity and features of communication technologies can
lead to patterns of use that produce feelings of over-
load, the relationship between stress and the use of those
technologies is likely to be more complex than purely
materialistic accounts suggest. Prior research on technol-
ogy shows that social processes are usually as important
as—if not more important than—a technology’s material
properties for shaping its use and consequences (e.g.,
Barley 1986; Fulk et al. 1987; Rice and Aydin 1991;
Orlikowski 1992, 2000). With few exceptions (Dabbish
et al. 2005, Renaud et al. 2006, Mazmanian et al. 2006),
however, the role of the social forces in shaping peo-
ple’s experience of their use of communication tech-
nologies in their lives has received scant attention. Even
less well understood are how norms and interpretations
might entwine with the material properties of these tech-
nologies to shape when and why people experience
them as sources of stress. Sociomaterial accounts of a
technology’s use—accounts that weave together rather
887

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
888 Organization Science 22(4), pp. 887–906, © 2011 INFORMS
than segregate social, symbolic, and material realities—
would advance our understanding of why people use and
experience technologies as they do (Orlikowski 2007,
Leonardi and Barley 2008).
Based on a study of the communicative activities of
members of a high-technology firm, this paper examines
people’s emotional experience of the communication
technologies they use. Specifically, we ask if—and if
so, how—e-mail and other communications technologies
contribute to the stress people experience. In answering
this question, we provide a more theoretically nuanced
understanding of the relationship between communica-
tion technologies and stress by showing that one cannot
understand how media induce stress without considering
how properties of technologies become entangled with
social norms, interpretations, and the flow of daily work.
We begin by reviewing previous research on commu-
nication technologies and the experience of stress and
overload. We focus primarily on e-mail because it is the
communication technology that has been most clearly
associated with stress. Using a mixed-methods research
approach, we find that because of e-mail’s features and
the norms and meanings associated with its use, our
informants blamed e-mail for the stress they experi-
enced regardless of the amount of time they worked and
regardless of the fact that other communication activities
also exacerbated their workload and the stress they felt.
In short, e-mail was not just a source but also a symbol
of overload.
Prior Research on Communication
Technologies and Stress
Scholars have proposed two paths by which the use of
communication technologies generates feelings of being
overloaded and overwhelmed. One school of thought,
primarily situated in the work–life literature, holds that
e-mail, pagers, cell phones, and mobile devices occa-
sion stress because they make it easier for work to spill
into times and places formerly reserved for family and
self (Murray and Rostis 2007). For example, Chesley
(2005) and Chesley et al. (2003) found that respondents
who reported more frequent use of cell phones, pagers,
and computer-mediated communications reported more
negative “work-to-family spillover.” Based on self-report
data, researchers (Duxbury et al. 2006, Towers et al.
2006) found that the more people relied on communi-
cation technologies, the more they worked from home
and locations other than their office. Furthermore, the
more they used these technologies, the more likely they
were to report feeling burned out. These studies rest on
variants of the argument that e-mail and other communi-
cation technologies produce stress by enabling work to
spill into other domains of life, thereby extending work
hours and making it more difficult to disengage from
work and fulfill family obligations (Major et al. 2002,
Boswell and Olson-Buchanan 2007).
A second school of thought, primarily focused on
“e-mail overload,” argues that e-mail and other commu-
nication technologies create feelings of stress primar-
ily because they increase the total amount of work that
people must handle and, hence, the time they spend
working. Research has revealed four ways that e-mail
creates additional work. First, because e-mail is easier
to send than written letters and memos, people report
that with e-mail, they handle more communications and
spend more time sorting and filing messages (Dawley
and Anthony 2003, Bälter 2000, Bellotti et al. 2005). In
the past, administrative aides assumed responsibility for
filing and storing most documents. E-mail pushed these
tasks onto everyone, regardless of their managerial or
professional status.
Second, e-mail creates additional work by making it
relatively costless for people to make requests that divert
attention away from the task at hand (Manger et al. 2003,
Belotti et al. 2005, Thomas et al. 2006). Dabbish et al.
(2005, p. 696) estimate that a third of all messages con-
tain “request(s) for action” that “cause people to shift
gears and to add new tasks to their current stack.”
Third, e-mail interrupts concentration. Jackson et al.
(1999, 2001b, 2003) documented the incremental time
required to resume work following an e-mail session.
In the organization that they studied, employees had set
their e-mail to signal the arrival of new messages every
five minutes. In addition to the time these employees
spent reading and managing messages, they required,
on average, 64 seconds to resume work. Jackson et al.
(2001a, p. 55) note that this rate of checking e-mail
caused 96 interruptions in an eight-hour day and, hence,
roughly an hour and a half of recovery time per day.
González and Mark (2004, 2005) suggest that the
situation is potentially worse than Jackson et al. esti-
mate. González and Mark shadowed 36 employees in
two organizations over a period of several days using
time logs in which they recorded the minute-by-minute
sequence of activities in which their informants engaged.
Based on these data, the researchers discovered that
although people might resume work shortly after an
interruption, they often did not return directly to the
same task. It took informants, on average, 25 minutes
to return to the original task, during which time they
engaged in an average of 2.3 other activities (Mark et al.
2005). Thus, e-mail creates distractions that extend con-
siderably the time it takes to complete a given task.
Finally, the use of e-mail to perform tasks for which
the programs were not designed requires additional time
to compensate for the programs’ inadequacies (Bellotti
et al. 2003, Renaud et al. 2006). Researchers have found
that people use e-mail not only to communicate, but also
to coordinate: for example, keeping track of tasks, dis-
tributing documents, managing contact lists, and orga-
nizing information (Whittaker and Sidner 1996; Belotti
et al. 2003, 2005). Consider how much time people

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
Organization Science 22(4), pp. 887–906, © 2011 INFORMS 889
spend reconciling responses when using e-mail to sched-
ule meetings between multiple parties or when locat-
ing documents from e-mail archives. When assigned to
geographically distributed teams (Herbsleb and Mockus
2003, Herbsleb et al. 2000, Hinds and Bailey 2003) and
to multiple teams (O’Leary et al. 2008, Mortensen et al.
2007), people are even more likely to use e-mail to
coordinate.
The common denominator that cuts across the two
foregoing bodies of research is that e-mail and other
communication technologies induce stress by extending
the time that people work, but the explanations differ.
According to studies situated in the work–life litera-
ture, the use of communication media extends time by
allowing people to continue working after leaving the
workplace. According to technology studies, communi-
cation technologies increase the total amount of work
that people confront. Thus, for different reasons, these
two streams of research arrive at the same conclusion:
the more people depend on and use e-mail and other
electronic media, the longer they work, and the longer
they work, the more likely they are to feel overloaded
and overwhelmed.
These bodies of research suffer from three short-
comings. The first is methodological. With one excep-
tion (Thomée et al. 2007), no study of e-mail over-
load directly measures stress. Instead, researchers docu-
ment the amount of time that processing e-mail requires
and then infer stress by appealing to a large body of
research that establishes that working long hours cor-
relates with stress (e.g., Cooper and Marshall 1976,
Sparks et al. 1997, Moen and Yu 2000). Thomée et al.
(2007) directly assessed e-mail’s contribution to stress
and found no relationship between time spent on e-mail
and stress, although they did find a correlation between
time spent doing e-mail and depression among women
(but not men).
Although students of work and family have measured
stress, their measures of technology use are ambigu-
ous or confounded. Duxbury et al. (2006) and Towers
et al. (2006) assessed the link between technology use
and stress by asking their respondents to indicate the
degree to which each of several technologies caused
stress and affected work–life balance. Hence, their mea-
sures of stress were not conceptually independent of
technology use. Chesley (Chesley et al. 2003, Chesley
2005) relied on well-established indicators of stress but
combined using e-mail with other uses of the Internet to
create a measure of computer use. To the degree that her
respondents used e-mail for work and the Internet for
recreation, Chesley’s finding of no association between
computer use and overload may have resulted from these
activities canceling each other out. Because of these
methodological problems, the existing literature on com-
munication media and stress provides weak grounds for
strong inference.
Second, most studies of communications media and
stress have not grappled with the possibility that people
use electronic media to gain control over their work and
reduce overload. A number of scholars have suggested
that e-mail might actually reduce stress by allowing peo-
ple to avoid meetings, telephone tag, and face-to-face
interactions that waste time (Berghel 1997, El-Shinnawy
and Markus 1998, Renaud et al. 2006). Phillips and
Reddie (2007) argue that e-mail might reduce stress
by providing a sanctioned way of procrastinating. Even
studies that report a link between stress and e-mail
contain findings that suggest that the opposite may
hold under certain circumstances. For example, Chesley
et al. (2003, p. 221) found positive associations between
the use of electronic media and work–life balance for
women.
Conflicting evidence on the deleterious consequences
of electronic communication has also been a recurrent
theme in recent research on BlackBerries and other
mobile e-mail devices. For example, the profession-
als that Mazmanian et al. (2006) encountered in their
study of a private equity firm claimed that BlackBerries
enhanced their sense of being up to date, in touch, and in
control of work. Yet their “experience of almost constant
connectivity increase[d] their electronic dependence, and
generate[d] compulsive routines of chronic checking,
escalation of commitment, reduced time for reflection
and increased stress in the longer term” (Mazmanian
et al. 2006, p. 9). Jarvenpaa and Lang (2005), Middleton
and Cukier (2006), and Murray and Rostis (2007) report
comparable results in their studies of BlackBerry users.
Gergen (2002) and Green (2002) show a similar paradox
among heavy cell phone users.
A third shortcoming of most research on communi-
cation technologies and stress is inadequate attention
to social and cultural dynamics. Researchers usually
attribute stress and overload to such material properties
as e-mail’s asynchrony, the cell phone’s mobility, time
spent communicating, and the volume of messages sent
and received. Although studies occasionally allude to the
influence of norms and interpretations, with a few excep-
tions, the thesis is rarely developed (Mazmanian et al.
2006). For example, in attempting to reconcile why some
people reported that e-mail and BlackBerries increased
stress and others claimed they reduced stress, Duxbury
et al. (2006, p. 320) speculate, “It is not the technol-
ogy per se that contributes to or alleviates stress but
how the technology is used” that matters. When Dabbish
et al. (2005) found, contrary to their expectations, that
individual differences explained little variation in the
likelihood that people would respond to e-mail, they sug-
gested that response patterns might reflect social norms.
Finally, Dabbish and Kraut (2006) noted that how peo-
ple interpret the volume of e-mail they receive may be
more important for explaining e-mail overload than the
actual quantity of messages. In other words, the notion

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
890 Organization Science 22(4), pp. 887–906, © 2011 INFORMS
that social dynamics might at least partially account for
why and when people find electronic media stressful has
usually been proposed as an afterthought.
The tendency for research on communication tech-
nologies to blame stress on the material features of
the technologies stands in sharp contrast to recent
research on how technologies shape other aspects of
work (Barley 1986, Fulk et al. 1987, Orlikowski 1992,
Poole and DeSanctis 1990, Fulk 1993, Markus 1994).
Studies of technology and organizing have shown that
social influence (Fulk 1993, Rice and Aydin 1991),
the transference of interpretations from one domain to
another (Barley 1988, Orlikowski and Gash 1994), sit-
uated improvisations (Orlikowski 1996), and the nego-
tiation and renegotiation of roles (Zuboff 1988, Barley
1990) typically shape a technology’s effects. Orlikowski
(2007), Orlikowski and Scott (2008), and Leonardi and
Barley (2008) have recently argued that contemporary
research on technology and work has become so ori-
ented to social and interpretive dynamics that the role of
technology’s constraints and affordances has faded from
view. To regain balance, both Orlikowski and Barley
argue for a more integrated or sociomaterial approach.1
As Orlikowski (2007, p. 1437) put it,
[The sociomaterial] view asserts that materiality is
integral to organizing, positing that the social and the
material are constitutively entangled in everyday life.
A position of constitutive entanglement does not privilege
either humans or technology000Instead, the social and the
material are inextricably related—there is no social that is
not also material, and no material that is not also social.
Orlikowski (2007) illustrated the sociomaterial view by
drawing on the research of Mazmanian et al. (2006) on
BlackBerry users. She attributed the BlackBerry users’
ubiquitous and often obsessive e-mail checking to the
interweaving of the device’s material capabilities (par-
ticularity its mobility and software configurations that
continually push e-mail to the user) with social norms,
interpretations, and individual proclivities that led users
to carry the devices everywhere, keep them constantly
activated, and respond to e-mail regardless of place
or time.
Sociomaterial analysis urges researchers to pay atten-
tion to the concurrency and interweaving of three sets of
pressures: the technology’s material attributes that con-
strain or afford particular behaviors, the social norms
and cultural interpretations that shape how people use a
technology, and what we call “quasi-material” parame-
ters. By quasi-material we mean those aspects of a con-
text that are separate from a technology’s features and
that may ultimately be social, but which users do not see
as social and which they treat as objective constraints.
Quasi-material parameters include the volume of mes-
sages the people receive, the times at which messages
typically arrive, and the distribution of correspondents
across time zones. Consider a team distributed between
North America and India. When North Americans are
awake, Indians are asleep. Most messages from India
arrive in the United States in the middle of the night.
North Americans treat the timing of the Indians’ mes-
sages as hard reality, even though in theory a firm could
mandate that one group must work at night to force tem-
poral entrainment. In other words, the effects of time
zones are experienced as a natural or material fact over
which workers have no control, even though time zones
are partially the product of social conventions that could
be contravened.
This study examines whether and how communica-
tion technologies evoke feelings of stress among users
from a sociomaterial vantage point. We use qualitative
data to document how members of an organization inter-
preted their experience of new media, and we triangulate
those interpretations against quantitative data to high-
light the social, material, and quasi-material aspects of
the phenomenon. The quantitative data also offer us a
direct measure of stress and allow us to point to sources
of stress that were not emphasized in the informants’
accounts because they were not figural in their work
culture.
We discovered that our informants blamed comm-
unications-related stress exclusively on the volume of
e-mail they handled and the extra time that e-mail added
to their workday. We also found that processing e-mail
bolstered their sense of being able to cope with their
work. Yet our quantitative data show that other media
also contributed to feelings of overload and that media-
related stress appeared to be independent of workload.
Unpacking the sociomaterial reality of e-mail allowed us
to make sense of these contradictions and to build theory
about the relationship between communication technol-
ogy and people’s experience of stress in daily life.
Methods
Site and Sample
We collected the data for this study between October
2001 and March 2002 as part of a larger program
of research on how people use communication devices
to construct their availability to others and manage
their work lives. Seventy-nine employees of a com-
pany known internationally for its workstations and
servers participated. We drew participants from three
departments because they represented different tasks and
occupations.
The first group consisted of top-level escalation engi-
neers (39 individuals) responsible for resolving customer
problems with hardware and software whose solutions
had evaded first-, second-, and third-level support staff.
The problems were of the type that could bring a cus-
tomer’s business to a grinding halt. Members of the esca-
lation group were available to customers 24 hours a day,
7 days a week. Technical writers (13 individuals), who

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
Organization Science 22(4), pp. 887–906, © 2011 INFORMS 891
penned documentation for the company’s equipment
and software, were the second group. As has become
increasingly true for technology companies, the writers
distributed much of the material they authored via web-
sites. Marketing personnel (27 individuals) responsible
for developing markets for the company’s products made
up the third group. Much of this group’s effort involved
planning and developing the company’s Web presence
and designing online interfaces for existing and potential
customers. Our sample included both managers (29%)
and nonmanagerial employees (71%).
Members of all three groups were concurrently
involved in a number of project teams that cut across
functional areas and imposed their own demands and
deadlines. To fulfill expectations, members of all groups
often brought work home. Although each group’s tasks
differed, all were composed of knowledge workers who
were regularly called on to innovate or unravel novel
problems. The complexity of each type of work cre-
ated interdependencies that required constant coordina-
tion and communication.
The average respondent worked 9.4 hours each day,
two-thirds of which (6.4 hours) were spent commu-
nicating. Respondents spent 34% of their communica-
tion time in a combination of meetings and encounters.
E-mail accounted for another 31%. Phone calls and tele-
conferences accounted for 16% and 14%, respectively,
with the remaining 5% of their communication time
allocated across the use of pagers, voicemail, video-
conferences, and instant messaging technologies. E-mail
use was ubiquitous. Escalation engineers used e-mail
to communicate with each other and with customers to
diagnose and solve technical problems. Technical writ-
ers used e-mail to talk with the technical specialists
whose knowledge they translated into documentation
and specifications. The marketers used e-mail to coordi-
nate strategy, interact with customers, and develop mar-
keting materials. Managers also coordinated and moni-
tored their groups’ activities primarily through e-mail.
The individuals we studied had worked for the com-
pany, on average, six years. The average age was 40.
Forty-nine percent were women. Sixty-four percent were
married. Nearly a third had at least one child who was 18
years old or younger. Because the company actively pro-
moted telecommuting, a third of the respondents worked
from home one or more days a week. Everyone had
access to e-mail, land phones, and voicemail, and all had
computers at home. All used the same e-mail client that
offered a number of options for managing e-mail, includ-
ing the ability to define filters and to decide whether
the recipient would be notified of new messages as they
arrived. Eighty-six percent of participants carried cell
phones, 68% used laptops, and 39% (all of whom were
escalation engineers) carried a pager.2
Data Sources
Because our goal was to examine the relationship
between stress and the use of communication technolo-
gies from a vantage point that differs from previous
research, and because we sought to deepen our under-
standing of the mechanisms that might underlie that rela-
tionship, we collected both quantitative and qualitative
data (see Edmondson and McMannus 2007 on combin-
ing methods to build intermediate theory). Combining
both types of data is valuable because it not only allows
one to confirm common findings across methods (Jick
1979), but just as importantly, one can identify dynamics
obscured by one data source or another (Bernard et al.
1985). We collected our quantitative data through com-
munication logs and surveys and qualitative data through
interviews.3
Communication Logs and Surveys. We gave respon-
dents a log book in which we requested that they
record all communication activities that occurred dur-
ing the course of two workdays. Additionally, we asked
respondents to log all work-related communications that
occurred in the morning before they came to work, in
the evening after they left work, and on one weekend
day. Half of the respondents logged communications for
a consecutive Thursday, Friday, and Saturday, and the
other half logged communications for a consecutive Sun-
day, Monday, and Tuesday.
Before logging began, we assembled the members of
each group to instruct them on procedures for complet-
ing their logs. As part of the training, we had them
log communications that had occurred on the day of
training to reveal questions that might occur while they
completed their logs. We instructed respondents to log
events at the time they took place or shortly thereafter,
stressing the importance of being exhaustive and accu-
rate. During the training, respondents completed a brief
questionnaire that asked demographic questions as well
as questions about which communication devices they
used. The questionnaire also included items that assessed
the level of stress respondents were experiencing and
their capacity to cope with the demands of work.
Respondents logged the full gamut of communica-
tion events: face-to-face encounters including meetings,
e-mail sessions (defined as any time a respondent looked
at or sent an e-mail), voicemail sessions (defined as
any time a respondent listened to or recorded a voice-
mail message), calls made or received on land and
cell phones, teleconferences, videoconferences, pages,
and instant messages.4Respondents recorded the time
each event began, the time it ended, the location where
the event occurred, whether the event was planned or
spontaneous, and the number of other people involved.
For e-mail and voicemail sessions, respondents logged
the number of messages they received, the number to
which they responded or forwarded, and the number

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
892 Organization Science 22(4), pp. 887–906, © 2011 INFORMS
they initiated. At the end of each day, respondents
recorded whether the day was atypical in any way
that affected their communications. Almost no respon-
dents acknowledged an atypical day. For each workday,
respondents recorded the times they arrived at and left
the workplace.5After logging for three days, respon-
dents sent their logs directly to the researchers using a
prepaid envelope to ensure confidentiality of their data.
Interviews. After receiving completed logs, we inter-
viewed 40 respondents in depth. We selected informants
to ensure that we captured the perspectives of all occu-
pational groups, men and women, managers and non-
managers, people with and without children, and people
who worked from home as well as those who worked
in traditional offices. We designed an interview protocol
of open-ended questions to gain a deeper understanding
of how informants made sense of the media they used,
how they thought about communicating, and how work-
related communications fit into and affected their lives
on and off the job. Prior to interviewing each informant,
we reviewed his or her communication log. With these
data in front of us and before asking the questions on
the protocol, we asked informants to explain patterns in
their logs and to clarify ambiguous entries.
The interview protocol did not include explicit ques-
tions about stress or overload, because we did not want
to bias informants’ descriptions of their experiences with
technologies. Thus, when informants talked about stress
or made connections between their use of a technol-
ogy and stress, they did so in the context of responses
to broad questions about boundaries between work and
home or general queries about technology use. For
example, questions like “Tell me about how you think
about using e-mail” often triggered emotional responses
that included talk of overload.
Interviews were conducted via telephone and were
tape recorded with the informants’ consent. No infor-
mant refused to grant permission to record. We asked
each informant the same questions in the same order.
We transcribed the tapes and entered the transcriptions
into Atlas.ti, a software package designed for qualitative
analysis of text and other documents.
Measures
Dependent Variables. Because previous research im-
plies that e-mail exacerbates stress by increasing the
amount of time that people work, we explored three
dependent variables: time worked, overload, and coping.
We measured the number of hours worked by summing
two values. The first was the length of the respondent’s
workdays, defined as the number of hours that elapsed
between the time the respondent reported starting and
ending work on each of the two workdays logged. The
second was the number of hours that the respondent
communicated about work outside the boundaries of the
workday, which we measured as the sum of the dura-
tion of all work-related communications that occurred
(1) in the morning before the respondent reported start-
ing work, (2) in the evening after the respondent reported
ending work, and (3) on the weekend day that the
respondent logged. Note that our measure is likely to be
conservative, because we do not have measures of how
many hours respondents spent before work, after work,
or on the weekends doing work-related tasks that did not
involve communication.
Stress researchers have long conceptualized stress
as multidimensional. General overload and a perceived
sense of mastery/coping are two of the most important
concepts that stress researchers routinely assess. General
overload is usually defined as the perception of being
emotionally overwhelmed by life’s events and demands.
Coping/mastery typically refers to having the psycholog-
ical resources and strategies to withstand or overcome
stressors (Lazarus and Folkman 1984). Researchers have
developed a number of scales to assess both concepts.
We measured overload using the emotional exhaus-
tion subscale of Maslach and Jackson’s (1981) well-
validated burnout inventory. We measured coping with
items from the coping/mastery scale employed by the
1992 National Study of the Changing Workforce (Fam-
ilies and Work Institute 1992). Because overload is a
measure of stress and coping is a measure of being able
to handle stress, the two should be negatively corre-
lated. Together, the two scales allowed us to capture the
contradictory experiences of doing e-mail—namely, that
doing e-mail could simultaneously exacerbate overload
while making people feel as if they were coping more
effectively.
The first column of Table 1 displays the items that
comprised both scales. These items appeared on the
Table 1 Factor Analysis of Overload and Coping Items
Scales and items Factor 1 Factor 2
Overload
Felt emotionally drained from your work 0070878 −0033125
Felt used up at the end of the workday 0076844 −0029197
Tired when you get up in 0071556 −0016867
the morning and have to face
another day on the job
Felt burned out or stressed from 0075864 −0041161
your work
Frustrated by your job 0075191 −0018573
Coping
Felt confident about your ability −0011606 0058418
to handle your personal problems
Found that you could not −0029776 0062184
cope with all the things you
had to do (reverse scored)
Felt difficulties were piling −003119 0075761
up so high that you could not
overcome them (reverse scored)
Variance explained 2094 1073

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
Organization Science 22(4), pp. 887–906, © 2011 INFORMS 893
questionnaire, which instructed the respondents to indi-
cate how frequently they had experienced “each of the
following feelings over the last three months” using a
five-point Likert scale whose values ranged from “never”
to “sometimes” to “very often.” To ensure that the two
scales were coherent and distinct, we submitted the items
from both scales to principle factor analysis using a vari-
max rotation. Columns 2 and 3 of Table 1 contain the
loadings of the items on the two factors that emerged
from the analysis. Overload items loaded strongly on the
first factor, whereas coping items loaded on the second,
indicating that the two concepts were empirically dis-
tinct. Equally important, items that load positively on
one factor also load negatively on the other: people who
were better able to cope reported less overload, and vice
versa. Factor scores based on the factor analysis served
as measures of both overload and coping.
Independent Variables. Because the respondents indi-
cated which medium they used for each communica-
tion event and the times at which the event began and
ended, we could construct for each medium two indica-
tors of its use: (1) the number of events that occurred via
that medium during the two workdays that the respon-
dents logged and (2) the total time that respondents spent
using that medium during those days. All number and
time variables were calculated for communication events
that occurred during the confines of the respondent’s
workday because we were interested in how communi-
cation during the workday affected the total time that
people reported working, the degree to which they expe-
rienced overload, and the extent to which they felt they
could cope.
We also reasoned that different types of face-to-
face events would make different demands on people’s
time and might, therefore, contribute differentially to
the experience of stress. Accordingly, we divided face-
to-face events into two types based on whether the
respondent reported that the event had been planned.
Doing this enabled us to distinguish between events that
had different cultural meanings. Meetings were planned,
and encounters were unplanned, face-to-face events. For
each, we calculated the number of events in which the
respondent had participated and the time that he or she
spent participating.
The number of e-mails handled was simply the sum
of the number of messages that a respondent reported
receiving and sending during the course of the work-
day; the latter included not only messages initiated but
also responses to e-mails received. Time spent handling
e-mail was the total number of hours consumed by read-
ing and writing e-mail during the workday. The number
of phone calls handled was the sum of calls made and
received on land and cell phones during the course of
the workday.6Time on phone recorded the total num-
ber of hours that respondents reported talking on phones
during the workday.
Although employees used phones to participate in
teleconferences, we asked respondents to distinguish
between teleconferences and other types of phone calls
when logging their activities. Teleconferences are equiv-
alent to meetings except that the respondent participated
by phone. We coded as teleconferences those events
that the respondents labeled as teleconferences as well
as any cell phone or land phone call that the respon-
dent recorded as involving more than one other partici-
pant. The number of teleconferences was the sum of all
events coded as a teleconference in which the respondent
participated during the workday. Time spent teleconfer-
encing recorded the total number of hours that these
teleconferences consumed.
Demographic Variables. Researchers have long known
that managers use telephones more frequently than non-
managerial employees and that they also have a prefer-
ence for face-to-face communications (Mintzberg 1975,
Kotter 1982). Some studies have also found that man-
agers report more stress than employees who have no
managerial responsibilities, but others have found the
reverse or no effect (Maslach et al. 2001). We indicate
being a manager with a dichotomous variable.
Prior research has shown that gender, occupation, mar-
ital status, and having children influence how long peo-
ple work, the stress that they report, and how they
experience and handle stress (Bogg and Cooper 1994,
Cooper et al. 2001, Jacobs and Gerson 2004, Cooper
et al. 2001, Maslach et al. 2001). Accordingly, we con-
structed a number of variables that captured these demo-
graphic effects. Contrary to our expectations, none of
these demographic variables, including gender and occu-
pation, were significantly related to time spent work-
ing, overload, or coping after controlling for media use
among our respondents. Thus, we do not describe these
analyses or report results using these variables here.7
Table 2 reports the means, standard deviations, and
correlations for all variables. The correlation matrix indi-
cates no problems with multicolinearity outside of the
correlations that one would expect by definition: correla-
tions between the number of communication events for
each medium and the time spent using that medium.
Data Analysis
Quantitative Data. We used hierarchical, ordinary
least squares regressions to assess whether being a man-
ager, the number of communication events employing
various media, and the time spent using various media
influenced the length of respondents’ workdays. Because
the number of events using a medium and the time spent
using that medium are highly correlated, we conducted
separate regressions for each measure of medium use.
We performed similar analyses for overload and for cop-
ing, except in these regressions we included the length of
the respondents’ workdays as an independent variable.

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
894 Organization Science 22(4), pp. 887–906, © 2011 INFORMS
Table 2 Means, Standard Deviations, and Correlations Among All Variables
Variable NMeans Std 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1 Total hours worked 74 1909 3032 1000
2 Overload 72 0006 0087 0015 1000
3 Coping 72 −0002 0084 0008 −0016 1000
4 No. of encounters 74 7008 6053 −0002 0021 0020 1000
5 No. of meetings 74 2031 2021 0024∗∗ 0020 0000 0017 1000
6 No. of e-mails handled 74 185 151 0018 0019 0029∗∗∗ 0039 −0012 1000
7 No. of phone calls 74 9039 6044 0036∗∗∗ 0010 0010 0016 0015 0009 1000
8 No. of teleconferences 74 1066 2003 0035∗∗∗ 0016 −0006 −0003 0019 −0005 0029∗∗ 1000
9 Time in encounters 74 1036 1028 0012 0014 0016 0045∗∗∗ 0023∗∗ 0023∗∗ −0011 0001 1000
10 Time in meetings 74 2041 2069 0023∗∗ 0026∗∗ 0008 0011 0086∗∗∗ −0010 0007 0014 0029∗∗∗ 1000
11 Time doing e-mail 74 400 2063 0028∗∗∗ 0037∗∗∗ −0006 0009 −0014 0044∗∗∗ 0031∗∗∗ 0015 0004 −0013 1000
12 Time on phone 75 1090 1071 0042∗∗∗ 0017 0000 −0022 0011 0000 0052∗∗∗ 0035∗∗∗ −0019 0007 0038∗∗∗ 1000
13 Time in teleconferences 74 1065 2029 0048∗∗∗ 0016 −0007 −0011 0026∗∗ −0006 0016 0088∗∗∗ 0005 0028∗∗ 0014 0043∗∗∗ 1000
14 Manager 75 0029 0016 0028∗∗ −0032∗∗∗ 0012 0029∗∗∗ 0008 0022∗∗ 0015 0020 0032∗∗∗ 0021 0021 0018 1000
Note. Total hours worked and all number and time variables are calculated over the two workdays logged.
∗∗p≤0005; ∗∗∗ p≤0001.
Qualitative Data. We analyzed our interview data in
multiple phases, following an inductive process that
involved teleporting back and forth among the qualita-
tive data, the quantitative data, and emerging categories
and relationships (Strauss and Corbin 1990). The stan-
dard approach to analyzing textual data is to read and
code the entire corpus of text numerous times (Lofland
and Lofland 1984). Initial readings serve primarily to
familiarize researchers with the substance of the data and
to suggest potential themes. During subsequent readings,
analysts develop increasingly refined coding categories.
Typically, these categories are nested to form hierarchi-
cally structured mappings of substantive domains. We
describe each phase of data analysis, acknowledging that
the analytic processes that characterize each phase are
rarely confined to that phase. Nevertheless, explicating
the broad phases of the analysis can speak to the rigor
of the method and the integrity of the inferences drawn.
Phase 1: The authors and an additional research assis-
tant read through all transcripts to develop a general
sense of the data. Each researcher created his or her own
list of preliminary codes. We then met to compare anal-
yses to determine which codes to use during the second
phase of analysis. The initial codes mapped the general
topics in our protocol. For example, we had one code
that flagged passages for each communication medium
that the respondents logged and codes for general topics
such as “home life,” “accessibility,” and “children.” At
this point, we had no code for stress or overload.
Phase 2: During the second phase of the analy-
sis, we developed second- and third-level codes that
captured the various ways informants used and talked
about e-mail, cell phones, land phones, teleconferences,
face-to-face interactions, pagers, instant messaging, and
voicemail. Second-level codes flagged whether passages
referred to norms, emotions, evaluations, practices, or
reasons for using for each technology. Third-level codes
captured distinctions within these categories. For exam-
ple, one set of third-level codes identified the various
norms that informants invoked to explain how peo-
ple used each medium in their organization. A second
set of third-level codes inventoried the emotions infor-
mants associated with using each medium. A third fam-
ily of third-level codes pointed to informants’ positive
and negative evaluations of the medium’s utility. The
fourth set of third-level codes referred to communica-
tion practices or how informants used the media in their
lives. Practices tended to be specific behavioral tac-
tics or strategies. A fifth set of codes highlighted rea-
sons for using or not using one medium rather than
another. For e-mail and phones, we further distinguished
between reasons for using the medium outside the work-
day, specifically in the morning, in the evening, and on
weekends. Table 3 illustrates this level of coding for
e-mail.

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
Organization Science 22(4), pp. 887–906, © 2011 INFORMS 895
Table 3 Examples of Codes Applied to Passages About E-mail in Interview Transcripts
Technology Second-level code Third-level code Example
E-mail Norm People should be
responsive
I can get a response in an hour, I think that’s fair000I won’t resend something
unless it’s been over a day. The other thing is, it depends on the content
of the e-mail or whatever I need. If it’s something that’s not urgent, then
what do I care whether it’s a day or two days or one hour? But if the
server goes down and I send them an e-mail or I call them and I get a
voice message and I don’t hear back from them for half an hour, then I’ll
get more stressed. Because that, to me, is something important.
(Marketing, website manager)
E-mail Norm OK to e-mail at night It’s pretty rare [to get work-related calls at home]. People don’t make a habit
of calling me at home. It tends to be more e-mail. Because it’s
asynchronous, right? You don’t have to bother other people at home.
(Marketing, business planner)
E-mail Emotion Guilt for not
responding
There have been times that we were ready to go out and I get an e-mail and
it’s like someone from Europe with a problem 000I feel somewhat guilty if I
don’t respond. (Escalation engineer)
E-mail Emotion Fear of falling
behind
Often e-mail overload gets to me. You get to the end of a day and you have
to spend six out of ten hours in meetings and the rest of it on conference
calls and you haven’t been able to deal with the e- mails and you got
more than normal that day and they’re all just piling up and you feel like
they all need you to do something. That’s the point where I end up either
working extra hours or come in on the weekend to get on top of it.
Because you just know that if you have to come in on Monday to that
level of e-mail, it’s not going to be a good week. (Escalation manager 2)
E-mail Emotion Fear of missing
something
If e-mail starts building to a level where it’s just at saturation point, then
there’s the chance that an e-mail will get missed or you just won’t do what
you need to do. (Escalation manager 2)
E-mail Evaluation
negative
More work, overload Years ago we didn’t use e-mail. It added a step in my day is what it did. I
mean it definitely puts me in touch with other resources and makes it
easier to have various conversations. But it’s an added item. And so this
is a negative. (Workstation technologist)
E-mail Evaluation
positive
Keeps us in contact So I mean [checking e-mail is] like the top, number one priority. That’s the
way you stay connected, you know, not only with the team in the Bay
area, but also with the various people in Europe and Japan, in particular.
(International Web manager)
E-mail Practice Answer on arrival When you’re sitting there working on something and you hear that little ring,
it’s hard not to go over and see what it is. I don’t know how many times I
check it in a day, probably, I don’t know. It seems to be all day. (Technical
writer)
E-mail Practice Keep inbox clean I try to keep my inbox as clean as possible, although recently I haven’t been
able to drop it below 200. I normally would like to keep my e-mail inbox
below 100. I try to keep the ones that I’m working on the inbox and just
look at the new ones coming in. (Program manager)
E-mail Practice Respond to boss
and staff
immediately
If it’s from my boss, I’ll always read that. If it’s from my coworkers, I’ll read
that. And then it kind of goes down from there. (Escalation engineer)
E-mail Reason for
using e-mail
versus other
media
Create a record E-mail is good because it’s a written record of technical things, especially in
computers. It might be a list of commands or a certain way to do things.
That translates well to e-mail, which doesn’t translate well over the phone.
It’s a lot easier for them to maybe take a screen shot or cut and paste
some text into a mail message. So that’s certainly a case where I would
use e-mail over the phone. (Escalation engineer)
E-mail Reason for
doing e-mail
on weekend
Handle backlog of
work
I’m usually on [e-mail on Sunday] three or four hours continuously, and that’s
sort of to sort of clean the plate out. (Web manager)
E-mail Reasons for
doing e-mail
in the evening
Others in later time
zone
[I sometimes log into e-mail after dinner] because it’s uninterrupted. So it’s
e-mailing out. And sometimes Asia will respond back and they sort of
seem to like that, so I’m sort of encouraged to continue it because they’re
so many hours out of sync. They start coming alive 5:00 or 6:00 at night,
so if I’m on responding to them, it’s pretty helpful. (Web manager)

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
896 Organization Science 22(4), pp. 887–906, © 2011 INFORMS
Concurrent with developing these codes, we also
began exploratory analyses of the quantitative data to
reveal patterns among the larger sample of respondents.
These early analyses were invaluable in helping elimi-
nate several lines of inquiry and focus our subsequent
qualitative analysis on explicating patterns and rela-
tionships that appeared to be robust. For example, we
initially suspected that we would detect occupational
differences in patterns of technology use. However, the
quantitative analyses revealed no differences. We there-
fore combed the interview data in search of occupational
differences, only to find that they too revealed no con-
sistent differences in how members of the three occu-
pational groups thought about, experienced, or used the
media we examined. The second phase of coding, there-
fore, led us to abandon this line of inquiry in subsequent
analyses.
Phase 3: During the third phase of qualitative anal-
ysis, the authors applied the second- and third-level
codes to the entire corpus of interview data. Each of
the authors initially coded the same subset of transcripts
using the codes developed in the second phase. We
then met to compare our codings, discuss discrepancies,
and refine our understanding of when a code should
be applied. Each author then coded a second common
subset of transcripts to assure ourselves that we were
applying codes consistently. Having assured ourselves
that disagreements were now minor, each author sepa-
rately coded a third of the remaining transcripts.
Phase 4: In the final phase of qualitative analysis,
we identified patterns and relationships in the inter-
view data and triangulated those against the quantitative
data from the survey and logs. For each second-level
code, we wrote analytic memos that summarized pat-
terns. For example, we developed a memo that summa-
rized “norms about e-mail use,” which included counts
of the number of informants who referred to a par-
ticular norm, such as responsiveness. Another analytic
memo summarized all mentions of e-mail practices, such
as the practice of “answering e-mail on arrival.” All
analytic memos included illustrative quotes and, impor-
tantly, noted exceptions to identified patterns, which led
to additional analyses that helped us understand the rea-
sons for the exceptions.
We triangulated the patterns and tentative theories that
emerged from these analyses with the quantitative data,
which enabled us to explore further, affirm, and some-
times refute our understandings. Similarly, when our
quantitative analysis pointed to relationships between
variables, we turned to our interview data to deepen our
understanding of these relationships. For example, our
quantitative analysis pointed to a relationship between
amount of time people spent doing e-mail and their sense
of overload. Our interview data enabled us to probe
informants’ subjective experiences of e-mail, including
the emotions and social norms they associated with
using the medium. Together, these two forms of data
enabled us to assemble an account that combined social
and material factors to explain how people experienced
the technologies they used.
E-mail and Stress
Our informants used a variety of communication tech-
nologies, and many told stories about how phones,
pagers, and other devices blurred the line between office
and home or disrupted their leisure. Yet they complained
only about e-mail. For instance, an escalation manager
we interviewed told us that his cell phone, PDA, and
pager, which he carried in pouches on what he called
his “nerd belt,” had allowed him to participate in a con-
ference call while touring Universal Studios with his
family:
With my cell phone, two-way pager, and Palm Pilot, I can
work anywhere. I’ve worked escalations in Disney World.
In fact, I can remember exactly where I was because
it made such an impression on me. I think it was the
Popeye thing. It was actually in Universal Studios. So the
kids are going down the whatever and I’m on the phone
with an engineer talking to him about a problem with a
storage array. Then we did a conference call with some
people from the field service organization and came up
with an action plan.
Later in the interview he told us that he often stopped
his car to “work escalations on the side of the road”
with his cell phone. Rather than view such interruptions
as problems, he took pride in his accessibility, noting
that being constantly available by phone and pager was
part of his job and that he appreciated the flexibility that
mobile devices provided. E-mail, however, was another
story. He was bothered by the volume of e-mail that
he received. “E-mail is a fire hose,” he told us. “You
have got so many requests coming at you from so many
directions that it can be overwhelming. Even at the best
of times it is hard to manage.”
This manager’s complaints about e-mail, like those of
most other informants, mirrored the literature’s explana-
tion for why e-mail contributes to overload. Informants
worried about how to handle the sheer volume of e-mail
they received and the extra work it created. For example,
they told us that e-mail created extra work because it
often contained requests that led them to turn their atten-
tion to tasks that they had not planned on performing:
What [really] throws me off is—we have a manager that
sends us stuff through e-mail and says—okay, you all
need to work on this right now. And that happens pretty
often. It raises my blood pressure. It changes my focus.
It stresses me out. But you’ve got to do what you’ve got
to do. (System administrator in escalation group)
Informants also insisted that e-mail led them to work
longer hours. For instance, an escalation engineer admit-
ted that she came to the office early and ate lunch at her
desk to stay on top of the flow of messages:

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
Organization Science 22(4), pp. 887–906, © 2011 INFORMS 897
The number of e-mails that come in on any given day
is between 150 and 200. So you have to find a way to
just deal with the sheer volume. I tend to sit at my desk
over lunch, if there’s no other opportunity. There tends
to be three main times when I deal with e-mail—early in
the morning, lunchtime, and before I go home, typically
between 5:00 and 6:00.
The majority of informants told us that they worked
from home in the morning, in the evening, and on the
weekend primarily to clear their inboxes so that they
would not be overwhelmed by e-mail when they arrived
at work. The communication logs indicated that nearly
60% of our respondents handled work-related e-mail
from home at some point during the three days that they
recorded their communications. Most informants who
did not do their e-mail from home told us that they had
stopped doing so because the practice had led to conflict
with their spouse, significant other, or children. Rather
than jeopardize their relationship with loved ones, they
accepted a ban on doing work-related e-mail at home.
Although some informants recalled that they brought
work home before they had e-mail, they felt that e-mail
made the boundary between office and home more per-
meable. As a marketing manager put it,
Years ago, when we didn’t have all this technology, you
could take a briefcase and work at home. But it was
harder. Now there’s a complete blurring of the lines. It
could be nighttime, could be whenever, you have access
to your files, you have access to your e-mail. You can
work from anywhere. So the general theme for me over
time has been a theme of the work hours increasing.
Our informants’ tendency to blame e-mail for over-
load stood in sharp contrast to how they understood and
experienced other communications activities. The clos-
est informants came to evaluating other media nega-
tively was to say that meetings and teleconferences were
often a “waste of time.” Some told us that they actually
liked teleconferences because during a teleconference
they could “multitask”: they could process their e-mail
while on the phone. One technical writer was so enthu-
siastic about teleconferencing that she saw it as one of
the benefits of her job:
We have such a huge volume of e-mail that it’s unbeliev-
able. So much work comes through e-mail. Lots of times
I’ll be on teleconferences where they’re talking about
things that are not relevant to me, but I know that they
will get to something that’s relevant to me. So I just start
doing e-mail while I’m on the phone. If I was in the [con-
ference room], I wouldn’t be able to do that. But because
I’m on the phone, they can’t see me doing e-mail, so I’m
able to get work done. And that is just awesome! It’s one
of the things I love the most about my job.
It was not uncommon for our informants to teleconfer-
ence from their office into meetings that were being held
in the same building.
The data from our survey and communication logs
support the informants’ interpretations of e-mail and
stress, but complicate their belief that e-mail was the pri-
mary culprit. The regression results in first and second
columns of Table 4 ask whether various communication
activities and being a manager affected the total amount
of time that respondents reported working on the week-
days they logged. The regression in the first column uses
variables that measure the number of communications
in which respondents engaged using each medium. The
regression in the second column uses variables that indi-
cate the time that respondents spent with each medium.
As our informants contended in the interviews, the first
regression shows that the number of e-mails handled did
extend the time our respondents reported working. But
in contrast to our interviewees’ accounts, e-mail was
not the only problematic medium. The results show that
making phone calls and participating in teleconferences
also extended the hours they worked.
Similarly, the data in the third and fourth columns of
Table 4 support and at the same time complicate our
informants’ stories about overload. The regression in the
third column asks whether overload was influenced by
the number of hours that respondents worked, being a
manager, and the volume of communication using each
medium. The fourth column presents the relationship
between overload and the time spent using each medium.
The results corroborate our informants’ claim that the
time they spent doing e-mail was positively associated
with feeling overloaded. In contrast to what interviewees
believed, however, the regressions suggest that overload
was independent of the hours that respondents worked.
Furthermore, although no informants mentioned time
spent in meetings as a source of stress, the regressions
indicate it was.
In sum, our informants’ accounts in the interviews
and our analysis of respondents’ communication logs
and survey responses concur that the more e-mail they
handled, the longer they worked, and the more time
they spent doing e-mail, the more overloaded they felt.
However, the regressions contradict our informants’ and
the literatures’ assumption that the relationship between
e-mail and stress is mediated by the amount of time
spent working. Instead, our analysis suggests that e-mail
is related to stress regardless of how much time people
work. Furthermore, even though our informants did not
attribute stress to other communication technologies, the
quantitative data show that teleconferences and phone
calls were associated with working longer hours and that
time spent in meetings exacerbated overload.
To understand why e-mail was the only medium that
our informants blamed for the stress they experienced
requires delving into how e-mail’s material properties
interacted with the specific anxieties that e-mail evoked,
the norms that governed its use, and the temporal distri-
bution of communicative acts over the course of a day.

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
898 Organization Science 22(4), pp. 887–906, © 2011 INFORMS
Table 4 Effects of Communication Media on Time Worked, Overload, and Coping
Hours worked Overload Coping
(1) (2) (3) (4) (5) (6)
Variables Number Time Number Time Number Time
Intercept 16083∗∗∗ 16069∗∗∗ 0031 −0042 −00592 −00666
Time worked — — 0001 −0002 00001 00028
Manager 0007 −0032 0039 0019 −00738∗∗∗ −00836∗∗∗
No. of meetings 0032 0005 00042
No. of encounters −0010 0002 00012
No. of e-mails handled 0001∗∗ 0000 00002∗∗
No. of phone calls 0014∗∗ −0001 00018
No. of teleconferences 0039∗∗∗ 0005 −00024
Time in meetings 0016 0009∗∗ 00056
Time in encounters 0029 0002 00141
Time doing e-mail 0022 0013∗∗∗ −00016
Time on phone 0046 −0001 00087
Time in teleconferences 0046∗∗∗00002 −00061
N74 74 72 72 71 71
R20028∗∗∗ 0034∗∗∗ 0015 0025∗∗ 0025∗∗∗ 0023∗∗
∗∗p≤0005; ∗∗∗ p≤0001.
Our data show that it was the entangling of these factors
that led informants to experience and interpret e-mail as
stressful independent of how much time they worked.
The Anxiety of E-mail
Forty-five percent of our interviewees explicitly associ-
ated the volume of e-mail they received with a loss of
control, which they articulated in terms of two anxieties:
the fear of falling behind in one’s work and the fear of
missing important information. Both anxieties were tied
to the technology’s asynchrony, which enabled people to
send messages at any time without disturbing the recip-
ient and which allowed messages to accumulate in the
recipient’s inbox until processed. Informants spoke of
the fear of falling behind using an imagery of a moun-
tain of messages that piled up unanswered, making them
feel overloaded and out of control. An escalation man-
ager likened the experience to being at the mercy of a
runaway assembly line:
Sometimes I get an e-mail from somebody that says,
“What are you doing answering e-mail on the week-
end?” It’s one of my methods of stress management.
I could work a normal 40-hour work week and then I’d
be stressed out on the weekend and in the evenings. It’s
less stressful for me to put in the hours because I don’t
have a backlog building up. The stacking up bothers me a
lot. I start feeling bad when the mountain is building. You
know, it’s like I Love Lucy, where she’s got the choco-
lates coming at her on the conveyor belt. I start stuffing
the chocolates in my shirt and in my mouth. It’s like,
“Oh no, this is not good.”
A program manager for a distributed team explained
how her peace of mind hinged on the state of her inbox:
“I like to keep my in box really low, because in the past
I’ve had the problem where it’s been so out of control
that you can never get control back.” A Web technologist
expressed similar sentiments: “I’m very, very organized
with my e-mail. I have my little folders, and if my inbox
gets over 100, I start being really frustrated. I feel like I
can’t prioritize.” In short, many informants associated a
clean inbox with being in control.
Informants also worried that in the mass of unopened
e-mail lay crucial information that, if missed, would
affect their ability to stay on top of their work and
threaten their aura of competence. An escalation engi-
neer explained that this was particularly true in this com-
pany because “that’s how [the company] conveys all
their updates. So if you miss an e-mail, you might miss
something important.” A project manager who received
between 150 and 200 e-mails a day concurred:
I feel like I need to deal with each day’s e-mail or oth-
erwise you are just guaranteed to lose or overlook some-
thing. If I don’t have access to my e-mail, like if I’m
ever off campus [out of the office], that’s when I feel sort
of not in control, like I’m missing something. It’s like
something could be happening with my project that I’m
not aware of.
The fear of missing something important prevented
the majority of our informants (75%) from using filters
and other features designed to screen and reduce the
volume of e-mail.
My e-mail is usually way out of control. One of the guys
I work with is just so organized and has like 50 e-mails in
his mailbox. And I’ve usually got about 2,000. He says,
“Let me show you how to do filters.” So he set one up
for me and I said—that’s really cool, and then he walked
away. And then one of the contractors that I worked with

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
Organization Science 22(4), pp. 887–906, © 2011 INFORMS 899
on the customer reference database kept saying, “You’re
not responding to my e-mails.” I said, “I’m not getting
e-mails.” She said, “I’m sending you stuff.” I finally real-
ized the filter he set up was for anything that would come
from this woman and it was all getting filed and I wasn’t
seeing it. I haven’t used filters since. (Marketing editor)
Similarly, people told us they were reluctant to “unsub-
scribe” from large e-mail distribution lists for fear of
missing an important item of information. A technical
writer explained, “You can subscribe and unsubscribe
from some of these groups, it’s just that you feel you
may miss something that you would like to know.”
Like other informants, a vice president feared that not
reading every e-mail would compromise her reputation.
She shared with us what she feared others might say
about her:
In business it’s important to be up to date, and so when
you’re in a meeting, if you haven’t seen what was due or
you haven’t seen this or that, you’re “unorganized,” “not
up to date.” Being surprised is not a good thing. So you
want to look on top of it and be on top of it, because if
you’re not, “you’re trying to fake it.”
In short, informants felt compelled to keep up with
their e-mail regardless of whether it meant that they
worked longer hours or took work home. Doing e-mail
eased informants’ anxieties and allowed them to feel as
if they were in control. The regressions in the fifth and
sixth columns of Table 4 corroborate this claim and are
consistent with the paradox of doing e-mail found in the
emerging literature on mobile devices. The regression in
the fifth column shows that the number of e-mails peo-
ple processed was related to their sense of coping. The
data also suggest that managers apparently had more
difficulty coping than respondents without managerial
duties did, a finding that is reconfirmed by the regres-
sion in the sixth column, which uses measures of time
on technologies as the independent variables. In light of
the foregoing discussion, it would seem that handling
more messages bolstered the respondents’ sense of cop-
ing because processing e-mail reduced the size of their
inbox and allayed the anxiety of allowing e-mail to go
unanswered. Interestingly, the more time people spent
doing e-mail, the more overloaded they felt, but the more
messages they handled, the more they felt they could
cope. Of course, any experience of relief was, at best,
fleeting, because new e-mail would soon arrive.
Norms About E-mail
Whereas fears of falling behind and missing informa-
tion motivated informants to handle all messages so
they would feel in control, social norms pressured them
to do so quickly. Although, in theory, e-mail’s asyn-
chrony should have granted recipients the leeway to
respond at a time that was convenient for them, our
informants described strong cultural expectations about
not keeping senders waiting. What constituted an accept-
able wait, however, varied. Some informants told us that
they expected responses within hours. Others claimed
the outside limit was a day. Still others felt timing
was contingent on subject matter or the sender’s prior
responsiveness.
Nevertheless, all informants felt obligated to be
responsive and expected the same of their coworkers,
however they might define it:
I’m very good about responding to e-mail. I rarely let
things sit without a response for more than a day. I’ll
respond saying either I’ve done it or that I’ll look into
it or whatever. Even if I have not completed an action,
I’ll let them know that I’m working it. So because of my
personal thing, I expect people to respond to me within a
day or so when I send them e-mail. (Marketing manager)
Those who answered their e-mail quickly enhanced their
reputation by doing so. As a senior manager explained,
colleagues who were responsive were seen as “really
sensitive” and “really caring.”
Some informants understood that their own behavior
was responsible for creating and perpetuating the norm
of responsiveness. As one escalation engineer explained
to us,
I think different people have slightly different expecta-
tions of when you read e-mails. I think it is based on your
previous levels of response. I typically respond within
reasonable time scales and, therefore, people have that
expectation of me. But the flipside to that is that I expect
people also to respond to my e-mail within reasonable
time scales.
Informants who adhered to the norm became indig-
nant when a coworker responded more slowly than they
thought reasonable:
It really drives me crazy when you send an e-mail to
somebody and they don’t respond. Even just, “I’m work-
ing on it. I got your e-mail, I’m working on it.” I at least
try to do that! So I get kind of miffed when I don’t hear
back. I’m always pleasantly surprised if I hear back from
somebody within an hour or two. I’m really appreciative.
Generally, I expect to hear something back in a day.
(Writer and editor for the marketing group)
To ensure that their coworkers responded to messages
in a timely manner and to shame those who did not,
informants resorted to a number of practices designed
to enforce responsiveness. The first practice was to label
e-mails as “urgent.” But informants told us that when
the strategy was overused, coworkers learned to ignore
it. The second practice was to follow an e-mail with a
phone call to signal to the recipient how important the
message was:
If I need to get in touch with someone, I might leave an
e-mail and then I’ll call. If they’re not in the office, I’m
going to leave a voicemail as well and just say, “Hey, I
left you an e-mail.” Or “Hey, read your e-mail.” There

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
900 Organization Science 22(4), pp. 887–906, © 2011 INFORMS
are some things that would be much too lengthy to talk
about over the phone, but I’ll use that [phone call] to
accentuate an important e-mail I might have sent.
(Escalation engineer)
The third tactic for enforcing the norm of responsiveness
entailed copying coworkers or bosses to create a measure
of accountability or to shame unresponsive coworkers
into changing their ways:
So one thing that I have to do with certain people is to put
another peer on the e-mail, and then I’ll get a response.
I try not to do a lot of one-upping, you know, like put
their manager on an e-mail—I don’t like doing that. I’ll
do that a half a dozen times a year. But I will put peer
managers on e-mails to elicit a response.
(Web manager, marketing group)
A second widely shared norm that shaped how infor-
mants experienced e-mail was rooted in the medium’s
asynchrony. People were allowed—even expected—to
send messages at any time of the day or night, particu-
larly if they worked with teams distributed across time
zones. This norm about when one could send e-mail con-
trasted sharply with informants’ expectations that they
should not call coworkers late at night or in the morning
unless a crisis made it impossible to avoid:
People don’t make a habit of calling me at home. It tends
to be more e-mail. So if I’m working, it tends to be
e-mail. Because it’s asynchronous, right? You don’t have
to bother other people at home, for example. (Marketer)
Even though senders felt free to send e-mail any-
time and were absolved of guilt about interrupting their
coworkers at night, it did nothing to absolve receivers
from the obligation to respond.
Flow of Daily Communication Events
Informants faced a conundrum in their attempts to keep
their e-mail under control. Most told us that they tried to
handle e-mail as it arrived, to maintain a sense of being
responsive and on top of their work.
I like to respond to [e-mail] right away. It’s easier to send
them out because I have so many e-mails coming in. It’s
easier for me to just keep things up than to come in and
have 20 that I have to answer. I’d rather check it every
couple of hours, you know, respond to two or three, and
be done with it. Then I know I’m up to date.
(Technical writer)
Our survey data support the informants’ claim that
they answered e-mail as it arrived; only 12% of respon-
dents said they tried to handle e-mail in batches. Con-
taining anxiety about e-mail and meeting expectations
of responsiveness were particularly difficult, however,
when the flow of other activities created quasi-material
pressures that precluded answering messages as they
arrived. Most of our informants worked on distributed
teams, with members in Asia, Europe, and the oppo-
site coast. Our informants were often asleep when team
members in other time zones were sending messages
during their own workday. Consequently, many infor-
mants started their mornings facing a loaded inbox:
The first thing I do when I get into work is to check my
e-mail, because there’s such a huge flow of e-mail. There
are a lot of teams and vendors we work with. There’s just
so much communication flying around. I always need to
check my e-mail and make sure that there’s no problem
that happened overnight, especially because some of our
vendors are in Europe and Asia, and things might hap-
pen during their day and then I won’t find out until my
morning. I need to see if there are any crises that I need
to take care of. E-mail often shapes the early part of my
morning. (Technical writer)
Working across time zones was not the only way
in which the natural flow of daily activities prevented
employees from answering e-mail as it arrived. An esca-
lation manager explained,
Often e-mail overload gets to me. You get to the end
of a day and you have to spend six out of ten hours
in meetings and the rest of it on conference calls, and
you haven’t been able to deal with the e-mails, and you
got more than normal that day, and they’re all just pil-
ing up, and you feel like they all need you to do some-
thing. That’s the point where I end up either working
extra hours or come in on the weekend to get on top of
it. Because you just know that if you have to come in
on Monday to that level of e-mail, it’s not going to be a
good week. So yeah, there are times where I get stressed
by it. If e-mail starts building to a level where it’s at sat-
uration point, then there’s the chance that an e-mail will
get missed and you won’t do what you need to do.
Unlike other informants, this manager seemed to rec-
ognize why she found e-mail so stressful: it built up
unanswered when she was occupied by other activities.
This buildup, coupled with expectations for immediate
response, meant that she frequently approached the end
of her day with the anxiety that her work was out of
control and that her obligations remained unmet.
This manager’s insight into how e-mail fit into the
flow of a day is supported by the patterns displayed in
Figure 1, which plots the frequency of e-mail sessions
and the frequency of the combination of phone calls,
meetings, and teleconferences by hours of the day. The
figure shows that respondents experienced two peaks of
e-mail activity: at 9 a.m. and between 3 p.m. and 4 p.m.
The morning peak, which was partially the result of
working across time zones, occurred just as other com-
munication activities reached a crescendo. E-mail use
subsequently fell and did not reach its second peak until
the late afternoon, when people could return to their
inboxes after other communications had begun to sub-
side. Because respondents felt pressure to answer all
e-mail and felt out of control when they did not, suc-
cessfully handling the messages that accumulated during
the day contributed to their sense of coping with the

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
Organization Science 22(4), pp. 887–906, © 2011 INFORMS 901
Figure 1 Distribution of E-mail, Phone Calls, Meetings, and
Teleconferences During the Day
0
20
40
60
80
100
120
140
123456789101112131415161718192021222324
Hour of the day
Frequency
Other media
E-mail
demands of their work, even as the time they spent doing
so contributed to their sense of overload.
If other communication activities led respondents to
push e-mail to the end of the day, and if respondents
experienced this accumulation of e-mail as particularly
stressful, then one might expect respondents who par-
ticipated heavily in other communication activities to
be especially susceptible to e-mail-induced stress. Said
differently, time doing e-mail should be significantly
associated with overload primarily for those respondents
whose days were consumed with other activities. To
explore this possibility, we calculated the proportion of
the day that each respondent spent in the combination of
meetings, encounters, teleconferences, and phone calls.
We split the sample at the median to create two groups:
high and low users of other media. For each group sepa-
rately, we regressed time spent doing e-mail on overload.
Although on average members of the two groups han-
dled the same volume of e-mail (163 versus 173 mes-
sages, t= −0035, p=72), time spent doing e-mail was
significantly associated with overload only for respon-
dents who participated heavily in other communication
activities.8
Understanding e-mail’s position in the flow of daily
communications helps us untangle the contradiction
between our interviewees’ contention that e-mail was
their primary source of overload and our quantitative
data that indicate that meetings and teleconferences were
just as troublesome. When informants spent significant
portions of their days in other communication activities,
e-mail built up and became figural just as their work-
day was about to draw to a close. They also knew that
the backlog would grow by the next morning because of
messages from other time zones. Facing a backlog at the
end of the day and anxieties about falling behind and
missing important information, coupled with the norm
for immediate response, led informants to conclude that
they were overloaded and that additional work would be
required to regain control.
In short, rather than attending to how much time tele-
conferences and meetings consumed or to the added
demands from working across time zones, interviewees
focused on their inbox as the salient source of over-
load and the target of their complaints. In this way,
e-mail became symbolic of the cumulative demands that
stretched work beyond what informants could reason-
ably manage in the course of a day.
Discussion
Sociomaterial distinctions between e-mail and other
technologies help explain why e-mail became a symbol
of overload and other communication activities escaped
blame. E-mail’s material features, specifically those that
stored messages in people’s inboxes until they handled
them, enabled responses to be temporally decoupled
from messages sent, which meant that people could send
e-mail at any time of the day or night ostensibly with-
out disturbing their recipients. These material proper-
ties of e-mail combined with the quasi-material temporal
rhythms of the respondents’ work to ensure that time
away from continuous processing resulted in a build-
up. Respondents who worked across time zones woke
up to accumulated mail from coworkers whose day had
already begun. Respondents who spent a significant por-
tion of the day on the phone and who attended meetings
and teleconferences returned at the end of the day to
swollen inboxes. Thus, the inbox served as a continuous
and tangible reminder of how overloaded one was.
Backlogs of e-mail made salient social norms about
responsiveness and anxieties about losing control. Infor-
mants feared that if they did not handle their mount-
ing e-mail, they would fall behind or miss important
information for which they would be held account-
able. Expectations that obligated them to answer e-mail
quickly further exacerbated these anxieties. Because of
these norms and anxieties, informants did not feel they
could ignore backlogs of e-mail. Consequently, to main-
tain a sense of control, they either extended their work-
day or did e-mail at home in the evening, in the morning,
and on weekends.
Informants’ experiences of meetings and telecon-
ferences were vastly different. Because meetings and
teleconferences were synchronous media, they left no
material reminder of unaccomplished work. Informants
spoke about norms requiring attendance at meetings and
teleconferences, but once a meeting or teleconference
was over, both the opportunity and obligation to attend
disappeared. Although people might need to make an
effort to learn what they had missed, informants held
that only a small portion of any particular meeting or
teleconference was relevant to them. This was why they
preferred teleconferences over meetings. Because their

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
902 Organization Science 22(4), pp. 887–906, © 2011 INFORMS
telephones were equipped with muting switches, infor-
mants could put their phones on mute and catch up on e-
mail while listening for cues that signaled that something
in the meeting was relevant for them. In other words,
teleconferences were welcomed because they offered an
opportunity to regain control over an inbox. The organi-
zation’s norms reinforced multitasking. Not only was the
practice widespread, but there was no sanction against
attending meetings remotely even when they occurred
in the same building. The only proscription was not to
be caught off guard when directly asked a question or
when the meeting moved to a topic on which one had
an obligation to speak.9
Because phones were synchronous, like meetings,
they too required co-presence for communication to
occur. Social norms entwined with the phone’s syn-
chrony to make it unlikely that informants would view
phone calls as a source of overload. Although, in theory,
informants could have made phone calls at any time of
the day, they experienced strong norms about when such
calls were appropriate. In particular, calling coworkers at
home late at night or in the morning was taboo, except
in the case of a perceived crisis.
In short, one cannot account for why e-mail was
singled out as a symbol of stress without considering
the unique entanglement of material, social, and quasi-
material factors surrounding its use. Absent e-mail’s
asynchrony, there would have been no buildup of mes-
sages for informants to confront. Absent social norms
about responsiveness and anxieties about falling behind
or missing an important message, respondents could
have ignored backlogs, especially at the end of the day.
Absent the flow of work across time zones and the dis-
tribution of communication activities during the course
of a day, people could have responded more easily to
e-mail messages as they arrived, thereby ameliorating
pileups at the beginning and end of the day.
Like any study, ours has limitations. First, because
we collected our data in 2001 and 2002, before Black-
Berries and other mobile e-mail devices were widely
diffused, the mix of media in use today is likely to
be different. It is important to remember, however, that
smartphones are typically used to handle not only phone
calls but e-mail anytime, anywhere. We suspect, there-
fore, that our findings are, at worst, conservative. We
would expect mobile e-mail to exacerbate both the sense
of being overloaded and the sense of being in greater
control, because e-mail can be processed continuously
with these devices. Furthermore, mobile e-mail should,
if anything, amplify norms for quick response because
senders could expect that recipients always have access
to e-mail. This seems to be precisely what emerging
studies on the use of mobile devices show (Jarvenpaa
and Lang 2005, Mazmanian et al. 2006, Middleton and
Cukier 2006).
Second, because our study did not entail observation
of individuals’ daily work, we cannot speak to the top-
ics that people discussed when communicating, nor can
we address the interplay between communication activ-
ities and the amount and nature of other types of work.
It is possible that the overload that people reported was
partly contingent on the other tasks they were perform-
ing. More important, because we did not observe or ask
about unfinished tasks, it is also possible that the dis-
tribution of unfinished work on the days we collected
data may have affected both the number of communica-
tions in which respondents engaged as well as the over-
load that they felt. No informant, however, mentioned
such work during our interviews, even though some must
have had unfinished work. This raises the intriguing pos-
sibility that in some settings e-mail may operate as a
symbol of all unfinished work.
Third, our data come from employees of one engi-
neering organization. Samples drawn from other settings
might reveal that e-mail’s relationship to stress would
vary across contexts with different preferences for tech-
nology, different norms, and different flows of work. Our
results might have been different had we studied a sales
organization in a single geographical location with a
“phone culture.” In such a setting, voicemail might have
acquired some of the symbolic attributes that e-mail had
in this organization.
Finally, we have no way of estimating how accurately
our respondents logged their communications. It seems
unlikely that respondents would have recorded events
that did not happen. It is possible, however, that they
may have omitted communication events that occurred
or inaccurately estimated their duration. Given the rel-
ative frequency of the various types of communica-
tions and the physical demands of logging, we suspect
that e-mail sessions or phone calls would constitute the
bulk of omitted events. Together, these two considera-
tions imply that errors in logging would have biased our
results in a conservative direction, because there is no
reason to believe that people who processed relatively
few e-mail messages would have been more likely to
omit their occurrence. On the contrary, it would seem
more reasonable for people who received the most mes-
sages to have failed to record some of their e-mail ses-
sions. If this were true, in the absence of omissions, the
relationship between e-mail and reported overload would
have been even stronger.
Implications
Our story complicates standard conceptions of why
e-mail and other communication media create feelings of
overload. Explanations in both the work–life and tech-
nology literatures blame the technologies’ material fea-
tures. The work–life literature tends to attribute overload
to features that enable people to send and receive work-
related e-mail anywhere at anytime, which allows work

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
Organization Science 22(4), pp. 887–906, © 2011 INFORMS 903
to spill into other domains of life. Technology studies
usually attribute overload to the volume of messages that
people receive and to the extra time it takes to handle
them, the tasks associated with them, and the interrup-
tions they create. These are important sources of stress,
and our study confirms that the volume of e-mail did
increase the length of the respondents’ workday. More-
over, the more time respondents spent handling e-mail,
the more they felt overloaded. In contrast to earlier
research, however, the extra time people spent work-
ing, either inside or outside the office, did not appear to
mediate the relationship between e-mail and the experi-
ence of overload. Instead, e-mail appeared to be related
to stress, regardless of how many hours respondents
worked. As we have shown, to explain why people expe-
rience e-mail as stressful regardless of the amount of
work it generates requires taking into account the simul-
taneous influence of social norms and interpretations, the
temporality of work flow, and the patterns of use enabled
by the material features of the medium.
Furthermore, our discovery that e-mail functioned as
a symbol is important both theoretically and practically.
With few exceptions (e.g., Barley and Knight 1992,
Meyerson 1994), researchers have not examined the
symbolic aspects of either stress or e-mail. Recognizing
e-mail’s symbolism illuminates why workers attended so
closely to e-mail’s consequences while overlooking how
social norms, the structure of their work, other technolo-
gies, and the flow of their workday also contributed to
their sense of being overloaded and losing control.
As a symbol, e-mail became the interpretive scapegoat
for the workers’ perceptions that they were expected
to do more than they could reasonably accomplish in
a day. It provided them with a culturally sanctioned
rhetoric of complaint about overload as well as a tan-
gible ritual for regaining control: to cope with over-
load, trim your inbox. Although students of technology
have long pointed to the importance of interpretations
for understanding a technology’s uses and effects (e.g.,
Orlikowski 1992), the possibility that technologies can
be transformed into powerful cultural symbols, and that
as symbols these technologies have broader implications,
has not been investigated.
The symbolic nature of e-mail may well extend
beyond the site we studied. Anecdotal evidence for this
can be found in the popular press, where e-mail is often
singled out as the culprit that has destroyed the qual-
ity of life. For example, writing for information indus-
try professionals, Adam (2002, p. 89) proclaims that
“managers have reported that e-mail causes them more
stress than either conflict with the boss or dealing with
customer complaints.” On April 22, 2005, news media
across Britain and North America rushed to report on
a study that claimed e-mail was more distracting than
smoking pot (British Broadcasting Corporation 2005,
Cable News Network 2005, Orlowski 2005).
Admitting that e-mail serves as a symbol of general
overload implies that attempting to ameliorate overload
by redesigning e-mail’s material features or by chang-
ing how people use those features is unlikely to reduce
the stress associated with it. Consider the recommenda-
tion that users process e-mail in batches to reduce the
frequency of interruptions. Our informants were techno-
logically sophisticated: they worked for one of the most
prominent computer companies in the world and many
had technical expertise in software design. Yet few used
filters, and almost none chose to answer their e-mail
in batches. We suspect that our informants did not fil-
ter or batch because such tactics were of little use for
relieving the stress they felt. Overload was not a mat-
ter of the number of messages they received: it was the
joint product of the time they spent handing messages,
the anxieties they felt, the norms of responsiveness they
accepted and reproduced, and a daily pattern of com-
munication activities that they could not control. To be
truly successful, any attempt to redress e-mail overload
would have to address this sociomaterial entanglement
head on.
Moreover, the increased pressure on people’s home
lives caused by e-mail’s capacity to blur boundaries and
accumulate may be experienced more intensely by some
groups of people. Those who have primary responsibility
for dependents or other obligations outside of work that
would prevent them from dealing with their inbox in the
morning and evening will likely suffer additional anxiety
from the accumulation of messages over time. Whether
people with ongoing responsibilities outside of work suf-
fer additional overload from their mounting inboxes or
are penalized at work for not being continually accessi-
ble or responsive is a topic for future research.
The sense of being overwhelmed and overloaded is
also likely to become more common and severe as orga-
nizations place employees on multiple teams and make
greater use of distributed work in their efforts to glob-
alize (Hinds and Kiesler 2002, O’Leary et al. 2008).
Distributed teams increase the need for employees to
participate in phone calls and teleconferences and to
exchange e-mails across time zones. Our data show that
e-mail, teleconferences, and phone calls added hours to
the workday. Furthermore, our analysis indicates that
teleconferences and working across time zones created a
backlog of messages at the beginning and end of the day,
which contributed to people’s sense of being overloaded.
To the degree that e-mail masquerades as a simple mate-
rial cause while functioning as a symbol of overload,
employees and organizations are unlikely to recognize
and address the larger problem: new patterns of work
that crowd days and create unrealistic expectations about
response time. To the degree that e-mail’s symbolic force
diverts attention from the stress created by the demands
being placed on a downsized and globalized workforce,
it serves as a red herring.

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
904 Organization Science 22(4), pp. 887–906, © 2011 INFORMS
Acknowledgments
This research was supported by funding from the Ford Foun-
dation to study “Work, Technology, and the Everyday Lives
of Women and Men” (990-0867), the National Science Foun-
dation under an SGER grant to study “Technology and the
Social Construction of Availability” (NSF 0328662), and the
Stanford/GM Collaborative Research Laboratory.
Endnotes
1Sociotechnical system theorists, writing in the 1950s and
1960s, were the first to argue that technological change
implicated both social and material phenomena (Trist and
Bamforth 1951, Emery and Marek 1962, Rice 1963). How-
ever, sociotechnical systems theorists quickly came to focus
almost entirely on social interventions, in particular, the use of
autonomous work groups (Pasmore 1988). Although one could
use “sociotechnical” as a synonym for “sociomaterial,” we pre-
fer the latter term because it has fewer historical connotations.
2Our study occurred just as mobile e-mail devices were emerg-
ing. Only one respondent had a BlackBerry, and only 22%
had ever used instant messaging. Videoconferencing technol-
ogy was available, but respondents almost never used it.
3In keeping with the conventions of sociology, we refer to
the people we studied as respondents when speaking about
our quantitative data and as informants when discussing data
collect through interviews.
4Readers interested in obtaining a copy of the communication
log should contact the authors.
5For most people, the workplace was an assigned office in a
building owned by the company. But even people who worked
from home had designated spaces where they worked and
could easily determine when they started and when they ended
their workday, even if they had never left home.
6We initially distinguished between land and cell phones
because people used cell phones in more settings than they
used land phones and because some respondents used cell
phones and land phones for different purposes and audiences.
We combine them for the present analysis because we found
that distinguishing between land and cell phones made no dif-
ference in the analysis.
7Readers who are interested in seeing regressions that include
these variables should contact the authors.
8Readers interested in seeing these regressions should contact
the authors.
9Although informants justified such multitasking by its per-
ceived ability to help them cope with the flood of e-mail, in
reality, doing e-mail while teleconferencing may not have been
as effective as they thought. Research indicates that although
multitasking is becoming more prevalent (Reinsch et al. 2008),
it may actually impair performance. In a recent series of
eight laboratory experiments, Ophir et al. (2009) showed that
heavy multitaskers were more easily distracted; had more trou-
ble storing, organizing, and remembering information; and
had more difficulty filtering out irrelevant information. Iron-
ically, they also had more difficulty switching between tasks
effectively.
References
Adam, R. 2002. Is e-mail addictive? Aslib Proc. 54(2) 85–94.
Alvarez, L. 2005. Got 2 extra hours for your e-mail? New York Times
(November 10), http://www.nytimes.com/2005/11/10/fashion/
thursdaystyles/10EMAIL.html.
Bälter, O. 2000. Keystroke level analysis of email message organiza-
tion. Proc. CHI 2000 Conf. Human Factors Comput. Systems.
ACM, New York, 105–112.
Barley, S. R. 1986. Technology as an occasion for structuring: Evi-
dence from observations of CT scanners and the social order of
radiology departments. Admin. Sci. Quart. 31(1) 78–108.
Barley, S. R. 1988. On technology, time, and the social order: Tech-
nically induced change in the temporal organization of radio-
logical work. F. A. Dubinskas, ed. Making Time: Ethnographies
of High-Technology Organizations. Temple University Press,
Philadelphia, 123–169.
Barley, S. R. 1990. The alignment of technology and structure through
roles and networks. Admin. Sci. Quart. 35(1) 61–103.
Barley, S. R., D. B. Knight. 1992. Toward a cultural theory of
stress complaints. B. M. Staw, L. L. Cummings, eds. Research
in Organization Behavior, Vol. 14. JAI Press, Greenwich, CT,
1–48.
Bellotti, V., N. Ducheneaut, M. Howard, I. Smith. 2003. Taking email
to task: The design and evaluation of a task management cen-
tered email tool. Proc. CHI 2003 Conf. Human Factors Comput.
Systems. ACM, New York, 345–352.
Bellotti, V., N. Ducheneaut, M. Howard, I. Smith, R. E. Grinter.
2005. Quality versus quantity: E-mail-centric task management
and its relation with overload. Human-Comput. Interaction 20(1)
89–138.
Berghel, H. 1997. E-mail—The good, the bad, and the ugly. Comm.
ACM 40(4) 11–15.
Bernard, H. R., P. Killworth, D. Kronenfeld, L. Sailor. 1985. The
problem of informant accuracy: The validity of retrospective
data. Annual Rev. Anthropol. 13(1) 495–517.
Bogg, J., C. L. Cooper. 1994. An examination of gender differences
for job satisfaction, mental health and occupational stress among
senior U.K. civil servants. Internat. J. Stress Management 1(2)
159–172.
Boswell, W. R., J. B. Olson-Buchanan. 2007. The use of communi-
cation technologies after hours: The role of work attitudes and
work–life conflict. J. Management 33(4) 592–610.
British Broadcasting Corporation. 2005. “Informania” worse than
marijuana. BBC News (April 22), http://news.bbc.co.uk/2/hi/uk
_news/4471607.stm.
Cable News Network. 2005. E-mails “hurt IQ more than pot.” CNN.com,
(April 22), http://www.cnn.com/2005/WORLD/europe/04/22/
text.iq/.
Chesley, N. 2005. Blurring boundaries? Linking technology use,
spillover, individual distress, and family satisfaction. J. Marriage
Family 67(5) 1237–1248.
Chesley, N., P. Moen, R. P. Shore. 2003. The new technology climate.
P. Moen, ed. It’s About Time: Couples and Careers. Cornell
University Press, Ithaca, NY, 220–241.
Cooper, C. L., J. Marshall. 1976. Occupational sources of stress:
A review of the literature relating to coronary heart disease and
mental ill health. J. Occupational Psych. 49(1) 11–28.
Cooper, C. L., P. J. Dewe, M. P. O’Driscoll. 2001. Organizational
Stress: A Review and Critique of Theory, Research and Appli-
cations. Sage, Thousand Oaks, CA.
Dabbish, L. A., R. E. Kraut. 2006. Email overload at work: An analysis
of factors associated with email strain. Proc. ACM Conf. Comput.
Supported Cooperative Work. ACM, New York, 431–440.
Dabbish, L. A., R. E, Kraut, S. Fussel, S. Kiesler. 2005. Understand-
ing email use: Predicting action on a message. Human Factors
Comput. Systems: Proc. CHI’05. ACM, New York, 691–700.

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
Organization Science 22(4), pp. 887–906, © 2011 INFORMS 905
Dawley, D. D., W. P. Anthony. 2003. User perceptions of e-mail at
work. J. Bus. Tech. Comm. 17(2) 170–200.
Duxbury, L. E., I. Towers, C. Higgins, J. A. Thomas. 2006. From 9
to 5 to 24/7: How technology has redefined the workday. W. K.
Law, ed. Information Resources Management: Global Chal-
lenges. Idea Group Publishing, Hershey, PA, 305–332.
Edmondson, A. C., S. E. McManus. 2007. Methodological fit in
management field research. Acad. Management Rev. 32(4)
1155–1179.
El-Shinnawy, M., L. M. Markus. 1998. Acceptance of communica-
tion media in organizations: Richness or features? IEEE Trans.
Professional Comm. 41(4) 242–253.
Emery, F. E., J. Marek. 1962. Some socio-technical aspects of automa-
tion. Human Relations 15(1) 17–25.
Families and Work Institute. 1992. 1992 National Study of the Chang-
ing Workforce. Families and Work Institute, New York.
Fulk, J. 1993. Social construction of communication technology.
Acad. Management J. 36(5) 921–950.
Fulk, J., C. W. Steinfield, J. Schmitz, J. G. Power. 1987. A social
information processing model of media use in organizations.
Comm. Res. 14(5) 529–552.
Gergen, K. J. 2002. The challenge of absent presence. J. E. Katz,
M. Aarhus, eds. Perpetual Contact: Mobile Communication,
Private Talk, Public Performance. Cambridge University Press,
Cambridge, UK, 227–241.
Golden, A. G., C. Geisler. 2007. Work–life boundary management and
the personal digital assistant. Human Relations 60(3) 519–551.
González, V. M., G. Mark. 2004. “Constant, constant, multitasking
craziness”: Managing multiple working spheres. Human Factors
Comp. Systems: Proc. CHI’04. ACM, New York, 113–120.
González, V. M., G. Mark. 2005. Managing currents of work:
Multi-tasking among multiple collaborations. ECSCW 2005:
Proc. Ninth Eur. Conf. Comput.-Supported Cooperative Work.
Springer, New York, 143–162.
Green, N. 2002. On the move: Technology, mobility, and the media-
tion of social time and space. Inform. Soc. 18(4) 281–292.
Herbsleb, J. D., A. Mockus. 2003. Formulation and preliminary test
of an empirical theory of coordination in software engineering.
ESEC/FSE’03. ACM, New York, 138–147.
Herbsleb, J. D., A. Mockus, T. A. Finholt, R. E. Grinter. 2000.
Distance, dependencies, and delay in global collaboration.
Proc. ACM Conf. Comput.-Supported Cooperative Work (CSCW
2000). ACM, New York, 319–328.
Hill, J. E., A. J. Hawkins, M. Ferris, M. Weitzman. 2001. Finding
an extra day a week: The positive influence of perceived job
flexibility on work and family balance. Family Relations 50(1)
49–58.
Hinds, P. J., D. E. Bailey. 2003. Out of sight, out of sync: Understand-
ing conflict on distributed teams. Organ. Sci. 14(6) 615–632.
Hinds, P. J., S. Kiesler. 2002. Distributed Work. MIT Press,
Cambridge, MA.
Jackson, T. W., R. Dawson, D. Wilson. 1999. Improving the commu-
nications process: The costs and effectiveness of email compared
with traditional media. Proc. 4th Internat. Conf. Software Pro-
cess Improvement, Res., Ed. Training 4INSPIRE’995,Heraklion,
Crete, 167–178.
Jackson, T., R. Dawson, D. Wilson. 2001a. Case study: Evaluating the
use of an electronic messaging system in business. Proc. Conf.
Empirical Assessment Software Engrg. ACM, New York, 53–56.
Jackson, T., R. Dawson, D. Wilson. 2001b. The cost of email inter-
ruption. J. Systems Inform. Tech. 5(1) 81–92.
Jackson, T., R. Dawson, D. Wilson. 2003. Reducing the effect of
email interruptions on employees. Internat. J. Inform. Manage-
ment 23(1) 55–65.
Jacobs, J. A., K. Gerson. 1998. Who are the overworked Americans?
Rev. Soc. Econom. 56(4) 422–459.
Jacobs, J. A., K. Gerson. 2004. The Time Divide: Work, Family and
Gender Inequality. Harvard University Press, Cambridge, MA.
Jarvenpaa, S. L., K. R. Lang. 2005. Managing the paradoxes of mobile
technology. Inform. Systems Management 22(4) 7–23.
Jick, T. D. 1979. Mixing qualitative and quantitative methods: Trian-
gulation in action. Admin. Sci. Quart. 24(4) 602–611.
Kaufman-Scarborough, C. 2006. Time use and the impact of tech-
nology: Examining workspaces in the home. Time Soc. 15(1)
57–80.
Kotter, J. P. 1982. The General Managers. Free Press, New York.
Lazarus, R. S., S. Folkman. 1984. Stress, Appraisal, and Coping.
Springer, New York.
Leonardi, P. M., S. R. Barley. 2008. Materiality and change: Chal-
lenges to building better theory about technology and organiz-
ing. Inform. Organ. 18(3) 159–176.
Lofland, J., L. H. Lofland. 1984. Analyzing Social Settings:
A Guide to Qualitative Observation and Analysis. Wadsworth,
Belmont, CA.
Major, V. S., K. J. Klein, M. G. Ehrhart. 2002. Work time, work inter-
ference with family, and psychological distress. J. Appl. Psych.
87(3) 427–436.
Manger, T., R. A. Wicklund, O.-J. Eikeland. 2003. Speed commu-
nication and solving social problems. Communications 28(3)
323–337.
Mark, G., V. M. González, J. Harris. 2005. No task left behind? Exam-
ining the nature of fragmented work. Human Factors Comput.
Systems: Proc. CHI’05. ACM Press, New York, 321–330.
Markus, M. L. 1994. Electonic mail as the medium of managerial
choice. Organ. Sci. 5(4) 502–527.
Maslach, C., S. E. Jackson. 1981. The measurement of experienced
burnout. J. Occupational Behav. 2(2) 99–113.
Maslach, C., W. B. Schaufeli, M. P. Leiter. 2001. Job burnout. Annual
Rev. Psych. 52(1) 397–422.
Mazmanian, M., W. J. Orlikowski, J. Yates. 2006. CrackBerries: The
social implications of ubiquitous wireless e-mail devices. C.
Sørensen, Y. Yoo, K. Lyytinen, eds. Designing Ubiquitous Infor-
mation Environments: Socio-Technical Issues and Challenges.
Springer, New York, 337–344.
Meyerson, D. E. 1994. Interpretations of stress in institutions: The
cultural production of ambiguity and burnout. Admin. Sci. Quart.
39(4) 628–653.
Middleton, C. A., W. Cukier. 2006. Is mobile email functional or
dysfunctional? Two perspectives on mobile email usage. Eur. J.
Inform. Systems 15(3) 252–260.
Mintzberg, H. 1975. The manager’s job: Folklore and fact. Harvard
Bus. Rev. 53(July/August) 49–61.
Moen, P. 2003. It’s About Time: Couples and Careers. ILR Press,
Ithaca, NY.
Moen, P., Y. Yu. 2000. Effective work/life strategies: Working cou-
ples, work conditions, gender and life quality. Soc. Problems
47(3) 291–326.

Barley, Meyerson, and Grodal: E-mail as a Source and Symbol of Stress
906 Organization Science 22(4), pp. 887–906, © 2011 INFORMS
Mortensen, M., A. W. Woolley, M. B. O’Leary. 2007. Conditions
enabling effective multiple team membership. K. Crowston,
S. Sieber, E. Wynn, eds. Virtuality and Virtualization. IFIP
International Federation for Information Processing, Vol. 236.
Springer, Boston, 215–228.
Murray, W. C., A. Rostis. 2007. Who’s running the machine? A the-
oretical exploration of work, stress and burnout of technologi-
cally tethered workers. J. Individual Employment Rights 12(3)
249–263.
Nippert-Eng, C. E. 1996. Home and Work: Negotiating Boundaries
Through Everyday Life. University of Chicago Press, Chicago.
O’Leary, M. B., A. W. Woolley, M. Mortensen. 2008. Multiple team
membership: Productivity and learning effects for individuals,
teams, and organizations. Working paper, Sloan School of Man-
agement, MIT, Cambridge, MA.
Ophir, E., C. Nass, A. D. Wagner. 2009. Cognitive control in media
multitaskers. Proc. Natl. Acad. Sci. USA 106(37) 15583–15587.
Orlikowski, W. J. 1992. The duality of technology: Rethinking
the concept of technology in organizations. Organ. Sci. 3(3)
398–427.
Orlikowski, W. J. 1996. Improvising organizational transformation
over time: A situated change perspective. Inform. Systems Res.
7(1) 63–92.
Orlikowski, W. J. 2000. Using technology and constituting structures:
A practice lens for studying technology in organizations. Organ.
Sci. 11(4) 404–428.
Orlikowski, W. J. 2007. Sociomaterial practices: Exploring technol-
ogy at work. Organ. Stud. 28(9) 1435–1448.
Orlikowski, W. J., D. C. Gash. 1994. Technological frames: Making
sense of information technology in organizations. ACM Trans.
Inform. Systems 12(2) 174–207.
Orlikowski, W. J., S. V. Scott. 2008. Sociomateriality: Challenging the
separation of technology, work and organization. Acad. Man-
agement Ann. 2(1) 433–474.
Orlowski, A. 2005. Email destroys the mind faster than marijuana—
Study. Register (April 22), http://www.theregister.co.uk/2005/04/
22/email_destroys_iq/.
Pasmore, W. 1988. Designing Effective Organizations: Sociotechnical
Systems Perspective. John Wiley & Sons, New York.
Phillips, J. G., L. Reddie. 2007. Decisional style and self-reported
email use in the workplace. Comput. Human Behav. 23(5)
2414–2428.
Poole, M. S., G. DeSanctis. 1990. Understanding the use of group
decision support systems: The theory of adaptive structuration.
J. Fulk, C. W. Steinfield, eds. Organizations and Communication
Technology. Sage, Newbury Park, CA, 173–193.
Reinsch, N. L., Jr., J. W. Turner, C. H. Tinsley. 2008. Multicommu-
nicating: A practice whose time has come? Acad. Management
Rev. 33(2) 391–403.
Renaud, K., J. Ramsay, M. Hair. 2006. “You’ve got email!” 000Shall
I deal with it now? Electronic mail from the recipient’s perspec-
tive. Internat. J. Human-Comput. Interaction 21(3) 313–332.
Rice, A. K. 1963. Productivity and Social Organization: The Ahmede-
bad Experiment. Tavistock, London.
Rice, R. E., C. Aydin. 1991. Attitudes toward new organizational
technology: Network proximity as a mechanism for social infor-
mation processing. Admin. Sci. Quart. 36(2) 219–244.
Richtel, M. 2003. The lure of data: Is it addictive? New York Times
(July 6), http://www.nytimes.com/2003/07/06/business/the-lure
-of-data-is-it-addictive.html.
Robinson, J. P., G. Godbey. 1997. Time for Life: The Surprising Ways
Americans Use Their Time. Pennsylvania State University, Uni-
versity Park.
Schor, J. B. 1993. The Overworked American: The Unexpected
Decline of Leisure. Basic Books, New York.
Sparks, K., C. L. Cooper, Y. Fried, A. Shirom. 1997. The effects of
hours of work on health: A meta-analytic review. J. Occupa-
tional Organ. Psych. 70(4) 391–408.
Sproull, L. 2000. Computers in U.S. households since 1977. A. D.
Chandler, J. W. Cortada, eds. A Nation Transformed by Infor-
mation. Oxford University Press, London, 257–280.
Strauss, A. C., J. Corbin. 1990. Basics of Qualitative Research:
Grounded Theory Procedures and Techniques. Sage, Newbury
Park, CA.
Stross, R. 2008. Struggling to evade the e-mail tsunami. New
York Times (April 20), http://www.nytimes.com/2008/04/20/
technology/20digi.html.
Thomas, G. F., C. L. King, B. Baroni, L. Cook, M. Keitelman,
S. Miller, A. Wardle. 2006. Reconceptualizing e-mail overload.
J. Bus. Tech. Comm. 20(3) 252–287.
Thomée, S., M. Eklöf, E. Gustafsson, R. Nilsson, M. Hagberg.
2007. Prevalence of perceived stress, symptoms of depres-
sion and sleep disturbances in relation to information and
communication technology (ICT) use among young adults–An
explorative prospective study. Comput. Human Behav. 23(3)
1300–1321.
Towers, I., L. E. Duxbury, C. Higgins, J. Thomas. 2006. Time thieves
and space invaders: Technology, work and the organization.
J. Organ. Change Management 19(5) 593–618.
Trist, E. L., K. W. Bamforth. 1951. Some social and psychological
consequences of the longwall method of coal-getting. Human
Relations 4(1) 3–38.
Valcour, P. M., L. W. Hunter. 2005. Technology, organizations and
work-life integration. E. E. Kossek, S. J. Lambert, eds. Work
and Life Integration: Organizational, Cultural, and Individ-
ual Perspectives. Lawrence Erlbaum Associates, Mahwah, NJ,
61–84.
Whittaker, S., C. Sidner. 1996. Email overload: Exploring personal
information management of email. Human Factors Comput. Sys-
tems: Proc. (CHI’96). ACM Press, New York, 276–283.
Zuboff, S. 1988. In the Age of the Smart Machine. Basic, New York.
Stephen R. Barley is a professor of management science
and engineering and the codirector of the Center for Work,
Technology and Organization at Stanford’s School of Engi-
neering. He received his Ph.D. in organization studies from
the Massachusetts Institute of Technology. His research inter-
ests include work and occupations, technological change, and
corporate power in representative democracies.
Debra E. Meyerson is an associate professor of education
and (by courtesy) organizational behavior at Stanford Univer-
sity. She received her Ph.D. in organizational behavior from
Stanford University. Her research focuses on technologies and
the construction of work-life boundaries, the construction and
disruption of gender bias in organizations, change agency
within traditional institutions, and philanthropic foundations as
social change agents.
Stine Grodal is an assistant professor of strategy and inno-
vation at Boston University. She received her Ph.D. in man-
agement science and engineering from Stanford University.
Her research investigates the use of symbolic resources and
technology within nascent fields.