Content uploaded by Armand V. Cardello
Author content
All content in this area was uploaded by Armand V. Cardello on Oct 15, 2016
Content may be subject to copyright.
Content uploaded by Armand V. Cardello
Author content
All content in this area was uploaded by Armand V. Cardello on Oct 15, 2016
Content may be subject to copyright.
“A Pace Not Dictated by Electrons”:
An Empirical Study of Work Without Email
Gloria J. Mark1, Stephen Voida1
Department of Informatics
University of California, Irvine
Irvine, CA 92697-3440 USA
{gmark, svoida}@uci.edu
Armand V. Cardello
U.S. Army Natick Soldier R, D & E Center
Natick, MA 01760 USA
armand.cardello@us.army.mil
ABSTRACT
We report on an empirical study where we cut off email
usage for five workdays for 13 information workers in an
organization. We employed both quantitative measures
such as computer log data and ethnographic methods to
compare a baseline condition (normal email usage) with our
experimental manipulation (email cutoff). Our results show
that without email, people multitasked less and had a longer
task focus, as measured by a lower frequency of shifting
between windows and a longer duration of time spent
working in each computer window. Further, we directly
measured stress using wearable heart rate monitors and
found that stress, as measured by heart rate variability, was
lower without email. Interview data were consistent with
our quantitative measures, as participants reported being
able to focus more on their tasks. We discuss the
implications for managing email better in organizations.
Author Keywords
Email; multitasking; interruptions; empirical study; sensors
ACM Classification Keywords
H.5.3 [Information Interfaces and Presentation (e.g., HCI)]:
Group and Organization Interfaces; K.4.m [Computers and
Society]: Miscellaneous.
INTRODUCTION
Less than a year ago, a New York Times article drew an
analogy between zombies and emails: you keep killing
them (or deleting them), and they never stop coming [20].
This tongue-in-cheek commentary brings to light a broader
issue associated with being continually connected to ICTs
and email in particular—we just can’t seem to keep up. The
wide adoption of ICTs in the workplace enables people to
stay continually up-to-date and connected with others, but
critiques as in the New York Times article raise important
questions about how continual connectivity might impact
aspects of our psyche and behavior such as creating stress
and distraction.
Does being continually connected provide benefits to
information workers? Wacjman and Rose [29] claim that
being continually connected gives agency to workers;
information workers are constantly negotiating how to
manage information, when to respond to communications,
and how to prioritize tasks and information. Email is
currently considered the tool de rigueur of the workplace,
cf. [3]. However, being continually connected on email has
also drawn criticism, expressed by Turkle, who writes “we
don’t do email, our email does us” [28]. Bolstering this
view is a result from a study of email use where 45% of
participants associated email with a loss of control [2].
Though we might surmise that we are overburdened with a
huge volume of email, how much attention does email
demand? Wacjman and Rose [29] found that, on average,
information workers engage in more mediated
communication each day than face-to-face communication,
of which email is the most common. The picture that
Jackson et al. [19] found is rather dismal: they found that
70% of emails were attended to within six seconds of
arriving. Email may not be distracting if workers quickly
returned to their interrupted task; however, it took an
average of 64 seconds to resume an interrupted task.
Email is a multi-functional tool in the workplace; not only
is it used for communication, but it also serves as a to-do
list and personal information management tool [3], for
archiving information, as a mechanism to foster
coordination and collaboration among colleagues, and as a
source for assigning and delegating tasks [30]. Barley et al.
[2] summarize this multi-functional character of email by
claiming it to be a symbol for work.
In this paper, we address the issues of email and
distractions. Whereas email use has been studied through in
situ observation [3, 25] and surveys [10], we approach
email usage from an inverted perspective. We investigate
how people perform work in the absence of email—a world
free of the “zombies” that vex so many in information
work.
1 Authors Mark and Voida contributed equally to this work and
share primary authorship.
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. To copy otherwise,
or republish, to post on servers or to redistribute to lists, requires prior
specific permission and/or a fee.
CHI’12, May 5–10, 2012, Austin, Texas, USA.
Copyright 2012 ACM 978-1-4503-1015-4/12/05...$10.00.
EMAIL AND MULTITASKING
Studies of information workers have consistently revealed
that they experience high levels of multitasking and
interruptions [8, 12, 21]. Interruptions can be beneficial, for
example, in gaining relevant information, addressing the
task at-hand [21], or as a way to exert control over work by
choosing when to respond to others [29]. However, some
argue that interruptions can have a negative impact on work
when they cause people to switch contexts or create
redundant work [21], if they occur at inappropriate times
[8], or if they lead users to forget their main task focus [9].
Evidence suggests that fragmented work patterns negatively
impact work productivity [24].
Observation of information workers reveals that they switch
tasks on average every three minutes [12]. Much switching
concerns email: studies show with consistency that people
spend about 23% of their time on email [8, 25], with an
estimate that people check email about 36 times an hour
[25]. This could be an overestimation, but even if halved,
and if true, it would comprise a large component of
multitasking. Further, people self-report that a high amount
of email is associated with a high level of stress [2, 10].
Different theories have been proposed to explain how email
creates overload for people. Managing the sheer volume of
email is one factor [29], but the more time spent on email,
the more likely it creates a feeling of overload [2]. Poor
email management strategies also play a role [10, 29], as
does the effort to keep track of separate email threads [3].
Recipients generally need to meet the task demands of the
sender, which can exacerbate this sense of overload [25].
Despite the wide attention that email overload has received,
it is still unclear to what extent email actually contributes to
multitasking. Does email lead people to focus less on the
task at-hand? In this paper, we address the following
research questions related to email and multitasking:
• How is multitasking affected in the absence of email? Is
multitasking fueled by email, or would multitasking
occur at the same rate in today’s information work
climate without email use? Does email lead to the
increased fragmentation of work?
• Though stress has been self-reported to correspond to
email use, to date no study has ever directly measured the
effect that email has on stress. Stress associated with
email has been based on self-reports through surveys or
logging [2, 10, 25]. Yet there may be biases in self-
reports of stress—for example, in overestimating it. We
therefore ask the question: how much stress does email
use actually cause? We measure stress directly during in
situ work using mobile wearable sensors.
There are several reasons why it is important to take an
inverted perspective and investigate how work might
change when email is removed. First, it enables us to
understand whether it is possible to create an environment
in which people can focus more closely on their tasks.
Since people self-interrupt to check email often [19, 21,
29], would they be able to spend longer durations on tasks
when email is not available and when keeping up-to-date
with incoming email is no longer a concern? Ultimately, we
are interested in providing support to avoid work
fragmentation, given the negative consequences associated
with this problem [24]. Second, we are interested in
learning whether—and how much—stress might be reduced
without email. While it is not realistic for an organization to
eliminate email usage, this study can shed light on how
email management and organizational email policies might
be improved.
Our method to directly address the role of email in
multitasking was to design an intervention in which
multitasking is observed both with and without email. It is
only by comparing a baseline condition (the status quo, i.e.,
typical email usage) with an experimental manipulation (the
absence of email) that we can directly examine the effects
email plays on multitasking and stress in the workplace.
RESEARCH SETTING
The study was conducted in a large scientific research
organization located on the east coast of the United States.
For the current study, participants were recruited by
inviting them to information sessions. A total of 13
participants volunteered for the study (6 females, 7 males,
mean age = 46). Additionally, 36 of their colleagues (13
females, 23 males) in their workplace social network
volunteered to provide end-of-day surveys during the study.
Our study was approved by institutional review boards at
our home institutions and at the field site.
All of our participants were information workers whose job
titles included chemical engineer, materials scientist,
psychologist, biologist, food technologist, and research
administrator; four of the 13 worked as managers or
mentors with significant supervisory responsibilities. Their
self-reported job characteristics revealed, on average (using
a 5-point scale, where 5 was high), relatively high levels of
creativity (M = 4.5, SD = 0.52), autonomy (M = 4.3,
SD = 0.9), and day-to-day task variety (M = 4.4, SD = 0.5).
All participants had some level of postsecondary education,
and averaged employment in their current job for 8.4 years.
Methodology
Our study was a within-subjects design, with each
participant taking part in a three-day baseline data
collection phase and a five-day experimental (“No Email”)
condition. In both of these conditions, we collected data
using a combination of ethnographic methods and
automatic, computer log and sensor-based data collection.
On the first day of the study, we conducted an initial semi-
structured interview with each participant to ask about his
or her existing multitasking and email usage strategies.
After installing a custom window activity logging
application on his or her computer and providing instruction
about how to use the heart rate monitor, a researcher spent
the rest of the day conducting an in-person ethnographic
observation of a typical, “baseline” workday. We also
began recording activity in the office using our logging
software to capture shifts in document and application
windows as a measure of how often people were
multitasking in their electronic work, a process that we
repeated for two additional days (and sometimes more), to
ensure that our system collected several days’ worth of
baseline data.
On the fourth day of the study, participants experienced the
No Email experimental condition—complete email cutoff
for a five-day period. During these days, we created a
Microsoft Outlook email-processing rule that would
automatically file away incoming messages without
triggering any pop-up notifications. We also asked the
participants to voluntarily refrain from sending work-
related email. However, because so many participants used
other facets of the Outlook software (e.g., contacts, the
calendar, SharePoint shared folders), we allowed
participants to use Outlook and access emails that had been
received prior to the onset of the experimental condition.
We recorded activity in the office for all five of the No
Email days using the window activity logging software. A
researcher returned on the final day of the experimental
condition to conduct a second day of ethnographic
observation. At the conclusion of the study, we conducted a
semi-structured interview in which we asked about the
participant’s activity management strategies and about their
experiences of not having email available. Post-study
interviews lasted about one hour.
Our use of ethnographic methods was intended to provide a
rich corpus of data for understanding the complex actions in
the workplace. To understand multitasking behavior in situ,
we needed to capture details of process and content of our
participants’ workplace activities. An alternative technique,
videotaping our participants, was not permitted by our field
site. Thus, we used a “shadowing” observation technique
similar to that used in previous time management and
multitasking studies [12, 21, 22]. A researcher sat with the
participant in his/her workspace and followed the
individual, whenever possible, to meetings or other
activities. The researcher recorded every action that the
participant performed to the second, such as opening a
computer application or making a phone call. Details about
each event were also noted. All interactions with colleagues
were documented, including details about the conversation
topic, documents used (if possible) and persons involved.
We conducted this in-depth observation for two days of the
study—the first day of the Baseline condition and the last
day of the No Email condition.
Our custom window activity logging application was
designed to collect real-time data about our participants’
multitasking behavior. This application was installed on
each participant’s computer workstation before we began
data collection, and it ran for the entire duration of the
study. The software recorded the frequency of switches
between document or program windows where electronic
work is carried out (after [18, 23]). In order to gather more
objective measures of stress and arousal, the participants
wore an electronic heart rate monitor while working—the
Garmin ForeRunner 301, consisting of a chest strap and
wristwatch receiver—for the duration of the study, i.e., in
both the Baseline and No Email conditions.
We administered survey instruments at the beginning and
end of the study, as well as at the end of each day. Pre-
study survey questions included a demographic
questionnaire and a polychronicity preference instrument
(IPV) [6], which measures the extent to which participants
prefer to work in a polychronic or monochronic manner. At
the end of each workday, we delivered several surveys
electronically from our data collection software, including
the NASA TLX workload scale [15].
Finally, in order to assess the impact of disconnecting a
colleague from email, we asked members of the
participants’ workplace teams to serve as peripheral
participants in the study. We asked these team members to
complete short surveys at the end of each day about their
own perceived productivity, their ability to get information
needed to accomplish work on that particular day, and their
success (or failure) in getting information that they needed
from the main participants, given the changes in their
accessibility via email.
RESULTS
Over the course of the study, we conducted nearly 137
hours of ethnographic observation. Our automated sensors
captured over 25,000 window changes and recorded over
1.6 million sensor events over a total of more than 700
hours of data collection. The heart rate monitors provided
readings of heart rate every few seconds; we received over
40,000 of these readings during the course of the study.
Sometimes illness, vacations, and technology failures
interfered with our study schedule. When these incidents
occurred, we rescheduled our informants so that they
experienced email cut off for a period of five consecutive
days. In some cases, this led to our extending Baseline or
No Email data collections beyond the typical 8-day
window.
We asked participants at the end of each day of the study if
they considered that day to be “normal” in terms of the
volume and type of work they did. We found no statistical
difference for participants in the number of days judged to
be normal between the Baseline and No Email conditions.
Approximately 65% of study days were reported as being
normal workdays.
Workplace Activity: Shadowing of Participants
We observed and classified a total of 5,643 activities
carried out by our 13 participants over the course of our in-
person shadowing sessions, comprising 71 hours of data
collected in the Baseline condition and nearly 66 hours in
the No Email condition.
In general, participants reacted well to being cut off from
email. Typically, participants stayed connected with their
colleagues during the No Email condition through a
combination of face-to-face meetings, telephone calls, and
by having team members notify them (mostly face-to-face)
of critical emails sent to the entire work group or
laboratory. Although participants checked email habitually
in the Baseline condition, by the fifth day of email cutoff,
we observed that participants rarely tried to check email.
The actions recorded from shadowing were categorized into
categories of activities, as shown in Table 1. A three-way
ANOVA of the duration of observed activities with
condition, participant, and activity type as factors reveals a
significant main effect of condition, F(1, 12) = 6.881,
p < .05, indicating that, overall, participants spent longer in
activities on average in the No Email condition than they
did in the Baseline condition. We also found a significant
main effect of activity type, F(6, 76) = 14.301, p < .001,
and a significant interaction of condition by activity type,
F(6, 69) = 8.753, p < .001, suggesting that our participants
allocated their time differently across different kinds of
activities when access to email was limited. We also
observed significant interactions of condition by
participant, F(11, 95) = 3.432, p < .001, and condition by
activity type by participant, F(60, 5474) = 4.849, p < .001,
suggesting that there were individual differences in how
participants reacted to changes in email availability and in
their strategies for multitasking in each of the conditions.
In order to better understand the effect of cutting off email
on activities, we ran a series of paired t-tests on the mean
durations that participants spent working in each type of
activity. We found significant differences in just two of the
activity types. First, participants spent significantly longer
on out-of-office activities in the No Email condition
compared to the Baseline condition: t(12) = −3.001, p < .05.
Second, we found that participants spent significantly more
time carrying out metawork (i.e., activities not connected
with any single project but associated with managing all of
them [12]) in the No Email condition than they did in the
Baseline condition: t(12) = −2.558, p < .05. This difference
might be explained by the fact that coordination work, often
carried out via email, is frequently interleaved with
metawork activities; without email, this coordination work
may be deferred, leading to larger contiguous blocks of
metawork.
Surprisingly, the duration of time spent in each session
reading, writing, or reviewing attachments to emails—
either newly-arrived or those archived in the participants’
stored folders—did not differ significantly between the two
conditions. Although this is a counterintuitive finding since
email was cut off, a potential explanation can be found in
the data we collected about the frequency (i.e., the total
number of activities observed) in which participants
engaged in each category of activity in the different
conditions. A three-way ANOVA of the frequency of
switching observed activities yielded a significant main
effect of activity type, F(7, 73) = 28.526, p < .001, as well
as significant interactions of condition by activity type,
F(6, 60) = 2.412, p < .05, and condition by participant,
F(11, 60) = 4.160, p < .001. We also ran paired t-tests on
these data to determine whether there were any differences
in the frequencies at which participants switched activity
types across the two conditions. The only activity type that
showed statistically significant differences was the email
activity; participants worked in their email client
significantly less frequently in the No Email condition
compared to the Baseline condition, t(11) = 3.964, p < .01.
Thus, though participants used their email client less often
when email was cut off, we found no difference between
the amount of time they spent using Outlook in either
condition (e.g., in the No Email condition, they read emails
received prior to the study).
Multitasking Measures: Window Switches
Next, we looked at the sensor data that were collected from
tracking computer window switches over the entire 8-day
duration of the study. To measure differences in
multitasking between the Baseline and No Email
conditions, we looked at the duration and frequency of
window switches. Table 2 shows, for each participant, the
means and standard deviations of the durations (in seconds)
that application and document windows were left open, as
well as the frequency of window switches (in switches per
hour) during each hour that our sensors collected data. We
counted all window switches, including when auxiliary
windows were invoked (e.g., reading a PDF attachment
from a past email). Seven extreme outliers were removed
from our set of over 25,000 data points.
For all but one participant, the mean window duration was
longer during the No Email condition compared to
Baseline. A paired t-test comparing each participant’s
Baseline
No Email
Activity Type
Mean
SD
Mean
SD
Out of office*
(work-related)
412.32
938.14
1195.24
2048.88
Computing tasks
52.47
90.52
50.32
82.81
Physical tasks
(e.g., reading,
jotting notes)
41.06
75.55
56.55
69.86
Communication
in the office
(excl. email)
84.82
256.94
60.15
103.68
Email (new/filed)
40.65
60.71
36.94
64.85
Metawork*
21.41
28.04
29.14
41.76
Other tasks
56.12
205.58
31.33
54.50
Overall
74.81
375.37
102.85
510.81
Table 1. Means and SDs of the observed durations (in seconds)
of types of activities, excluding personal breaks.
* = sig. difference between Baseline and No Email at p < .05.
window duration during the Baseline and No Email
conditions showed that participants spent highly
significantly longer times in a document or program
window in the No Email condition: t(12) = −5.00, p < .001.
Combining participants’ data, an ANOVA with condition
and participant as factors shows a highly significant
difference between conditions: F(1, 14) = 28.01, p < .0001.
There was no significant difference between participants:
F(12, 12) = 2.37, p > .05. There was, however, a significant
condition by participant interaction: F(12, 25094) = 3.18,
p < .0001. Thus, the data show that there was a significant
change in the length of time that participants’ windows
were active between conditions: in the No Email condition,
participants had individual application windows open for a
significantly longer duration before switching to another
window, compared to those in the Baseline condition.
We also examined the frequency of window switches.
Table 2 shows that all participants had a lower mean
frequency of window switches in the No Email condition. A
paired t-test comparing participants’ frequencies in Baseline
to No Email showed a highly significant difference:
t(12) = 6.50, p < .0001. Thus, participants in the No Email
condition switched their windows significantly less often
than in the Baseline condition. Combining participants’
data, an ANOVA showed a highly significant difference
between Baseline and No Email: F(1, 13) = 44.23,
p < .0001, with a significant difference between
participants: F(12, 12) =3.14, p < .05. There was also a
significant condition by participant interaction:
F(12, 903) = 1.97, p < .05. Thus, overall, the window
logging data show that participants in the No Email
condition had individual application and document
windows active for longer durations and switched between
them less often than they did in the Baseline condition,
when email was available.
Email and Stress: Heart Rate Monitoring
We also directly measured participants’ level of stress when
working with and without email through the use of heart
rate monitors. Participants wore the heart rate monitors full
time while at work. Heart rate variability (HRV) is widely
used as an indicator of mental stress (see [1, 4, 5] for
reviews). The lower the measure of HRV, the higher the
amount of stress that an individual experiences. The
correlation between stress and HRV is high; as arousal
increases, HRV decreases. The body responds to stressful
circumstances by regulating itself: the sympathetic nervous
system is very good at responding to stress. Similarly, when
people are not experiencing stress, HRV is higher, as the
heart rate fluctuates more, i.e., the body is not regulating
itself. A lowering of HRV has been associated with other
factors related to stress (e.g., depression [14] and anxiety
[31]). HRV has been used to measure a variety of
phenomena, including stress during computer work [17].
A recommended measure of overall HRV is the standard
deviation of heart periods [5]. Our goal was to measure
study participants’ stress during the Baseline and No Email
conditions and we did so by measuring HRV continually as
participants went about their normal work. Due to technical
difficulties, we were only able to capture data across both
the Baseline and No Email conditions from seven of our
participants. Table 3 shows the number of days for these
participants in which heart rate data were captured. In post-
study interviews, no one reported that wearing the heart rate
monitor was obtrusive or that it affected their behavior.
Nearly all participants reported that once they put the chest
strap on, they forgot about it. One participant wore the
monitor home once because she forgot that she was wearing
it. Another participant reported that he tended to “sit back
more” but did not believe it changed his behavior. A third
person reported that it was “a bit annoying,” but she wore it
on all of the study days except one.
We computed the HRV using the standard deviation of the
recorded heart rate, which is the standard way to measure
HRV [5]. Table 3 shows measures averaged for each
participant and for each condition over all the days that data
were collected. A paired comparisons t-test between the
Baseline and No Email conditions shows a very strong
trend that people in Baseline had lower HRV (i.e., higher
Baseline
No email
Subj.
# days
Mean
SD
# days
Mean
SD
2
6
76.35
8.43
5
79.29
10.60
3
4
88.99
9.82
3
93.70
9.85
4
3
71.31
7.06
3
76.07
5.36
5
5
74.92
23.72
5
75.61
25.93
6
3
95.88
18.80
8
100.00
23.39
8
2
77.15
8.75
7
78.99
13.03
11
3
62.26
9.40
4
67.22
18.22
Total
26
77.03
16.173
35
80.39
18.36
Table 3. Mean and SD of heart rate monitor data and
number of days that data were captured for each condition.
Baseline
No Email
Duration
Frequency
Duration
Frequency
P
M
SD
M
SD
M
SD
M
SD
1
76.4
466.7
36.2
34.3
106.1
621.6
12.7
21.2
2
72.6
329.7
27.2
26.0
214.0
729.9
9.1
14.1
3
79.4
450.1
39.2
33.5
121.9
536.2
25.4
20.8
4
53.2
166.9
53.9
34.5
109.8
322.7
24.6
17.7
5
133.8
554.8
17.5
18.5
216.6
882.3
8.2
13.1
6
55.3
275.9
44.2
32.7
83.4
285.0
24.1
23.3
7
67.4
397.9
45.1
33.1
115.7
596.9
25.8
28.1
8
100.6
402.1
26.7
23.0
156.8
821.5
16.2
41.5
9
78.8
369.9
40.1
26.1
181.6
658.1
9.0
12.2
10
53.8
181.0
47.5
29.5
198.4
828.1
16.2
25.2
11
130.5
421.9
23.4
18.8
180.7
560.5
17.5
17.1
12
99.4
400.7
25.5
24.1
126.8
441.9
24.1
17.2
13
62.3
438.2
42.7
33.3
56.7
156.4
30.6
32.0
M
75.5
394.3
37.1
31.4
131.9
568.1
18.2
23.5
Table 2. Mean and SD of window duration (in seconds)
and frequency of window switches (switches/hour in
which data were collected) for each participant.
stress) than those in No Email (Mean difference = −2.926,
SD = 3.425, t(6) = −2.260, p = .065). Levene’s test
conducted on the comparison of the standard deviations of
all Baseline data with all No Email data shows this
difference to be highly significant: SD(Baseline) = 16.17,
SD(No Email) = 18.36, F(1, 40409) = 170.86, p < .0001.
These results indicate that participants experienced less
stress when their email was cut off than in the baseline
condition when they had email access.
An ANOVA conducted on the mean heart rate with
Condition and Participant as factors shows that the measure
was significantly less in Baseline than in the No Email
condition: F (1, 40409) = 33.40, p < .001. There was also a
significant difference between participants,
F(6, 40409) = 206.34, p < .0001, and a significant
interaction of condition by participant, F(6, 40409) = 16.30,
p < .0001. Thus, mean heart rate was higher in the No
Email condition. While there could be a number of
explanations for this result, heart rate generally increases
with activity. This notion is consistent with our
observations, which showed that without email, participants
engaged in more “Out of office” activities (see Table 1).
Attitudes of Multitasking and Email: Surveys
Surveys were deployed to help us understand whether
immersion in an “email free” work environment might be
affected by—or influence—participants’ attitudes about
multitasking and feelings of productivity, time pressure,
and frustration. We asked participants at the beginning of
the study questions drawn from the Inventory of
Polychronic Values (IPV), an instrument designed to assess
the extent to which people in a culture: (1) prefer to be
engaged in two or more tasks simultaneously; and (2)
believe that their preference is the best way to do things [6].
Relative to the results reported by Bluedorn et al. [6], four
out of 13 participants were categorized as having a high
preference for polychronicity and nine participants
measured an average preference for polychronicity. There
was no significant relationship between polychronicity
preference scores and mean window durations in either the
Baseline or No Email conditions.
At the end of each day of the study, participants completed
the NASA TLX workload assessment [15], which measures
mental, physical, and temporal demands, as well as
performance, effort, and frustration. An ANOVA conducted
on participants’ responses between the Baseline and No
Email conditions showed no significant differences. These
findings suggest that foregoing the use of email at work
does not fundamentally change the subjective workload that
a person experiences—in either direction.
Colleagues’ Perspectives: End-of-Day Self-Reports
A person’s availability via email has the potential to impact
not only their own multitasking and stress levels, but also to
affect the ability of their colleagues to find information they
need to collaborate and to delegate tasks. For each of our
study participants, we recruited a group of 2–7 of that
person’s closest team members to answer a short series of
survey questions at the end of each day of the study. These
questions combined general queries about the success and
stress level of the colleagues’ day (e.g., “Rate the following
statement: I was able to get the information I needed to
conduct my work today” and “How productive do you think
you were today?”) and a question specifically geared to
determining the impact of the main participant’s availability
or unavailability via email (“Rate the following statement:
It was easy for me to reach [name] to get information I
needed from [him/her]”).
An ANOVA showed no significant difference in team
members’ responses between the Baseline and No Email
conditions for these questions. However, a trend
approaching statistical significance revealed that team
members agreed more strongly with the statement “I was
able to get the information I needed to conduct my work
today” in the Baseline condition (M = 4.53, SD = 0.28) than
in the No Email condition (M = 3.71, SD = 0.26), F(1, 47) =
3.221, p < .08.
QUALITATIVE DATA ANALYSIS
How did our informants feel when their email was cut off
for five days? The post-study interview data was open-
coded and themes were identified. Not surprisingly, nearly
all informants viewed email as “double-edged,” having pros
and cons. Attitudes towards email could be categorized in
two ways. To two informants, they viewed email primarily
as a communication tool in their work—these people
missed email the most when it was cut off. The second
category—the rest of the informants—expressed an attitude
that email was primarily a burden. A common reason
expressed by this latter group was due to the volume, e.g.:
My work has become how to manage email.
I have so many emails, I don’t even read them.
With email, it’s a train wreck… you can’t look away from it.
One informant’s attitude was extreme:
[Email is] an annoyance, a giant to-do list. I got to the
point where, being overwhelmed, I gave up…. That would
be nirvana—to tell everyone I’m wiping out my email….
It’s ruining my life, it’s interfering with my happiness.
Only one informant considered herself addicted to email.
She explained that she takes her BlackBerry on vacation but
then locks it in the hotel safe to keep from checking email.
One informant explained she was not addicted but would
not want to live without it. Still another admitted that at
times he felt addicted to email.
Lack of Agency
Another reason why informants viewed email as a burden
was that they felt that it affected their agency to work.
About half of the informants described that they felt like
they were not in control of their email, and consequently
not in control of their work. Attitudes were expressed as:
It ruled my life—that made me feel depressed, and now I
feel liberated [without email]…too much headache
trying to keep on top of everything.
I let the sound of the bell and the pop-ups rule my life.
[When email was cut off]: Felt it was under my control.
It prompts me. When you come back, it’s waiting.
In contrast, when asked how they felt about working
without email, nearly all informants described that their
pace was more relaxed. One informant described he felt
“liberated”; another informant expressed the biggest benefit
as “peace”; still another described it as “refreshing.”
Email, Communication, and Relationships
A theme that emerged was that when email was cut off,
face-to-face interactions with other colleagues in the
workplace increased. All informants reported that with
email cut off, they interacted with people more, both face-
to-face and by phone, whereas they would usually send
emails instead. They viewed this change as a benefit.
Nearly all informants described how they walked around
the workplace more to visit their colleagues for face-to-face
interactions in the No Email condition; some even walked
to other buildings on the campus. Some visits to colleagues
were by choice, as opposed to placing a telephone call2.
The lack of access to email enabled the informants to reflect
more deeply on how email affects their relationships with
their colleagues. The informants expressed that during the
time of email cut off, they became aware that the use of
email hindered their work relationships. Some
representative comments that express this idea were:
…Email can be a superficial blanket that distances you from
real relationships where you’re really working together.
I socialize and the social aspect became that much greater.
I was surprised at how much all that human interaction
came in to fill the vacuum [i.e., when having no email].
[Working without email] helps with one-on-one
relationships.
Email is easier, but getting up and walking around, it’s a
lot easier to talk face-to-face. You can pick up more vibes
of the relationship thing that you can’t do with email.
Focus
Our quantitative results showed that in the No Email
condition, people switched tasks less and spent longer
durations on each window screen before switching. The
interview data corroborate this result. A common theme
that was expressed by nearly all informants was that during
the time that email was cut off, they were able to spend
longer periods of time on a task and focus more intently on
their work. One informant described that during the study,
2 In our shadowing observation, we could not always follow
informants when they left the office; the activity of interacting with
others outside the office is characterized as “Out of office” in Table 1.
he learned that email was his main source of self-
distraction. The word “focus” was found often throughout
our interview data. Typical comments included:
It gave me more focus.
It gave me time to think about [work] more.
I was able to spend time actually doing tasks that had to
be done…. It was nice to be able to sit and work on a
manuscript for the whole morning.
When I didn’t have email, multitasking, I had three
projects done. I was more focused.
[I] wasn’t distracted by checking email. It was nice. I was
able to plan more what I was doing for a chunk of time.
I was writing a report, and I could focus on the report.
Only one person told us that she was surprised that, despite
expecting to be able to accomplish a larger number of
projects in the No Email condition, she did not.
Feeling Cut Off
The biggest disadvantage expressed by the informants when
they did not have email was that they felt “cut off.” Two
informants did not report feeling cut off (one reported that
he did not miss anyone); one informant felt cut off “a little
bit” but was “OK” with it. Only one informant mentioned
from whom she felt cut off (off-site customers). About half
the informants described it as a general sense of isolation:
You have that sense that there’s something really important,
there’s a deadline that you missed… not knowing.
The hardest thing was not being sure what I missed.
I felt disconnected for sure. When email is down, it makes
me crazy. I don’t feel as productive…it was difficult.
Yes—hands down—it isolates you as the one person
who’s not “plugged in.”
There was a little uneasiness…felt I was missing something.
I didn’t like knowing that information was coming or
not ‘cause there’s the expectation that you’re on email.
Thus, though most informants reported feeling cut off from
others, this feeling seemed to be grounded in a fear that
they were potentially missing out on information being
exchanged within the organization.
As a result of not using email in the workplace, the
informants could describe their actual experiences of how
email affects work, as opposed to their perceptions. The
interview data revealed that there were various ways that
email was reported to be burdensome. First, the social norm
that surrounds email use is primarily what causes it to be a
burden. The informants commonly expressed that there is a
norm or expectation that the email recipient will respond to
an email quickly. For example, one informant comes into
work two hours later than her colleagues. She described
how her colleagues, who arrived at work and sent out
emails two hours prior, expect an answer to their emails as
soon as she arrives. Second, email is a burden due to the
sheer volume of email that people receive. The informants
commonly reported not being able to keep up with their
email. They referred to important emails that get lost in the
flood of incoming messages. Third, the informants admit
that they lack self-control to not check email regularly. In
some cases, this lack of self-control translates into a lack of
agency in the workplace, as two informants described: a
“lack of control” and that it “rules their life.” Last, email is
a burden when it is used as a channel for delegating work.
Aside from the two informants who viewed email primarily
as a communication tool, to the other informants, it was a
channel for delegating tasks to others. One informant’s
experience is an illustration—this informant receives
“taskers” from his superiors. These refer to tasks that he
needs to do immediately. As a lab scientist, this interferes
with his ability to set up and run experiments without
interruption. When his email was cut off, the taskers
suddenly stopped. Though his superiors could have called
him on the phone or walked down the hall to delegate the
task to him, this did not happen. This experience led him to
believe that the taskers he had been receiving by email were
either not important or that the senders had taken initiative
to find the information themselves when he was off email.
DISCUSSION
We set out to discover whether email was a significant
component of multitasking behavior, and whether, by
removing email, people would focus more on their tasks
and multitask less. Our results show that without email,
people did multitask less and had a longer task focus, as
measured by a longer duration of time spent working in
each application window and a lower frequency of shifting
between windows. In other words, our data show that email
usage leads to more work fragmentation. Perhaps this is not
surprising, as other studies have shown how often people
check email [19, 25]. Our study confirms, that without
email, multitasking would be reduced. This raises a number
of other questions, such as: how much information sent
through email is actually necessary or desirable for work?
We also directly measured whether email causes stress. We
found a strong trend that stress, as measured by HRV, was
lower when email was removed. We expect that this result
would achieve significance with more subjects. If
workplace stress is detrimental to health, as some research
suggests [16], then our results may even suggest that email
could contribute to workplace health problems by raising
stress levels. This is a topic for further research.
It is possible that the out-of-office activity could have
affected HRV. Though HRV is reported higher in athletes
[7] and in older adults doing aerobic activity [27], there is
little data about the relationship between HRV and physical
activity or exercise in the general population. In our study,
any activity (e.g., walking to meetings) was a random effect
variable—it occurred in both conditions of our study, and is
therefore an unlikely contributor to the trend that we
observed between conditions.
It is difficult to disentangle the different sources of stress in
the workplace, so we can only speculate as to whether it is
the volume of email, social expectations and organizational
conventions associated with email, or the symbolism of
email as a delegator of work [2] that lead to the increase in
stress that we observed. It is noteworthy that we found
stress levels to change within a period of just five days
without email. This suggests that short “vacations” from
email may be sufficient to reduce stress in the workplace.
At the same time, the higher mean heart rate that we found
without email use is consistent with both our observations
and our interviews, which show that without email, people
take more frequent work breaks, during which they move
around the workplace more. This would suggest that not
only might reducing email reduce stress, but that reducing
use of email may even be good for health. Of course,
further research would be needed to examine this
hypothesis. Overall, our results suggest that the reduction of
email needs to be taken seriously by organizations and they
raise a number of potentially interesting issues for future
study.
Email and Non-mediated Interaction
One of the interesting findings that emerged from the
interview data was that not having email led to more face-
to-face interaction. On the whole, the informants reported
that they enjoyed their social life at work with their
colleagues more when email was cut off. Combined with
lower stress levels and the reports of being more relaxed,
this leads us to question more broadly how email is
affecting workplace relationships.
Interestingly, the colleagues in the participants’ work
groups did not report detrimental effects when their
colleagues were off email. Aside from reporting that it was
harder to get information from their colleague (which was
expected), their satisfaction, productivity, and stress levels
were not affected. As a result, we might hypothesize that
when a person upon whom one relies for information
becomes unavailable for a period of time, this will increase
the self-initiative of the information seeker to find the
information or to solve the problem on their own.
The Pace of Work
Our findings suggest that email speeds up the pace of work.
Participants switched between windows more often, and the
amount of time they spent in any one window before
moving on was shorter with email. At the same time, from
the interviews, nearly all informants reported that email
creates expectations that the receiver will respond quickly.
When we combine our quantitative and qualitative results,
we see support for the idea that email drives a cycle of
working at a faster pace. We cannot say whether this is
positive or negative for organizations; our study results do
not address whether a faster work pace is associated with
higher productivity or efficiency. What we can say from our
data is that the study participants overall preferred a slower
pace of work based on their reactions to work life without
email. Perhaps this is best expressed by the informant who
described not having email as allowing him to work at “a
pace not dictated by electrons.”
Implications for Organizations
Our results suggest several implications for organizations in
alleviating the burden of email on employees. First, our
results reinforce some existing recommendations that
reading email in batches might be efficacious [26]. Our
informants were off of email for five days. In the
interviews, most of the informants described that they
dreaded reading their email when they concluded
participation in the study. However, the process of catching
up was not as bad as they anticipated. One informant, for
example, said that when he returned to email, he read
through his inbox quickly and learned a strategy of
eliminating emails based on their subject header. He
claimed that this took far less time than handling emails by
checking them regularly throughout the day. This finding
illustrates a contrast to the survey responses reported by
Dabbish and Kraut [10]; our informants reported feeling
less overloaded with restricted email use. These differences
might be explained by the evolution of email management
strategies over the last several years, the specific culture of
email usage in the organization we studied, or the fact that
we were examining email usage in the broader context of
multitasking and information work. While we do not
suggest that reading email should be done in batches of
several days, organizations could consider experimenting
with establishing practices of exchanging email—at least
within particular work groups—at certain intervals, such as
first thing in the morning, after lunch, and in the evening.
Organizations could also try deferring informational emails
to a pull-oriented channel (e.g., an intranet website or RSS
feed), rather than sending email “blasts” that necessitate
action by every employee.
One of our intriguing results was that without email, our
informants consistently reported feeling relaxed and more
focused. This suggests that an organization should consider
the immaterial benefits for workers in restricting, filtering,
or actively managing the delivery of email.
Are we suggesting that organizations cut off email?
Absolutely not. We did learn a great deal from this study, as
did our participants. For example, informants reported that
when someone needed to reach them and email was not
available, they often used an alternative means. Perhaps if
communication is dispersed within an organization across
various media, e.g., IM, telephone, or face-to-face, it would
create less email traffic and may lower stress. Of course,
dispersed communication channels lead to other problems
such as not knowing which means to use to reach someone.
This is a topic of ongoing research.
Limitations
Although we studied a relatively small group of 13
participants (and surveyed their immediate colleagues), we
do not feel that this factor adversely affected the validity of
our findings, for two main reasons. First, this participant
size is consistent with previous, observation-only studies of
workplace multitasking (cf. [12]). Second, our combination
of qualitative and quantitative data collection techniques
provided an enormous amount of data from each participant
over almost two weeks of authentic, in situ work.
One of the main limitations of the study is that we collected
data at only a single field site, a decision that we made
because of these workers’ willingness to commit to a
demanding study schedule, including observation and the
use of heart rate recording devices, and the support that we
received from the field site administrators in allowing us to
ask employees to forego use of email for an entire week.
Even within this organization, we may have unintentionally
encountered some bias due to participant self-selection,
since only a portion of those people who voluntarily
attended a study information session consented to be in the
study. Thus, our participants may have self-selected on the
basis of their attitudes toward their email workload, concern
with multitasking in the workplace, or for some other
reason. From our initial interviews, we know that most
participants considered email to be a burden, at least some
of the time. We are also aware that some potential
participants excluded themselves from the study because
their jobs were solely focused on responding to email. We
would strongly advocate that similar studies be run in a
variety of different work contexts to help validate the
findings with information workers that may have different
practices and different relationships with the use of email.
We also encountered some technical difficulties with the
heart rate monitors that prevented us from capturing data
from all participants. However, our findings show a strong
trend that would very likely reach significance with
additional participants.
CONCLUSIONS
Our study has shown that there are benefits to not being
continually connected by email. Without email, our
informants focused longer on their tasks, multitasked less,
and had lower stress. It is an open question to what extent
the effects we found in our study might be sustainable. How
the benefits of reduced email usage might outweigh the
known benefits of email in reaching larger numbers of
people rapidly with information is not clear. What our study
suggests is that the tradeoffs among email usage, work
pace, stress, and collaboration need to be more closely
explored. There will always be new “zombies” lurking with
advances in information technology, and we must continue
to be vigilant in assessing the human costs that are incurred
when these advances are adopted in the workplace.
ACKNOWLEDGMENTS
This work was supported by the National Science Foundation
under award CNS-0937060 to the Computing Research
Association for the Computing Innovation Fellows Project,
as well as by the U.S. Army Natick R, D & E Center. We
would like to thank the members of the executive steering
committee at our field site for their support of our study and
our informants for their time (and patience).
REFERENCES
1. Acharya, U.R., Joseph, K.P., Kannathal, N., Lim, C.M.,
and Suri, J.S. Heart rate variability: A review. Medical
& Bio Engineering & Comp 44, 12 (2006), 1031–1051.
2. Barley, S., Myerson, D., and Grodel, S. E-mail as a
source and symbol of stress. Organization Science 22, 4
(2011), 887–906.
3. Bellotti, V., Ducheneaut, N., Howard, M., Smith, I., and
Grinter, R.E. Quality versus quantity: E-mail-centric
task management and its relation with overload. Human-
Computer Interaction 20, 1 (2005), 89–138.
4. Berntson, G.G. and Cacioppo, J.T. Heart rate variability:
Stress and psychiatric conditions. In M. Malik and A.J.
Camm (Eds.), Dynamic Electrocardiography, Blackwell
Publishing, Oxford, UK, 2007.
5. Berntson, G., Bigger, J., Eckberg, D., Grossman, P.,
Kaufmann, P., Malik, M., Nagaraja, H., Porges, S., Saul,
J., Stone, P., and van der Molen, M. Heart rate
variability: Origins, methods, and interpretive caveats.
Psychophysiology 34, 6 (1997), 623–648.
6. Bluedorn, A., Kalliath, T., Strube, M., and Martin, G.
Polychronicity and the Inventory of Polychronic Values
(IPV): The development of an instrument to measure a
fundamental dimension of organizational culture. J of
Managerial Psychology 14, 3/4 (1999), 205–231.
7. Boutcher, S.H., Nugent, F.W., McLaren, P.F., and
Weltman, A.L. Heart period variability of trained and
untrained men at rest and during mental challenge.
Psychophysiology 35, 1 (1998), 16–22.
8. Czerwinski, M., Horvitz, E., and Wilhite, S. A diary
study of task switching and interruptions. In Proc. CHI
2004, ACM Press (2004), 175–182.
9. Cutrell, E., Czerwinski, M., and Horvitz, E.
Notification, disruption, and memory: Effects of
messaging interruptions on memory and performance.
In Proc. INTERACT 2001, IOS Press (2001), 263–269.
10. Dabbish, L.A. and Kraut, R.E. Email overload at work:
An analysis of factors associated with email strain. In
Proc. CSCW 2006, ACM Press (2006), 431–440.
11. Fogarty, J., Hudson, S.E., Atkeson, C.G., Avrahami, D.,
Forlizzi, J., Kiesler, S., Lee, J.C., and Yang, J. Predicting
human interruptibility with sensors. ACM Trans.
Computer–Human Interaction 12, 1 (2005), 119–146.
12. González, V.M. and Mark, G. “Constant, constant,
multitasking craziness”: Managing multiple working
spheres. In Proc. CHI 2004, ACM Press (2004), 113–120.
13. González, V.M., Nardi, B., and Mark, G. Ensembles:
Understanding the instantiation of activities.
Information Tech & People 22, 2 (2009), 109–131.
14. Gorman, J.M. and Sloan, R.P. Heart rate variability in
depressive and anxiety disorders. American Heart
Journal 140, 4, Part B (2000), 77–83.
15. Hart, S.G. and Staveland, L.E. Development of NASA-
TLX (Task Load Index): Results of empirical and
theoretical research. In P.A. Hancock and N. Meshkati
(Eds.), Human Mental Workload. North-Holland,
Amsterdam (1988), 239–250.
16. Hewlett, S.A. and Luce, C.B. Extreme jobs: The
dangerous allure of the 70-hour workweek. Harvard
Business Review 84, 12 (2006), 49–59, 162.
17. Hjortskov, N. Rissén, D., Blangsted, A., Fallentin, N.,
Lundberg, U., and Søgaard, K. The effect of mental
stress on heart rate variability and blood pressure during
computer work. Eur J of Appl Phys 92, 1–2 (2004), 84–89.
18. Hutchings, D.R., Smith, G., Meyers, B., Czerwinski, M.,
and Robertson, G. Display space usage and window
management operation comparisons between single
monitor and multiple monitor users. In Proc. AVI 2004,
ACM Press (2004), 32–39.
19. Jackson, T., Dawson, R., and Wilson, D. Reducing the
effect of email interruptions on employees. International
J of Information Management 23, 1 (2003), 55–65.
20. Klosterman, C. My zombie, myself: Why modern life
feels rather undead. The New York Times, Dec. 3, 2010.
21. Mark, G., González, V., and Harris, J. No task left
behind? Examining the nature of fragmented work. In
Proc. CHI 2005, ACM Press (2005), 321–330.
22. Mintzberg, H. Structured observation as a method to
study managerial work. J Mgt Stud 7, 1 (1970), 87–104.
23. Nair, R., Voida, S., and Mynatt, E.D. Frequency-based
detection of task switches. In Proc. HCI 2005. Springer
(2005), 94–99.
24. Perlow, L.A. The time famine: Towards a sociology of
work time. Admin Sci Quarterly 44, 1 (1999), 57–81.
25. Renaud, K., Ramsay, J., and Hair, M. “You’ve got e-
mail!” … shall I deal with it now? Electronic mail from
the recipient's perspective. IJHCI 21, 3 (2006), 313–332.
26. Robbins, S. Tips for mastering e-mail overload.
Harvard Business School Working Knowledge (2004,
Oct 25). http://hbswk.hbs.edu/archive/4438.html.
27. Stein, P.K., Ehsani, A.A., Domitrovich, P.P., Kleiger,
R.E., and Rottman, J.N. Effect of exercise training on
heart rate variability in healthy older adults. American
Heart Journal 138, 3 (1999), 567–576.
28. Turkle, S. Alone Together. MIT Press, Cambridge, MA, 2010.
29. Wacjman, J. and Rose, E. Constant connectivity:
Rethinking interruptions at work. Organization Studies
32, 7 (2011), 941–961.
30. Whittaker, S., and Sidner, C. Email overload: Exploring
personal information management of email. In Proc.
CHI 1996, ACM Press (1996), 276–283.
31. Yeragani, V.K., Sobolewski, E., Igel, G., Johnson, C.,
Jampala, V.C., Kay, J., Hillman, N., Yeragani, S., and
Vempati, S. Decreased heart-period variability in
patients with panic disorder: A study of Holter ECG
records. Psychiatry Research 78, 1–2 (1998), 89–99.
