I'd Be Overwhelmed, But It's Just One More Thing to Do:" Availability and Interruption in Research Management

Abstract

Many CSCW projects dealing with individual availability and interruption filtering achieve only limited success. Perhaps this is because designers of such systems have limited evidence to draw upon; most data on interruption management is at least a decade old. This study uses an empirical sampling method and qualitative interviews to examine attitudes toward availability and interruption. Specifically, we analyze how corporate research managers spend their time and look at how their attitudes toward interruption relate to their various activities. Attitudes toward interruption are marked by a complex tension between wanting to avoid interruption and appreciating its usefulness. We conclude by discussing the implications of these findings for design, suggesting that the notion of socially translucent systems may be a fruitful approach.

Full text

“I’d Be Overwhelmed, But It’s Just One More Thing to Do:”
Availability and Interruption in Research Management
James M. Hudson
College of Computing
Georgia Institute of Technology
Atlanta, GA 30332-0280
jhudson@cc.gatech.edu
Jim Christensen, Wendy A. Kellogg, and Thomas Erickson
IBM T.J. Watson Research Center
30 Saw Mill River Road
Hawthorne, NY 10532
{ibmjim, wkellogg, snowfall}@us.ibm.com
ABSTRACT
Many CSCW projects dealing with individual availability
and interruption filtering achieve only limited success.
Perhaps this is because designers of such systems have
limited evidence to draw upon; most data on interruption
management is at least a decade old. This study uses an
empirical sampling method and qualitative interviews to
examine attitudes toward availability and interruption.
Specifically, we analyze how corporate research managers
spend their time and look at how their attitudes toward
interruption relate to their various activities. Attitudes
toward interruption are marked by a complex tension
between wanting to avoid interruption and appreciating its
usefulness. We conclude by discussing the implications of
these findings for design, suggesting that the notion of
socially translucent systems may be a fruitful approach.
Keywords
CSCW, availability, interruption, time management,
attention economy, managers, social translucence
INTRODUCTION
In recent years, there has been increased discussion of the
“attention economy” and “information overload.”
Essentially, these discussions suggest that the important
commodity in the current economy is no longer money or
other physical resources. Rather, it is an individual’s time
and attention. Due to limited time, attention is a limited
resource. Those who succeed will be those who best gain
others’ attention, or who most effectively deploy and
manage their own [7]. While these ideas are certainly not
new (e.g., [8]), technology seems to have exacerbated the
problem. Technology has allowed more and more
information and people to reach us than ever before. More
and more, individuals feel overwhelmed.
If technology is one of the leading causes of this problem, it
makes sense that technology could also be a solution.
While researchers have explored some approaches to
ameliorating the attentional demands of communication
technologies, however, they have met with limited success.
For example, Rodenstein, Abowd, and Catrambone [19]
designed a system prototype that allowed an individual to
filter interruptions in a lightweight fashion. Studies of this
prototype, however, revealed no significant performance
gains for those using the system. This is consistent with
later research showing that notification of an incoming
message, even when the message is ignored, is disruptive to
task performance [6]. In a somewhat more complex
approach, Milewski and Smith [15] built a telephone
system that allowed a caller to preview a callee’s self-
declared state before placing a call. Unfortunately,
Milewski and Smith weren’t able to test the usefulness of
their availability states. Instead, they discovered that users
of the system never seemed to remember to change their
availability state, rendering the preview ineffective.
One conclusion that might be drawn from this previous
work is that the strategy of requiring an overloaded,
attention-limited person to devote time to managing these
demands may not be the best approach. What is the
alternative? Instead of placing the burden on the overloaded
individual, the system could take on the management of
incoming demands for attention. Perhaps a system could
automatically filter interruptions for users; or, perhaps, it
could reveal the callee’s state to a potential caller without
requiring the callee to declare this information.
Regardless of the strategy pursued, it seems clear to us that
more information is needed. While there is a considerable
literature in this area, it is rather surprising to note that most
of the work is at least a decade old, and, clearly, the
technological terrain has changed quite significantly in that
time. Thus, in this study, we take a strongly empirical
approach and begin asking how it is that people really
spend their time, and how they view demands upon it.
Permission to make digital or hard copies of all or part of this work fo
r
p
ersonal or classroom use is granted without fee provided that copies
are not made or distributed for profit or commercial advantage and tha
t
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 2002, April 20-25, 2002, Minneapolis, Minnesota, USA.
Copyright 2002 ACM 1-58113-453-3/02/0004…$5.00.
Our study examines a group of managers in a corporate
research laboratory. While, in some regards, this group
does not fit typical management demographics (e.g., the
majority have doctorates), they still have the demands on
their attention and the frequent interruptions that
characterize management more generally. Thus, as we
minneapolis, minnesota, usa • 20-25 april 2002 Paper: Structure and Flow
Volume No. 4, Issue No. 1 97

proceed, we will show where our findings replicate earlier
literature with different populations.
Figure 1. For the pager study, each manager carried a
RIM Blackberry™ for the period of a week.
ON MANAGEMENT
A great deal is known about management – management
styles, management theories, techniques, leadership
qualities. Far less is known about how managers really
spend their time and how they cope with the numerous
forces competing for their attention.
Sproull [20] provides an early empirical study of
interruption. Sproull shadowed seven managers for 3-6
days each, taking notes on what they did. She found that a
manager’s day is dominated by brief, oral communication
(80%) and many short activities (an average of 58 activities
with a duration of 9 minutes). Activities were interrupted
21% of the time, but interestingly, managers spontaneously
interrupted themselves as often as others interrupted them.
Sproull’s data paint a picture of continuously multitasking
managers and she concludes, “The difficulty in
distinguishing between interrupted and interrupting activity
suggests that it may be misleading to think about
managerial attention in terms of tasks and interruptions”
[20, pg. 23]. This view accords with the results presented
here that show a simple view of interruption and
availability to be inadequate.
A decade ago, Panko [18] reviewed the available literature
on how managers spend their time. He found, not
surprisingly, that availability patterns varied by the
manager’s position in the corporate hierarchy; the higher a
manager was, the more time s/he spent in meetings of one
form or another. He found that managers spent
approximately 25-60% of their time engaged in some form
of communication. Half of the manager’s day is spent in
the manager’s own office. An additional ten percent of the
day is spent in the manager’s own department. Finally, he
found that the majority of meetings managers have are
dyadic face-to-face encounters.
Panko’s review has two limitations. First, it is a decade old,
and, as we’ve noted, communication technologies have
changed a lot in that time. Second, while it describes how
managers spend their time, it offers little insight into how
managers deal with interruption. Thus, while Panko
acknowledges that management is interruption driven, he
does not explore how this is manifested in the experience of
the manager or in managerial work practices. Some recent
research has pointed to the disruptiveness of interruption.
(e.g., [2, 3, 14]); however, there are also claims that
interruption can be beneficial (e.g., [4]), though, with the
exception of [16], no research is available. In our view, a
more thorough understanding of how people manage their
time and deal with interruptions, as well as the drawbacks
and benefits of interruptions, would be valuable in
informing the design of new communication technologies.
METHOD
The goal of this study was to understand how managers
spend their time and how their attitudes toward interruption
vary in relation to their activities. We used an adaptation of
Kubey and Csikszentmihalyi’s Experience Sampling
Method (ESM) [12], in which we randomly interrupted
managers with a survey delivered via a wireless pager.
Twelve managers in IBM Research were outfitted with a
RIM Blackberry™ for one week. Figure 1 shows the
device. At random intervals each day, the device
interrupted the participant (via a silent vibrating alarm), and
presented him or her with a brief survey. The interruptions
were constrained to be no less than thirty minutes and no
more than two hours apart. The survey consisted of eight
questions that could be answered by an experienced
participant in less than thirty seconds. Essentially, the
survey asked participants: “What are you doing right now?
Are you alone or with others? How would you feel about
someone interrupting you right now?” The eight questions
were designed to give us fine-grained answers about these
questions. The first four questions, which are most relevant
for this paper, are illustrated in Table 1. Our participants
informally reported the probes to be annoying, but curiosity
about the results encouraged sustained involvement.
When giving subjects the device, we conducted a short
training session to teach them how to use the device. The
study lasted for a week, and most subjects wore the paging
device from 8:00AM until 9:00PM. If participants did not
want to participate during personal hours, we modified the
times to page them only during normal business hours.
Once the participant completed the ESM portion of the
study, we conducted a follow-up interview that lasted
approximately thirty minutes. During the interview, we
asked questions aimed at determining unique constraints on
each individual, the individual’s attitude toward availability
and interruption, and the challenges they faced. The
interviews provided valuable insights into the responses
given to the ESM probes.
During the week of observations, the pager was
programmed to administer the survey ten times a day, but it
was possible for a participant to complete fewer probes. If
Paper: Structure and Flow
CHI
changing the world, changing ourselves
98 Volume No. 4, Issue No. 1

1. Right now, I am:
By Myself
Engaged with 1 other person…
This is a planned event.
This is an unplanned event.
Engaged with 2 or 3 others…
This is a planned event.
This is an unplanned event.
Engaged with many others…
This is a planned event.
This is an unplanned event.
2. I was engaged in:
Deep concentration
Reading / Writing
Watching / Listening
Interaction / Communication…
Face-to-face
Telephone
Email
Chat / I.M.
Other…
Eating
Traveling
Other…
3. This activity is:
Business
Personal
Other
4. How much time would you have for an
interruption?
It would be awkward to be interrupted.
I would prefer not to be interrupted.
I could be available for a few seconds to a
minute.
I could be available for minutes or longer.
Table 1. Questions 1 through 4 of our survey illustrate how it was designed to provide detailed information about a
manager’s state and attitudes.
there was no response to a probe, it would time out after
five minutes and the survey would be replaced with a
question that asked why the subject did not respond.
Because of these missed probes, the number of data points
for each individual varies. Subjects completed the survey
71% of the time overall and 80% of the probes during
business hours. In total, this provided 295 completed
surveys, 190 (64%) of which occurred when the subject
was engaged in a business activity. Of the missed probes
during business hours, slightly over half were because the
subject was too busy to respond. For the remaining missed
probes either the subjects did not have the paging device
with them or they did not notice the probe.
SUBJECTS
At IBM Research, there are three levels of management
between standard employees and the corporate vice
president (first-line, second-line, and third-line managers,
respectively). We solicited participants from all levels,
resulting in seven first-line managers, four second-line
managers, and one third-line manager. Two potential
subjects declined to participate because they did not want
to wear a pager or be interrupted.
With the exception of one first-line manager, all subjects
were male. Two managers (one first-line and one second-
line) were of European background. The remaining subjects
were of North American background. While there are some
suggestions that communication patterns vary along gender
[21, 23] and cultural lines [22], these issues could not be
examined here without broader participant demographics.
RESULTS AND DISCUSSION
Our results generally replicated earlier results in the amount
of time managers spend alone (42% in our data) and in how
much time they spent in communication with others (46%).
However, the combination of ESM and interview data also
revealed new results, particularly for the understanding of
managers’ attitudes towards interruption. We found a
fundamental tension between the disruptiveness of
interruption and its potential benefits. Also prominent in
our data was the importance to managers of maintaining a
sense of control over interruptions. The data also revealed
relatively consistent daily rhythms in attitudes toward
interruption. In this section, we explore these findings in
more detail.
Time Distribution
Of 190 probes in which the subject said that the current
activity was business-related, 43% of the time the subjects
were alone. The next largest activity was spontaneous,
unplanned dyadic meetings. This category accounted for
14% of the reported activities. In order of frequency, the
remaining activities were large planned meetings (12%),
planned meetings with 2 or 3 other people (9%), planned
meetings with one person (6%), and spontaneous meetings
with 2 or 3 people (4%).
Meetings that can occur through serendipitous encounters
seem to do so. As greater numbers of people are involved,
there appears to be a steep drop off in how easily an
unplanned meeting can occur. In these cases, it becomes
necessary to plan meetings in order to ensure that all can be
present. Therefore, dyadic communication favors
spontaneity while larger meetings require planning.
When managers are alone, most of the time is spent reading
and writing. Of 73 responses in this category, 68% of the
time the manager was involved in either reading or
writing
1
. “Deep concentration” (our term) was another
important activity, taking 45% of the manager’s alone time.
(Managers indicated being involved in both deep
concentration and reading/writing 34% of the time.) Other
important activities included traveling (10%), doing email
(10%), and organizing (3%). While these numbers suggest
a surprisingly small amount of time devoted to email, it is
likely that some email activity was reported as
“reading/writing.”
The type of communication channel(s) used for meetings
varied as a function of number of participants and whether
the meeting was planned or spontaneous. For two people,
all planned meetings were face-to-face, whereas impromptu
meetings took place via telephone 14% of the time.
Research has suggested (e.g., [17]) that physical proximity
is an important factor in successful collaboration. As
1
The survey allowed multiple answers to this question. Therefore, the
percentages do not sum to one hundred percent.
minneapolis, minnesota, usa • 20-25 april 2002 Paper: Structure and Flow
Volume No. 4, Issue No. 1 99

Figure 2. During business hours, managers spend a
significant amount of their time alone. Nearly half of their
day, however, is taken up with meetings and other
informal conversations.
Figure 3. Many large group meetings involve some form
of mixed modality among face-to-face, telephone, and
eMeeting participation. When managers reported
“passively attending” meetings, it was not possible to tell
which modalities were used.
meeting size increases, however, there is less reliance on
face-to-face communication. With three to four people,
14% of the planned meetings are telephone conferences. Of
eight spontaneous meetings this size, only one was a
telephone conference. Meetings with more than four people
often involved multiple media. Survey data indicated that
large face-to-face meetings involved telephone interaction
23% of the time. Large telephone conferences, on the other
hand, involved a face-to-face component 75% of the time.
In approximately one third of the cases, managers indicated
for question one that they were engaged with many other
people. For question two, however, they indicated that they
were “watching/listening” rather than engaged in some
form of communication. For these cases when managers
passively attended meetings, we were unable to determine
which modalities were in use (Figure 3).
Levels of Management
Panko [18] suggests that higher-level managers spend more
time in meetings than their lower-level counterparts. While
our sample size at each level is too small to draw strong
conclusions, we did find this pattern in the likelihood that
managers were by themselves when probed. The ratio of
planned to spontaneous meetings in our data also seems to
increase as the level of management increases. Since this
pattern has not been previously reported, it warrants further
investigation.
Availability and Time
While there are no straightforward rules to specify a
manager’s attitude toward interruption when s/he is alone,
some general patterns do emerge. Each time the probe
occurred, we asked managers to identify how much time
they had for an interruption (Table 1, Question 4). The
possible responses were that it would be awkward to be
interrupted, that the manager would prefer not to be
interrupted, that the manager could spare a few seconds to a
minute, and that the manager could be available for minutes
or longer. When managers were alone, their openness to
interruption seemed to vary regularly based on the time of
the probe. Mapping the time of a probe to the manager’s
response to this question yielded strong clustering patterns
(Figure 4).
Figure 4. A summary of managers’ attitudes toward
interruption when alone. The exact pattern for any
individual might differ from that shown here (lighter
shading indicates greater openness toward interruption).
Paper: Structure and Flow
CHI
changing the world, changing ourselves
100 Volume No. 4, Issue No. 1

Figure 5. Attitudes toward interruption vary based on the activity in which managers are engaged. Each bar indicates
proportions of responses for each state. Bars represent 100% of the responses rather than absolute numbers. Note that
there was only one recorded occurrence of an unplanned large group meeting (far right).
At least within IBM Research, managers tend to arrive at
work between 8:00AM and 9:00AM and leave between
5:00PM and 6:00PM. In the morning, until about 10:30AM
or 11:00AM, it is typically awkward to deal with
interruption. Then, there is a lull during which the manager
could be available for minutes or longer that lasts until
about 1:00PM. Between 1:00PM and 2:00PM, managers
prefer not to be interrupted. From 2:00PM to 4:00PM,
managers once more become available for interruption.
Between 4:00PM and 5:30PM, managers again prefer not
to be interrupted. After 5:30PM, attitudes vary, but tend
toward availability if the manager reports still being
engaged in business activities.
Relation between Activity and Availability
While attitudes toward availability vary throughout the day,
some general patterns are also visible based on the activity
in which the manager is engaged. Figure 5 shows a
consolidated view of managers’ attitudes towards
availability in various states. If an event is planned,
interruption is generally more awkward than if it is
unplanned. The data suggest a correlation between the size
of the meeting and attitude, but larger meetings were too
rare to determine significance. While there is not much
difference between meetings with two or three other people
and meetings with more people in the awkwardness of
interruption, there were a few cases in the larger meetings
in which managers reported being available for minutes or
longer.
Interruption Driven
Despite what seems to be a general aversion to interruption,
managers commonly spoke of themselves as being
“interruption driven.” One manager described himself as
needing interruption:
I have sort of come to rely on interrupts. If I’m not
being interrupted, I don’t know what to do. I have
to generate an internal interrupt of some sort to get
me going.
While this manager might be an extreme case, many others
echoed a more conservative version of the same theme. For
example, one manager claimed interruption as a memory
aid:
One [benefit of interruption] is so that I don’t have
such a short-term memory load. I can deal with
something now and not have to deal with it later.
More importantly, however, managers view interruption as
part of their job description. As one manager said, “I guess
I see handling interrupts as part of what I do.” While
managers do talk of learning to deal with interruptions, they
have uniformly come to see them as an important part of
what they do. All managers viewed interruption as a
necessary part of their job. Beyond simply being part of the
job description, however, managers often reported deriving
benefits from interruptions.
In the interviews, managers suggested that benefits from
interruption arise in two ways. First, openness to
interruption allows the manager to deal flexibly with
problems before they become overwhelming. One manager
summarized this saying:
Being flexible enough to respond and respond
quickly to certain kinds of interruption, I find to be
useful in getting things done.
minneapolis, minnesota, usa • 20-25 april 2002 Paper: Structure and Flow
Volume No. 4, Issue No. 1 101

Second, the interruptions themselves often carry useful
information. This information assists other projects
serendipitously.
[It’s] useful to be open to interruption to – in an
informal way – pick up information or be able to
make a connection that you wouldn’t have
otherwise. Often [the] more relaxed or offhand
way… can be more effective than setting up an
appointment.
Previous research has also suggested possible benefits to
being interrupted [16]. Our results provide additional
support for the notion of beneficial interruptions. Some
interruptions, however, can be significantly more disruptive
than others. From a manager’s point of view, the most
disruptive interruptions are ones that do not come soon
enough. These are the interruptions that occur when
something has become a crisis; they are the interruptions
that occur late because the interrupter could not or declined
to reach the manager earlier.
Managing Interruption
Despite a desire to remain open to interruption, there are
times that managers feel they must have uninterrupted time
to accomplish their tasks. Therefore, we investigated ways
that managers seek to handle their incoming interruptions.
Since different communication technologies interrupt in
different ways, we were interested in how managers use
and deal with different communication media.
For face-to-face interruptions, managers use a variety of
social strategies. Most managers, when working alone,
work in a private office. While in the office, they often use
the door as a social signal regarding their availability. Some
managers tend to use the door in a binary mode:
Either my door is open, in which case I’m available,
or it’s closed, in which case I’m not there.
For others, however, the door is a continuum that is readily
interpretable by others:
I can’t tell you how many degrees of door openness
or closeness there are, but there are many degrees.
And people generally interpret those fairly well.
With social cues like this, however, managers sometimes
feel that they miss information. When they allow others to
make decisions about when interruptions are appropriate,
they lose control over handling the interruptions. This can
mean that the manager misses information. One manager
expressed this problem:
If someone decides not to bug me [not only] will I
not know that in most cases, but I might disagree
with their decision. In fact, I know this. I know that
there are times when people did not tell me about
this thing or the other thing because they said,
“Well, your door was closed. I didn’t want to
interrupt you.” And, I very strongly disagreed and
was unhappy about that decision.
Social cues based on physical artifacts such as doors do
work, but they may require discussion of the norms
surrounding them in order to be effective cues [11]. As
such, they are not predefined artifacts (or categories of
availability), but rather have culturally constructed meaning
[10]. Meanings tend to be specific to institutions and
individuals. Understanding these meanings requires a
process of enculturation into a community of practice [13].
Telephone and email are generally handled differently than
face-to-face interruption. With the advent of voice mail,
both telephone and email can be dealt with whenever the
manager chooses. Therefore, when managers are
attempting to work without interruptions, they will
frequently attempt to ignore potential incoming
interruptions, such as telephone calls and email messages.
Some managers find this easy to do, but others find that
they need to change their physical location, whether by
moving within the office or leaving it altogether. One
manager talked about the irresistible pull of incoming email
or a ringing telephone:
It’s like reaching for the chocolate or potato chips
you’re not supposed to have. I just don’t have
enough willpower to stop myself from reaching for
[the email] and seeing what it is.
Whether one considers this a matter of personal willpower,
as this manager expresses, or a common effect of our
powerful and ever-present communication technologies, if
the manager wants uninterrupted time, the temptation to
respond must be removed. In fact, this manager discovered
that by simply moving across the office to another desk
physically separated from email and the telephone, it was
easier to ignore interruptions and focus on the task at hand.
Managers also reported using a number of other tactics to
avoid interruption. Some managers block off time in their
own calendar so that it cannot be scheduled with other
things. Some simply lock themselves in their office and
ignore most types of interruption. Others stay home where
they can better control incoming interruptions. Whatever
technique they prefer, most managers only employ it for
part of the day. That way, they remain available on a daily
basis, but still find time to work without interruption.
Work/Life Balance
In a recent, three day online global summit among IBM
employees, discussants described two distinct approaches
to balancing work and personal life so that neither
overwhelms the other. One strategy involves drawing a
hard line. During work time, personal issues should not
interfere. During personal time, work issues should remain
behind. Only in cases of emergency does this line blur.
The other strategy emphasizes flexibility. Personal issues
are dealt with on work time and work issues can be dealt
with at home. Most individuals seem to feel that they align
better with one or the other of these approaches. Our data,
however, suggest a more complex picture for those who
favor a flexible approach. It seems that these individuals
only do so when they can control the flexibility. For
Paper: Structure and Flow
CHI
changing the world, changing ourselves
102 Volume No. 4, Issue No. 1

example, when work begins to encroach on home life, they
prefer to choose when to fit it in rather than automatically
accepting an interruption.
In response to the probe study, most individuals indicated
that on personal time, business interruptions should be
either critical to the manager or critical to the interrupter.
The interviews, however, told a different story. A number
of individuals stated that they try to remain flexible. As one
manager put it, “I guess I would want to be seen as
reachable.” Several managers, even those who draw a hard
line between work and home life, echoed the theme of
difficulties that arise when others cannot reach them:
I guess I would hate to think that there is work that
isn’t being done because people can’t reach me or
they feel stuck or in a quandary or upset or whatever
because they can’t reach me. … I guess it would be
good if I were more easily reached for that time that
I’m not sitting by my work computer.
During personal time, managers want to be accessible to
those who need their attention. At the same time, however,
they wish to maintain control over these interruptions. They
do not mind doing work at home, but they want it to fit into
the holes in their personal schedule rather than disrupting it.
One manager summarized this feeling by discussing the
type of connection that he would like to have to his
professional activities:
I would not mind being connected all the time, but
more on the email side than on the phone mail side.
… It’s probably more in a pull-mode connected
than in a push-mode connected. I would be perfectly
happy to have web access all the time and no
incoming inbox.
IMPLICATIONS FOR CSCW SYSTEMS
Many computer systems designed to alleviate the problems
of interruption and limited availability fail because the
intuitive notions of designers may not match the realities of
the problem. While the word interruption typically carries
negative connotations, we have seen that managers
experience an internal tension in their attitude toward
interruption. On one hand, there is little doubt that
interruption can be disruptive to the task currently
occupying the manager’s attention. On the other hand, the
interruption may bring news related to something else that
the manager views as important. Managers need
uninterrupted time to accomplish certain tasks, but view
interruptions as important to accomplishing certain higher
level goals. As one manager put it:
I’m not sure that having fewer interruptions would
really achieve a lot because part of being a senior
manager is dealing with all of the stuff that doesn’t
work as planned. It’s just a matter of life. It’s part of
my job to deal with large amounts of interruption.
… If it could be planned, then it would just work
[out].
Therefore, managers struggle with finding the balance
between entertaining useful interruptions and avoiding
distracting ones. Achieving such a balance is inherently
problematic.
CSCW software has traditionally supported a binary notion
of availability rather than the continuum in which managers
typically work. Users of these CSCW systems must declare
their attitudes toward interruption in advance. These
settings are then used to filter potential interruptions.
While this method can produce uninterrupted time, it does a
poor job of supporting the potential benefits of being
interrupted. Our findings suggest that designers of CSCW
systems should focus on making interruptions more
effective rather than on decreasing them.
How to make interruptions more effective, however, is a
challenge. This study offers a number of suggestions for
how CSCW systems can be designed to do this. First, there
is rarely a state during which an interruption would be
ideal. By their nature, interruptions will disrupt something.
This implies that systems should not be designed to queue
possible interruptions until the ideal time. Rather,
interruption should occur at the best relative time. As this
study suggests, there are periods of lull during the day and
certain states during which interruption is better received
than other times or states. While the details of this study
cannot be applied to all populations, the suggestion of
regular patterns of acceptable interruption times should be
explored for other CSCW audiences.
Another interesting implication of this study is that
technology has not significantly changed a manager’s daily
life. The discourse surrounding the concept of information
overload suggests that technology is exacerbating the
problem. This study, however, suggests that managers still
follow the same communication patterns that were
documented ten [18], twenty [20], or even thirty years ago
[5]. While changing technology certainly has effects on
society, it is not clear that technology is causing all of the
challenges that critics predict.
The complexities surrounding availability imply that the
acceptability of interruption is a socially constructed
phenomenon. Because of this tension between benefits and
disruptiveness, certain interruptions receive a higher
priority than other ones. More importantly, the priority of
an interruption relative to the current task varies.
Regardless of state or time, availability is handled
differently depending on the nature of the interruption.
Technological systems are rarely able to independently deal
with this sort of social construction [1]. Therefore, social
processes need to be designed into any system designed to
ease the challenge of limited attention. We believe the
notion of socially translucent systems [9] can provide one
way to approach this challenge. Creating awareness and
accountability through making behavior more visible – the
definition of a socially translucent system – can allow
social mechanisms to play more effective roles in
technology-mediated interruptions. For example, by
minneapolis, minnesota, usa • 20-25 april 2002 Paper: Structure and Flow
Volume No. 4, Issue No. 1 103

making information such as current activity, location,
historical patterns of activity, etc., visible in a way that
does not require vigilance and active maintenance on the
part of the recipient, potential interrupters can make better-
informed decisions about whether to interrupt. Of course,
as the managers in our study report, relying on social
mechanisms to manage interruption is not infallible and
does not always accord with their wishes. But, when they
fail, it tends to result in negotiation – that is, further social
interaction can be used to repair or gloss over problems. On
the other hand, when computer filtering fails, only anger
and sometimes a sense of helplessness results. Information
and attention are complex social processes that would seem
to require social solutions. Designing socially translucent
systems to manage interruption can help embed these
processes in the technologically-mediated systems through
which we interact.
ACKNOWLEDGEMENTS
We especially thank the busy managers that agreed to let us
constantly interrupt them for this study. We also thank the
individuals that provided useful feedback, especially the
Grapevine and Social Computing groups at IBM Research.
Thanks to Cheryl Kellogg for the quote used in the title of
this paper. Special thanks to Peter Malkin, Tracee Wolf,
Karrie Karahalios, Amy Bruckman, and the Electronic
Learning Communities research group at Georgia Tech.
REFERENCES
1. Ackerman, M. The Intellectual Challenge of CSCW:
The Gap Between Social Requirements and Technical
Feasibility. Human-Computer Interaction, 15 (2-3).
181-205.
2. Bailey, B.P., Konstan, J.A. and Carlis, J.V. The Effects
of Interruptions on Task Performance, Annoyance, and
Anxiety in the User Interface. in Proceedings of
INTERACT 2001, NCP, Tokyo, Japan, 2001.
3. Bailey, B.P., Konstan, J.A. and Carlis, J.V. Measuring
the Effects of Interruptions on Task Performance in the
User Interface. in IEEE Conference on Systems, Man,
and Cybernetics 2000, IEEE, Nashville, TN, 2000, 757-
762.
4. Cohen, D. and Prusak, L. In Good Company: How
Social Capital Makes Organizations Work. Harvard
Business School Press, Boston, MA, 2001.
5. Cohen, M.D. and March, J.G. Leadership and
Ambiguity: The American College President. McGraw-
Hill, New York, NY, 1974.
6. Cutrell, E., Czerwinski, M. and Horvitz, E. Notification,
Disruption, and Memory: Effects of Messaging
Interruptions on Memory and Performance. in Interact
2001 Conference Proceedings, Tokyo, Japan, 2001.
7. Davenport, T.H. and Beck, J.C. The Attention
Economy: Understanding the New Currency of
Business. Harvard Business School Press, Cambridge,
MA, 2001.
8. Derber, C. The Pursuit of Attention: Power and Ego in
Everyday Life (Second Edition). Oxford University
Press, New York, NY, 1979, 2000.
9. Erickson, T. and Kellogg, W.A. Social Translucence:
An Approach to Designing Systems that Support Social
Processes. ACM Transactions on Computer-Human
Interaction, 7 (1). 59-83.
10. Hall, E.T. The Silent Language. Anchor Books, New
York, NY, 1959, 1981.
11. Kristoffersen, S. and Ljungberg, F. An Empirical Study
of How People Establish Interaction: Implications for
CSCW Session Management Models. in Human
Factors in Computing Systems: Proceedings of CHI '99,
ACM Press, Pittsburgh, PA, 1999, 1-8.
12. Kubey, R.W. and Csikszentmihalyi, M. Television and
the Quality of Life: How Viewing Shapes Everyday
Experience. Lawrence Erlbaum, Hillsdale, NJ, 1990.
13. Lave, J. and Wenger, E. Situated Learning: Legitimate
Peripheral Participation. Cambridge University Press,
Cambridge, UK, 1991.
14. McFarlane, D.C. Coordinating the Interruption of
People in Human-Computer Interaction. in Sasse, A.
and Johnson, C. eds. Proceedings of Human-Computer
Interaction (INTERACT'99), IOS Press, 1999, 295-303.
15. Milewski, A.E. and Smith, T.M. Providing Presence
Cues to Telephone Users. in Whittaker, S. and Kellogg,
W. eds. Proceedings of the Computer Supported
Cooperative Work (CSCW) 2000 Conference, ACM
Press, Philadelphia, PA, 2000, 89-96.
16. O'Conaill, B. and Frohlich, D. Timespace in the
Workplace: Dealing with Interruptions. in Proceedings
of Human Factors in Computing (CHI '95), ACM Press,
Denver, CO, 1995, 262-263.
17. Olson, G.M. and Olson, J.S. Distance Matters. Human-
Computer Interaction, 15 (2&3). 139-178.
18. Panko, R.R. Managerial Communication Patterns.
Journal of Organizational Computing, 2 (1). 95-122.
19. Rodenstein, R., Abowd, G. and Catrambone, R.
OwnTime: A System for Timespace management. in
Human Factors in Computing Systems: Proceedings of
CHI '99, ACM Press, Pittsburgh, PA, 1999.
20. Sproull, L.S. The Nature of Managerial Attention. in
Advances in Information Processing in Organizations,
JAI Press, 1984, 9-27.
21. Sproull, L.S. The Nature of Managerial Attention.
Advances in Information Processing in Organizations,
1. 9-27.
22. Tannen, D. Gender Differences in Conversational
Coherence: Physical Alignment and Topical Cohesion.
in Tannen, D. ed. Gender and Discourse, Oxford
University Press, New York, NY, 1994, 85-135.
23. Tannen, D. That's Not What I Meant! How
Conversational Style Makes or Breaks your Relations
with Others. William Morrow and Co., New York, NY,
1986.
24. Tannen, D. You Just Don't Understand: Women and
Men in Conversation. William Morrow and Co., New
York, NY, 1990.
Paper: Structure and Flow
CHI
changing the world, changing ourselves
104 Volume No. 4, Issue No. 1