What makes mobile computer supported cooperative workmobile? Towards a better understanding of cooperative mobile interactions

globedeepMobile - Wireless

Nov 24, 2013 (3 years and 6 months ago)

84 views

Int.J.Human-Computer Studies 60 (2004) 737–752
What makes mobile computer supported
cooperative work mobile?Towards a better
understanding of cooperative mobile interactions
Gregor Schrott
a,
*,Johannes Gl.uckler
b
a
Institute of Information Systems,Frankfurt University,Mertonstr.17,Frankfurt D-60054,Germany
b
Institute for Economic and Social Geography,Frankfurt University,Dantestr.9,
Frankfurt D-60054,Germany
Abstract
Despite the high availability of mobile phones and personal digital assistants with online
capabilities,mobile computer supported cooperative work is still in its infancy.So far,only
little is known about the distinct attributes of mobile cooperative work in comparison to its
stationary counterpart.Across which dimensions does cooperation via mobile devices differ
from traditional hard-wired settings and what implications have to be drawn for future
research?To bring more light to this question,we conducted an experimental business-case at
Frankfurt University with 16 graduate students and analysed their collaborative behaviour
across mobile and non-mobile channels of communication over a 5 week period.We find that
mobile messages differed from stationary messages in terms of size and that the use of mobile
emails prevailed over stationary emails under conditions of stress.Finally,we found that the
social structure of mobile communication corresponded with the structure of stationary
communication.This indicates that mobile communication technologies support existing
communication relations rather than creating new relations.From the perspective of system
designers,these results may serve as practical insights into the user behaviour of mobile
technologies and might support the future development of mobile computer supported
cooperative work environments.
r 2003 Elsevier Ltd.All rights reserved.
ARTICLE IN PRESS
*Corresponding author.Tel.:+49-69-798-23318;fax:+49-69-798-28585.
E-mail addresses:gschrott@wiwi.uni-frankfurt.de (G.Schrott),glueckler@em.uni-frankfurt.de
(J.Gl.uckler).
1071-5819/$ - see front matter r2003 Elsevier Ltd.All rights reserved.
doi:10.1016/j.ijhcs.2003.11.006
1.Introduction
Mobile work scenarios in which business managers interact with their colleagues
in an ‘anytime,anywhere’ (Johansen,1988) fashion have become reality on a large
scale.Several factors have contributed to this development.First,immense efforts
have been undertaken towards organizational restructuring in order to reduce the
level of hierarchy within organizations and to create more permeable internal and
external boundaries.Unlike in the ‘old days’ employees do not have to be co-located
with their colleagues to interact.Instead,the vision of future work scenarios
constitutes settings in which globally distributed individuals may work together
simultaneously without sharing physical co-presence (Malone and Laubacher,1998).
The second factor that has contributed to this development has been the availability
of technological infrastructure.For example,global system for mobile communica-
tion protocol is currently applied in 190 countries of the world on a reliable basis and
the number of worldwide subscribers is expected to exceed the one billion mark at
the end of this year (GSM-World,2003).In Germany,for the first time in 2000 the
number of mobile subscribers was higher than the number of traditional wired
telephone users (IZT,2001).Third,the number of providers for mobile devices has
been increasing steadily.The variety of personal digital assistants with mobile access
and Smartphones has never been greater before.Fourth,the shrinking costs of data
transmission and the sponsoring of devices have additionally catalysed the
distribution of mobile technologies and amplified the growth of the worldwide
mobile market.Together with a strong emphasis on collaboration and group work
(Reichwald et al.,2000;Katzenbach and Smith,1993) these factors have led to
establish mobile cooperative work as an integral part of our daily working life.
Surprisingly,from a research perspective still little is known about the distinct
features of mobile employee collaboration in comparison to workstation-centric
computer supported cooperative work settings.While several technically orientated
publications have described innovative systems and prototypes for mobile
collaboration and made valuable suggestions on how to improve the usability of
mobile devices,the socio-technical perspective has not been equally developed
(Bellotti and Bly,1996;Fagrell et al.,1999;Fagrell and Ljungberg,2000;Wiberg and
Gr
.
onlund,2000).In order to contribute to a better understanding of the interaction
and communication processes within mobile computer supported cooperative work
settings,we focused on the following questions:
*
Do mobile messages differ fromstationary messages with regard to frequency and
size?
*
Under what conditions do users prefer mobile over stationary communication?
*
Does mobile communication imply or foster different structural communication
patterns,i.e.do people use different technologies to relate to different people?
We believe that a more comprehensive understanding of users’ behaviour will lead
to an improvement of systems development and its applicability.The paper is
organized in five sections.Section 2 discusses related work in the field of mobile
ARTICLE IN PRESS
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752738
computer supported cooperative work and develops the hypotheses for our
argument.Section 3 describes the experiment and the infrastructure used.After
that,Section 4 presents the findings and discusses the results.Section 5 summarizes
the argument and provides an outlook on possible further research.
2.Characteristics and consequences of mobile communication
Mobile computer supported cooperative work issues have been tackled
predominantly from a technical perspective.The availability of technical infra-
structure,devices and software applications certainly forms the prerequisite for
wireless collaboration in global system for mobile communication networks.
Recently,some economists and business administrators have started working on
the creation of valuable business models and the applications of economic principles
to cooperative multi-player scenarios in the mobile world.Apart from these two
strands of thought,social scientists have raised questions about how the people who
have started to use mobile devices integrate this technology into their daily life and
how mobile technology enables and constrains new patterns of interaction (e.g.Katz
and Aakhus,2002).One illustrative example is Muziko Ito’s analysis of how the
increasing availability of mobile communication infrastructure has influenced
the transformation of the Japanese youth (Ito,2001).In addition to studies on the
impact of communication technology on the social life world,more research is
needed on the integration of mobile communication technology into corporate
business processes and work relations.For the near future,the support and
assistance of corporate work environments is expected to remain one of the most
important applications to satisfy user needs.But what makes mobile computer
supported cooperative work different from its stationary counterpart?In this paper,
we will focus on three types of differences.First,we take a look at the frequency and
volume of message exchanges;second,we look at the conditions in which mobile
email communication is employed and third,we look at the social structure of
networks that mobile communication reflects between work collaborators.
2.1.Mobile vs.stationary electronic messages
Mobile devices differ from stationary personal computers (PCs) along various
attributes.Smaller displays (Brewster,2002),limited input-capabilities (James and
Reischel,2001) as well as smaller storage capacity and bandwidth (Varshney,1999;
Dunlop and Brewster,2002) affect the ergonomics of these devices to a great extent.
Pocket PCs are usually equipped with a 240 ￿320 pixels and 4096 colours display,
which is very small and limited in comparison to a desktop screen.To enter data,
users either have to use a small keyboard or type-recognition both incorporated into
the touch screen,which is much less comfortable than a conventional keyboard.The
batteries usually last for 4–5 h when extensively using the device.Storage is limited as
well.Despite the possibility of expanding the standard storage through media-cards,
most devices on the market currently offer 32–64 megabyte for data storage.
ARTICLE IN PRESS
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752 739
Another important factor is the limited bandwidth.When using plain global system
for mobile communication protocol or general packet radio services protocol,the
bandwidth is far smaller than the usual internet connection.All these factors make it
somewhat ‘inconvenient’ for the collaborators to use mobile devices for electronic
messaging.It seems obvious that limited comfort and capacity of mobile devices
have consequences on the intensity and structure of communication facilitated
across this medium.Therefore,one would expect messages that are sent via mobile
devices to differ from stationary emails with regard to frequency and size.The
limited input–output capabilities of mobile devices and other limitations mentioned
above are assumed to have an effect on the emails sent via these devices.We expect
users to write shorter messages over mobile channels.Apart fromdifferences induced
by different ergonomics,the situations in which mobile devices are used to exchange
email messages will play an important role,too.
2.2.Situation dependence
Several authors have described the specifics and advantages of mobile services in
general (B.ullingen and W
.
orter,2000;Gerpott,2001;Reichwald et al.,2002;
Kollmann,2001;M.uller-Veerse,1999;Schmitzer and Butterwege,2000;Wiedmann
et al.,2000;Wohlfahrt,2001;Zobel,2001).Apart fromthe independence of location,
instant connectivity is articulated as a core feature of mobile services compared to
conventional desktop applications.When looking at marketing campaigns from the
major mobile operators and research studies,one can easily see that it is on top of
users’ wishes to be able to receive emails formother colleagues without delay (Hyers,
2001).However,given the inconvenience of the application of this technology,
mobile communication over small devices is not yet capable of entirely replacing
stationary communication over conventional desktops.In turn,when communica-
tion is urgent and stationary infrastructure is not available,we expect mobile devices
to be used more frequently.More generally we suggest that under conditions of
stress and the need for immediate communication,the use of mobile technology for
email exchanges will increase.If the saying ‘time is money’ is right,then professional
collaborators will prefer mobile communication channels to save time and prefer
mobile exchanges of email if time is short.
2.3.Social structures of communication
Knowledge officers who are responsible for the successful management of
computer supported cooperative work environments have to care for various tasks.
Apart from the design of computer supported cooperative work platforms (Pinelie
and Gutwin,2000) and soft-skill matching for successful team formation (F
.
arber
et al.,2003),an analysis of the underlying communication structure is believed to
contribute to a better understanding of dynamics within these communication
networks.Furthermore,knowledge of the social structure of communication
relations would allow for a more subtle understanding of how to improve work
organization (Cross et al.,2001;Cross and Prusak,2002).However,analysing
ARTICLE IN PRESS
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752740
communication networks is a difficult endeavor,since collaborators are not to be
understood as isolated individuals but as members of a network that has a
characteristic structure of information flows.In order to identify the role of
individuals in a network one first has to know the overall pattern of communication
in the network.Social network analysis is a research methodology developed in
anthropology and structural sociology that offers formal techniques to study the
structural patterns of communication networks (Mitchell,1969).Its techniques
‘‘provide explicit formal statements and measures of social structural properties that
might otherwise be defined only in metaphorical terms’’ (Wasserman and Faust,1994).
We believe that it is necessary to analyse and understand the structure of
communication that evolves and is reproduced in collaboration networks in order to
be able to conclude practical interventions for knowledge management purposes.
Although a lot has been written about knowledge management and computer
supported cooperative work,little is known about the actual structure of interactions
and dynamics in mobile collaborate work environments.Hence,those contributions
can so far only be marginal to the real-world problems of designing collaborative
scenarios (Schoberth and Schrott,2001).One interesting issue at the interface
between technology and social communication is the question to what extent
technological changes affect social practice and communication structures.Do the
differences between mobile and stationary possibilities to communicate imply
different interactive patterns between the individuals involved?In other words:do
different communication media cause different social structures of communication?
3.The experiment
In order to be able to gain information about users’ behaviour in mobile computer
supported cooperative work settings,we designed a university experiment.Sixteen
graduate students were invited to join an experimental business case project at the
Department of Information Systems at Frankfurt University.For a period of
5 weeks they were given a project task to fulfil and finally present their result to a
professional audience.
3.1.Design
The experimental setting of the business case was as follows:SDD Inc.(Superior
Digital Devices),a fictional company,was a Helsinki-based enterprise with 3500
employees and annual revenues of $21 billion.Students had to imagine that they had
been recruited during a seminar held at University in 2000 and had now been
employed with SDD Inc.for 3 years.Unfortunately,they had lost track of each
other and were now split over several departments of SDD Inc.(marketing,
development,sales and international business).The major task for them was to
design,plan and organize a spontaneous product launch for the ‘‘BPA 998’’ (Better
Performing Assistant,a new personal digital assistant with telephone functionality)
at New Years Eve of 2002 in Frankfurt.Originally,the launch was planned for
ARTICLE IN PRESS
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752 741
CEBIT 2003 in March.But since the main competitor was going to present a new
product at the GSM World Congress in February 2003,i.e.to gain a first-mover
advantage,the presentation had to take place as soon as possible.Unfortunately,all
of their superiors were on vacation when the chief executive officer (CEO) called for
their help to plan this event.Each of four departments received instructions about
their specific tasks and the requirements for coordination with the other
departments.In addition,the students got a budget limit for the event,which they
had to split up according to their costs (e.g.catering,sound,speakers’ honoraria).
Instructions were designed in order to require an interactive work flow,as the
following example illustrates:The development department,e.g.had to specify the
technical characteristics of the new product in order to enable the marketing
department to identify the target customer group and to prepare an invitation list for
the product launch.The total number of participants then had to be communicated
to the international business department,which was responsible for the event
location.The sales department,for instance had to calculate the price of the product
according to the development department’s technical specification.Hence,all tasks
were highly interrelated and a lot of coordination was required from the students,
especially with regard to the allocation of the budget across departments.
In addition,rules and incentives were specified in order to render the experimental
project more realistic and in order to ensure student commitment.The overall task
was split up in several sub-tasks,which where enforced by the chief executive officer
(course supervisor).In case of project-failure,all students would not have been given
credits for the course.A continuous necessity to communicate across departments
was stimulated through external interventions:Re-scheduling deadlines and budget
reductions served as stress-factors that ensured ongoing inter-group coordination.In
addition to these sanctions,there was also a clear incentive for commitment to the
project.Students were informed ex ante that at the end a winner team would be
awarded an attractive prize.
3.2.Technical equipment and communication channels
Interaction between the individuals and the teams was the primary unit of analysis
of the project.The electronic communication was facilitated in two ways.Email
messages could be sent and received either via stationary computer facilities (private
or university lab) and a web-based communication portal (mobile service portal
(MSP),see below) or via mobile devices and a web-based communication portal
(MSP,see below).To enable the exchange of emails,students were provided with a
collaboration portal and mobile devices.We used a mobile communication portal
from T-Mobile (MSP,MSP) with its architecture being displayed in Fig.1.The
portal provides a typical collaboration functionality based on a Microsoft Exchange
2000 server,which is protected by firewalls,and allows for access using various
devices.Depending on the attributes of the devices,the MSP generates appropriate
style sheets to display the different contents.The portal allowed for access across
various channels (personal digital assistant,desktop,phone) and provided
collaboration functionality like email,public folders and short message service
ARTICLE IN PRESS
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752742
(SMS)-notification.For further information about the technical specification of the
portal see T-Mobile (2002).We customized the portal to the corporate structure of
the SDD Inc.Intranet to make the environment as realistic as possible for the
students.
In order to provide students with a 24h possibility to use electronic communica-
tion,everybody received a subscriber identification module (SIM)-card and a mobile
device (cf.Fig.2).Eight students received a T-Mobile MDA and eight students a
combination of Compaq iPaq H3870 and Nokia 6310 (cf.Fig.2).The T-Mobile
MDA had an integrated global system for mobile communication module and
allowed for transport control protocol/internet protocol connections to the server,
the Compaq 6310 was enabled for server connection by providing the students with a
Nokia 6310 phone with Bluetooth capability.At the initial meeting,we assured that
all students were able to get access to the portal and that no severe technical
problems remained.The SIM card as well as the mobile federations were free of
charge for the students.
ARTICLE IN PRESS
Fig.1.Architecture MSP.
Fig.2.T-Mobile MDA and Compaq iPaq H387/Nokia 6310 bundle with SIMcards.
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752 743
3.3.Data collection and methods
Both,mobile as well as stationary electronic messages were recorded electronically
by log file-analysis.The email log files went through several steps of data
transformation (Fayyad,1996;Mobasher et al.,2000).All communication flows
towards and fromthe chief executive officer were removed fromthe raw data,as well
as ‘delivery-failure-notices’ and administrative questions.Students were instructed to
respect a series of conventions on how to use these media:(1) the students were not
allowed to use any other than the project email address;(2) all electronic messages
had to be titled in a defined way in order to distinguish administrative questions
from content messages,as well as to identify emails sent through the mobile channel
more easily during the data analysis.Besides,students were asked to avoid
‘broadcasting’,which means spreading their messages to everybody in the group
although they were not directly affected by the message.
Apart from the electronic documentation of email messages,students were asked
to complete a questionnaire after the project had ended.Here,students evaluated the
different communication channels in terms of their usefulness and convenience.
Further,they indicated in which situations they used what kind of electronic
channel.We used these data in order to evaluate our results.
4.Results
The project was carried out during 5 weeks between November 11th and December
15th 2002.Although students were allowed to have face-to-face meetings,use mobile
telephones and the SMS,most communication was electronic.Because of the scarce
use of phone calls and SMS messages,these media were dropped out of the analysis
and focused exclusively on mobile and stationary email communication.
1
During this
period students interacted intensively within their teams as well as between teams:
2745 dyadic interactions were realized via desktop-access and 990 via the mobile
clients throughout the 5 weeks of observation.The number of dyadic interactions
counts all communication acts between two individuals.If student x for example,
sent a mail to student y and z,the number of dyadic interactions would be two.
A density analysis of both communication networks shows that stationary
electronic communication not only outweighed mobile communication in frequen-
cies but also implied a higher density of interrelations (Fig.3).
2
While stationary
ARTICLE IN PRESS
1
The small number of phone calls might be explained due to higher communication costs:While the
SIM cards where free of charge to the students and only allowed data transmission,voice transmission
(phone calls) had to be made through private phones and paid individually.The little number of SMS
might be due to the availability of mobile email access.When being equipped with mobile email access
including ‘‘always on’’-general packet radio services-functionality,the differences between SMS and using
the MDA to write short messages is very marginal.In addition to that,operating systems,such as Pocket
PC Phone Edition allow users to boot their mobile device within seconds.
2
The network structures have been visualized with the software package Pajek (Batagelj and Mrvar,
1998).
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752744
emails connected 72% of all possible pairs of collaborators within the project
groups,the overall density of the mobile-network was just half as much with a value
of 36%.
4.1.Mobile vs.stationary messages
The log files of stationary and mobile emails contained information on sender and
receiver,date of the mail,subjects and size of the message.In order to investigate the
difference between stationary and mobile messages we analysed the size of the
exchanged messages.Emails sent over desktop PCs were far bigger than the mails
sent via mobile devices.The average size of a desktop email was 119.5 kb while an
average mobile email sized only 4.1 kb,which is about thirty times less the volume
than that of desktop messages.Of course,the mobile service portal optimizes mobile
emails to reduce data costs by leaving out some graphical elements (buttons).
However,this cannot account for the significant difference in size.In sum,emails
sent over mobile devices were significantly smaller than emails sent over a desktop
PC.
In the questionnaires students reported that,given the inconvenience,they tried to
keep the messages as short as possible when they used mobile devices.In order to get
a better understanding of the criteria for the choice of the different media,students
were asked to assess the effectiveness of mobile and stationary messaging.
3
On a
four-point rating scale (with 4 indicating highest effectiveness) electronic messages
scored less than face-to-face meetings (3.81) which offered the maximum richness of
exchange.However,electronic messages were used far more frequently and thus
embraced a large part of the entire interaction within the project group.Therefore,
stationary emails scored 3.44 and did not fall much behind the perceived
effectiveness of meetings.In contrast,mobile emails scored clearly less with only
2.38 points on average.Interestingly,meetings and desktop emails only received high
marks (3 and 4),while the evaluation of mobile emails ranged from1 to 4,showing a
higher heterogeneity.
ARTICLE IN PRESS
Fig.3.Network graphs of communication channels.
3
For reasons of privacy,we did not analyse the content of individual email messages.
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752 745
Apart from evaluating the effectiveness of the media,students were asked to
briefly state the main reasons for using stationary and mobile emails.Mobile emails
were mainly sent for quick and spontaneous coordination.Especially when students
were ‘on the road’ mobile devices allowed themto be up to date and respond quickly
to changes.In all other cases,stationary emails outweighed mobile communication
because it was more convenient and allowed the enclosure of attachments,e.g.
product specifications.However,both media were qualified as weak with regard to
bargaining,decision taking,and developing creative ideas.In those instances,
students preferred to meet personally in order to make use of the increased richness
of information exchange.
4.2.Situation dependence
The target and work plan of the experiment were challenging for the participants.
They had to organize and coordinate themselves in a very professional manner to
complete their tasks successfully.Especially the splitting up of the global budget into
team-budgets and the fulfilment of deadlines required an intense and faultless
interaction between the students.Additional stress was imposed on the group by
shortening the budget at one stage and advancing the deadline for a milestone at
another stage.The following graph (Fig.4) gives an overview of the communication
activity along the duration of the seminar.
The dotted line indicates the number of dyadic messages sent over a stationary
desktop PC,the solid line reflects those which were sent over a mobile device.In
ARTICLE IN PRESS
0
50
100
150
200
250
300
350
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
project day
number of daydic messages
stationary emails
mobile emails
Fig.4.Electronic communication over time.
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752746
total,the number of desktop emails (2745) exceeded by far the number of mobile
emails (990).The distribution of email messages over time shows that the flow of
communication was not continuous at one level of intensity but varied considerably
in terms of frequency.Over the 5 week period a number of communication peaks can
be observed as well as a number of days nearly without any message exchanges.
However,the distribution of the two communication media was similar over time:
mobile and stationary emails followed the same temporal pattern and were
correlated with r = 0.68 (p o 0.01,N=33).Despite the similar temporal pattern,
however,there are various instances,in which the number of mobile messages
prevailed over stationary messages.Fig.5 illustrates the differences between the
number of interactions.White bars indicate the dominance of stationary emails,
black bars the dominance of mobile emails.
During the project period,the number of mobile messages exceeded stationary
emails on 5 days:November 13th 2002,November 15th 2002,December 4th 2002,
December 7th 2002 and December 16th 2002.What were the conditions that made
mobile communication more effective in these instances?Acloser look at the process
of task-assignment helps to reconstruct the dominant use of mobile devices.Usually,
the supervisor sent emails to the students providing themwith details on the task and
the corresponding deadline.Students were also informed that task completion was
necessary within the deadline to get the credit points for the course.This constraint
was introduced in order to enforce collaboration and interaction.
ARTICLE IN PRESS
-60
-10
40
90
140
190
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
project day
number of dyadicmessages
Fig.5.Daily frequency differences between stationary and mobile email messages.
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752 747
The students had to fulfil three main tasks.The first task was to group themselves
in four teams with four students each.After handing out the login-information for
the portal,the mobile devices and SIMcards on the evening of November 11th,the
students received an email,which asked themto build teams and report membership
by November 15th.The extensive use of mobile devices on November 13th (first
white bar in Fig.4) is a result of the students acquainting themselves with the mobile
devices and testing for the functionality of the portal.The second peak on November
15th was due to the fact that students wanted to get the response fromthe supervisor
about the process of team formation.Since they attended other courses in the
university,they checked with their mobile devices in order to be able to immediately
comment with their colleagues afterwards.The second task was to generate a
preliminary version of the event-plan for the product launch.The deadline for
sending the document to the CEO was Monday,December 1st,evening.Due to a
technical problem at the server farm of the provider,the mobile access to the portal
was not available for one day so that stationary emails prevailed clearly until
December 3rd.The third peak of mobile communication on December 4th was a
consequence of the fact that the draft of one teamwas rejected and a second deadline
for December 6th was set.The corresponding teamused their mobile devices in order
to quickly spread news and organize their work program for revision.Correspond-
ingly,on Sunday,7th the students also used their mobile devices to communicate
about the response about their submitted drafts.The last time that mobile
communication exceeded stationary emails was on the last day of the course.Again
the same pattern could be observed.After the deadline for the final paper on
December 15th had expired,students wanted to be able to quickly respond to any
response that would have affected their work.
In sum,mobile emails prevailed over stationary communication under conditions
of stress.Whenever response and mutual coordination were probable and had to be
managed quickly students used mobile devices in order to be alert and react without
delay.More generally it can be concluded that with increasing stress,collaborators
were more likely to use mobile devices.
4.3.Structural attributes
Apart from differences in frequency,we were interested to see whether the two
types of communication channels reflected different structural patterns of commu-
nication.
4
We used the quadratic assignment procedure in order to check for
systematic differences across the two communication relations and to see whether
different communication relations supported different patterns of interaction
(Krackhardt,1987).The quadratic assignment procedure proceeds as follows:first,
for two observed matrices,stationary and mobile emails,the Pearson correlation
coefficient is calculated.Then the rows and columns of the comparative matrix are
permuted to provide a new,mixed up matrix.Now both matrices are correlated
ARTICLE IN PRESS
4
For a detailed analysis and discussion of the results of the social network analysis see Gl.uckler and
Schrott (2003)
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752748
again and the r-value is stored away.This permutation-correlation process is
repeated an arbitrarily large number of times (here 2500).The individual r-values
fromthe permutations generate an empirical reference distribution against which the
observed original r-value is then compared.If fewer than 5%of the r-values fromthe
random correlations are larger than the observed r-value of the original correlation,
the correlation is qualified significant at the 0.05 level.
The quadratic assignment procedure analysis between mobile and desktop emails
shows an intermediate correlation coefficient (r = 0.418).In all cases the 2500
random permutations of the matrices did not produce a single correlation that
exceeded this observed r-value.Therefore,with the significance at level 0.01 the two
matrices are statistically associated with each other,which means both communica-
tion media reflect the same communication pattern.The students who committed
themselves to other students by using desktop emails,were most likely to also
commit their mobile emails to these students.Although communication spread
across different media,this does not imply that the different media served different
communication relations.Hence,communication exchange displayed a similar
structural pattern across the media.In response to the question raised in Section 2.5
this result shows that different communication media do not imply different social
patterns of interaction.Instead it seems that different media enrich existing
communication relations and increase the possibilities to reach others.Mobile
communication,therefore,reflects the same kind of social work relations and can be
employed in order to complement and support existing communication networks.
5.Summary and further research
Research about mobile computer supported cooperative work lags behind current
practices.Mobile interaction and communication processes of cooperative groups
are still poorly understood from a socio-technical perspective.To get a better
understanding about the distinct attributes of mobile communication processes,
especially exchange of emails in comparison to communication which are initiated
via desktop PCs we conducted a 5 week cooperative business-case experiment with
16 students at the Business-Administration and Economics Faculty,Frankfurt
University,Germany.All students were equipped with mobile devices and access to a
web-based communication portal.In contrast to many other experiments,our focus
was not exclusively on the usage of mobile channels of communication.We enabled
students to choose between various modes of communication,such as sending and
receiving emails via traditional desktop PCs,which allowed for a comparative
analysis of usage behaviour between these modes.
In the experiment we found that emails sent over mobile devices differ from
‘traditional’ emails.The higher the demand for voluminous message exchanges,the
less likely students chose mobile devices for communication.Interestingly,the use of
mobile communication was highest under conditions of stress.Shortly before and
after the deadlines for the submission of project reports the number of mobile emails
exceeded the number of emails sent via desktop PCs.These insights in user behavior
ARTICLE IN PRESS
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752 749
should assist designers and engineers to further understand the challenges of mobile
devices for human computer interaction (Dunlop and Brewster,2002).Future
research as well as system designers should pay special attention to user groups that
work in an environment of permanent interaction,changes and mutual recoordina-
tion.Despite the difference in size and application of mobile and stationary emails
we found that the different media did not imply different communication patterns.
The social structure of mobile communication,i.e.the structure of who
communicated with whom,did not significantly differ from the structure of
stationary email communication.
Notwithstanding,this experiment faces a number of limitations:first,due to the
limited availability of mobile devices and smart cards the number of students was
limited to 16.Second,students were not charged for using mobile connections.While
this may correspond with the provision of mobile communication in big
corporations,it would be interesting to see how different pricing schemes affect
the use of mobile devices.Third,while students had owned mobile phones for more
than 2 years and were technically interested,they had only little experience with
mobile devices for data transmission.Since the use of mobile devices was new for
most of the participants it would be helpful to see whether the use of mobile
communication increases substantially with increased routine.Fourth,students did
not have any additional equipment,as for instance external keyboards.Probably,
some of the limitations of user convenience can be reduced with additional tools and
will ‘ease the pain’ of mobile communication.Finally,the students in the experiment
always had the possibility to meet each other on campus and thus avoid mobile
communication.However,since in many real-world situations collaborators are
physically distributed over large distances,mobile communication may be the only
viable possibility to exchange information.So we expect mobile communication to
have a more pronounced role in distant work relations.Currently we plan to conduct
field studies at selected real-world enterprises and contrast our findings with these
studies.
Acknowledgements
The authors are grateful for the technical and financial support from T-Mobile,
which made this research project possible.
References
Batagelj,V.,Mrvar,A.,1998.PAJEK—program for large network analysis.Connections 21,47–57.
Bellotti,V.,Bly,S.,1996.Walking away from the desktop computer distributed collaboration and
mobility in a product design team.Proceedings of the ACM1996 Conference on Computer Supported
Cooperative Work (CSCW‘96),Boston,MA,pp.209–218.
Brewster,S.,2002.Overcoming the lack of screen space on mobile computers.Personal and Ubiquitous
Computing 6 (3),188–205.
ARTICLE IN PRESS
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752750
B.ullingen,F.,W
.
orter,M.,2000.Entwicklungsperspektiven,Unternehmensstrategien und Anwendungs-
felder im Mobile Commerce,Wissenschaftliches Institut f.ur Kommunikationsdienste,Diskussionsbei-
trag Nr.208,Bad Honnef.
Cross,R.,Prusak,L.,2002.The people who make organizations go—or stop.Harvard Business Review
80 (6),5–12.
Cross,R.,Parker,A.,Prusak,L.,Borgatti,S.P.,2001.Knowing what we know:supporting knowledge
creation and sharing in social networks.Organizational Dynamics 30,100–120.
Dunlop,M.,Brewster,S.,2002.The challenge of mobile devices for human computer interaction.Personal
and Ubiquitous Computing 6 (4),235–236.
Fagrell,H.,Ljungberg,F.,2000.A field study of news journalism:implications for knowledge
management systems.In:Proceedings of the Sixth Biennial Participatory Design Conference,CPSR
and ACMPress,New York.
Fagrell,H.,Kristoffersen,S.,Ljungberg,F.,1999.Exploring support for knowledge management in
mobile work.In:Proceedings of the Sixth European Conference on Computer-Supported Cooperative
Work,Copenhagen,Denmark.Kluwer Academic Publishers,Dordrecht,pp.259–275.
F
.
arber,F.,Keim,T.,Weitzel,T.,2003.Enhancing partner matching with recommender systems.Working
Paper No.2003-8,Institute of Information Systems,Frankfurt University.
Fayyad,U.M.,1996.Advances in Knowledge Discovery and Data Mining.AAAI Press,Menlo Park,CA.
Gerpott,T.J.,2001.Wettbewerbsstrategische Gestaltunsgelder f.ur Mobilfunkbetreiber auf Mobile-
Business-M
.
arkten.Information Management & Consulting 16 (2),34–43.
Gl.uckler,J.,Schrott,G.,2003.Knowing the flows is knowing who grows.An experimental exploration of
the impact of physical,electronic and mobile communication flows on individual and team
performance—a network approach.Working Paper,University of Frankfurt.
GSM-World,2003.Today’s GSM Platform.http://www.gsmworld.com/technology/gsm.shtml,
2003-06-20.
Hyers,K.,2001.Mobile Commerce—Wireless’ Newest El Dorado,Report,Cahners In-Stat Group.
Ito,M.,2001.Mobile phones,Japanese youth and the replacement of social contact.Working Paper,
http://www.itofisher.com/PEOPLE/mito/mobileyouth.pdf,2003-06-10.
IZT,2001.Entwicklung und zuk.unftige Bedeutung mobiler Multimediadienste.Institut f.ur Zufunfts-
studien und Technologiebewertung,Sekretariat f.ur Zukunftsforschung,Institut f.ur Arbeit und
Technik,Werkstattbericht Nr.49,Berlin.
James,C.L.,Reischel,K.M.,2001.Text input for mobile devices:comparing model prediction to actual
performance.In:Proceedings of the SIGCHI Conference on Human Factors in Computing Systems,
Seattle,Washington,United States.pp.365–371.
Johansen,R.,1988.Groupware:Computer Support for Business Teams.Macmillan,New York.
Katz,J.,Aakhus,M.,2002.Perpetual Contact.Cambridge University Press,Cambridge.
Katzenbach,J.R.,Smith,D.K.,1993.The discipline of teams.Harvard Business Review 71 (2),111–121.
Krackhardt,D.,1987.QAP Partialling as a test of spuriousness.Social Networks 9,171–186.
Kollmann,T.,2001.Ist M-Commerce ein Problem der Nutzungsl.ucke?Information Management &
Consulting 16,59–64.
Malone,T.W.,Laubacher,R.J.,1998.The Dawn of the e-lance economy.Harvard Business Review76 (5),
144–150.
Mitchell,J.C.,1969.The concept and use of social networks.Mitchell,J.C.(Ed.),Social Networks in
Urban Situations.Analyses of Personal Relationships in Central African Towns.University Press,
Manchester,pp.1–50.
Mobasher,B.,Cooley,R.,Srivastava,J.,2000.Automatic personalization based on web usage mining.
Communications of the ACM 43/8,142–151.
M.uller-Veerse,F.,1999.Mobile Commerce Report.Durlacher Corp.
Pinelie,D.,Gutwin,C.,2000.A review of groupware evaluations.Proceedings of WETICE,2000,
Workshops on Enabling Technologies:Infrastructure for Collaborative Enterprises,IEEE Computer
Society,Gaithersburg,MD,June 2000,pp.86–91.
Reichwald,R.,M
.
oslein,K.,Sachenbacher,H.,Englberger,H.,2000.Telekooperation—Verteilte
Arbeits- und Organizationsformen.Springer,Berlin.
ARTICLE IN PRESS
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752 751
Reichwald,R.,Meier,R.,Fremuth,N.,2002.Die mobile
.
Okonomie und Spezifika.In:Reichwald,R.
(Ed.),Mobile Kommunikation.Wiesbaden,Gabler,pp.3–16.
Schmitzer,B.,Butterwegge,G.,2000.Mobile commerce.Wirtschaftsinformatik 42 (4),355–358.
Schoberth,T.,Schrott,G.,2001.Virtual communities.Wirtschaftsinformatik 43 (5),517–519.
T-Mobile,2002.Mobile Access Portal and Mobile Service Portal—Whitepaper.
Varshney,U.,1999.Networking support for mobile computingk.Communications of the Association for
Information Systems (CAIS) 1(1),Article 1.
Wasserman,S.,Faust,K.,1994.Social network analysis.Methods and Applications.Cambridge
University Press,Cambridge.
Wiberg,M.,Gr
.
onlund,A.,2000.Exploring mobile CSCW.In:Svensson,L.,Snis,U.,S^rensen,C.,
F
.
agerlind,H.,Lindroth,T.,Magnusson,M.,
.
Ostlund,C.(Eds.),Proceedings of IRIS,Vol.23.
Laboratorium for Interaction Technology,University of Trollh
.
attan Uddevalla.
Wiedmann,K.-P.,Buckler,F.,Buxel,H.,2000.Chancenpotenziale und Gestaltungsperspektiven des M-
Commerce.Der Markt 39 (153),84–96.
Wohlfahrt,J.,2001.One-to-one marketing immobile commerce.Information Management & Consulting
16(2),S.49–53.
Zobel,J.,2001.Mobile Business und M-Commerce,Munich.
ARTICLE IN PRESS
G.Schrott,J.Gl
.
uckler/Int.J.Human-Computer Studies 60 (2004) 737–752752