Information technology to support electronic meetings

Abstract

As managers spend more of their time in meetings, the study of information technology to support meetings becomes increasingly important. Several unique systems to support meetings electronically have been developed in industry and universities. The PLEXSYS systems at the University of Arizona have been operational since 1985 and are now being implemented in industrial sites. This article proposes and defines a new term for information technology systems that support group meetings: electronic meeting systems (EMS). EMSs are more than group decision support systems (GDSS): they support more tasks than just decision making; they focus on communication. They move beyond the GDSS decision room, where groups must meet at the same time in the same place, to meetings that can be conducted across time and space. The article then presents a model of the EMS concept, which has three components: group process and outcomes; methods; and environment. Each of these components is explained in turn, and the implications derived from their study to date are discussed. Finally, the implementation of information technology for meeting support and its use in corporate settings will be addressed, as it has implications for productivity, meeting size, group member participation, and the role of the IS department.

Full text

Electronic Meetings
Information
Technology to
Support Electronic
Meetings
Keywords: Electronic meeting systems, group
decision support systems, group proc-
ess and outcomes, methods, soft-
ware, environments
ACM Categories: K.0, K.m, H.0, H.4.2, H.4.m
By: Alan R. Dennis
Joey F. George.
Len M. Jessup
Jay F. Nunamaker, Jr.
Douglas R. Vogel
Department of MIS
College of Business and Public
Administration
The University of Arizona
Tucson, AZ 85721
Abstract
As managers spend more of their time in meet-
ings, the study of information technology to sup-
port meetings becomes increasingly important.
Several unique systems to support meetings elec-
tronically have been developed in industry and
universities. The PLEXSYS systems at the Uni-
versity of Arizona have been operational since
1985 and are now being implemented in indus-
trial sites. This article proposes and defines a
new term for information technology systems
that support group meetings: electronic meet-
ing systems (EMS). EMSs are more than group
decision support systems (GDSS): they support
more tasks than just decision making; they focus
on communication. They move beyond the
GDSS decision room, where groups must meet
at the same time in the same place, to meet-
ings that can be conducted across time and
space. The article then presents a model of the
EMS concept, which has three components:
group process and outcomes; methods; and en-
vironment. Each of these components is ex-
,plained in turn, and the implications derived
from their study to date are discussed. Finally,
the implementation of information technology for
,meeting support and its use in corporate set-
tings will be addressed, as it has implications
[for productivity, meeting size, group member par-
ticipation, and the role of the IS department.
Introduction
Managers and knowledge workers spend a sig-
nificant proportion of their time working in
groups. Estimates of this proportion range from
60-70% for information systems (IS) managers
to 30-80% for general managers (Hymowitz,
1988; Ives and Olson, 1981; Mintzberg, 1983;
Mosvick and Nelson, 1987). Unfortunately, most
group meetings are not as productive as they
could be (Goldhaber, 1974; Hymowitz, 1988;
Mosvick and Nelson, 1987; Tubbs, 1984). One
Fortune 500 company estimated that it lost $71
million each year due to ineffectively managed
meetings (Mosvick and Nelson, 1987). Yet, even
though significant advances have been made to
enhance individual productivity through the use
of information technology (IT), comparatively little
has been done to improve group productivity.
Recently, however, there has been rapidly grow-
ing interest in the use of IT to support meetings
(Richman, 1987). While the concept has been
discussed for several years (Huber, 1984; Keen
and Scott Morton, 1978), most early efforts to
develop systems to support meetings met with
limited success (Kraemer and King, 1986). One
of the most promising early efforts was led by
G.R. Wagner at Execucom (Gibson and Ludl,
1988; Kull, 1982), although the system did not
survive. The University of Arizona’s first system
to support meetings, developed as part of the
ongoing PLEXSYS project, integrated some of
Wagner’s ideas in its design and implementa-
tion. The Arizona facility, which was also based
on some of the work of Paul Gray (Gray, 1981;
1983), was fully operational in MaTch, 1985. At
thi3 same time, researchers at the University of
Minnesota also began to investigate the poten-
tial of systems to support meetings (DeSanctis
and Gallupe, 1985).
Several other university and industry groups
have developed IT-based systems to support
meetings and group work (e.g., EDS, MCC,
Xerox PARC, Claremont Graduate School).
MIS Quarterly/December 1988 591

Electronic Meetings
Such systems can be grouped into two broad
classes: group decision support systems (GDSS)
and computer-based systems for cooperative
work (CSCW) (Figure 1). The distinction
tween these two types of systems is in the pri-
mary type of group support they each were de-
signed to provide. GDSSs are more task-
oriented in that they provide a means for a group
to work on and complete a task, such as reach-
ing a decision, planning, or solving problems.
CSCWs, on the other hand, are more driven by
communication needs. They provide a means
for small groups to communicate more efficiently,
enabling them to jointly create or critique a docu-
ment, for example. The distinctions between
these two classes of systems are blurring. Some
software tools developed as part of the
PLEXSYS project exemplify the common area
between GDSS and CSCW. In time, we believe,
these two classes of systems will completely over-
lap, representing a single class of IT systems
to support electronic meetings.
The purpose of this article is to propose a con-
ceptual model of IT-based systems to support
meetings, based on experiences in the
PLEXSYS project. (See Appendix A for a brief
history of the PLEXSYS project, and Appendix
B for a description of the PLEXSYS environ-
ments). First, in light of the blurring of the dis-
tinctions between GDSS and CSCW, and in light
of the changing technological focus from com-
puters to IT in general, a new term is proposed
and defined: information technology to support
electronic meetings, or electronic meeting sys-
tems (EMS), for short. Then a model of the EMS
concept is presented. The model is divided into
three component parts: group process and out-
comes; methods; and the EMS environment.
Each of these components is then discussed
separately, beginning with a conceptual model,
moving to a discussion, and ending with impli-
cations. The article ends with a discussion of
the broader implications that can be drawn from
the discussion of the EMS concept.
Group Decision
Support Systems
GDSS CSCW Computer-Supported
Cooperative Work
Group
Support Systems
GSS CSS
Collaboration
Support Systems
Information Technology to Support Meetings
Figure 1. The Progression to EMS
592 MIS Quarterly~December 1988

Electronic Meetings
Definition of EMS
Webster defines a meeting as, "an act or proc-
ess of coming together." The definition does not
imply that only one type of task is performed
in a meeting (e.g., decision making), nor does
it imply that the people participating in the meet-
ing must come together in a central location at
a specific time. It says only that they come to-
gether. Since the definition of a meeting is so
broad, it makes sense to use a term that
matches the definition to designate the informa-
tion technology systems that support meetings.
The term we propose is electronic meeting sys-
tems (EMS).
EMSs are systems that use information technol-
ogy to support the group work that occurs in
meetings. EMSs combine the task-orientation of
GDSS and the communication-orientation of
CSCW. GDSS has been defined as an inte-
grated computer-based system to facilitate the
solution of an unstructured or semi-structured
task by a group that has joint responsibility for
performing it (DeSanctis and Gallupe, 1985);
EMS is that and more. It also enhances com-
munication among group members (Bui, 1987;
Bui and Jarke, 1984; DeSanctis and Gallupe,
1987; Huber, 1984; Kraemer and King, 1986;
Loy, et al., 1987). EMS provides an additional
communication channel; it enhances communi-
cation by adding structure, either implicitly or ex-
.plicitly; and it can provide a complete recording
of the group session to aid productivity in sub-
sequent sessions (Karon, 1987). EMS also
moves beyond the decision-making function im-
plied in the term GDSS, since meetings involve
more than just decision making; for example,
they can also involve problem structuring, idea
generation, idea organization, planning, creat-
.ing, and even the elicitation of knowledge in the
construction of expert systems (Nunamaker, et
,al., 1988b). In addition, EMS moves beyond the
decision room, where groups must meet at the
,same time in the same place, to meetings that
san be conducted across time and space. Group
,members can be located in different places and
in different times, yet still work together to ac-
,’omplish some common purpose.
Therefore, we define EMS as:
An information technology-based environ-
ment that supports group meetings, which
may be distributed geographically and tem-
porally. The IT environment includes, but
is not limited to, distributed facilities, com-
puter hardware and software, audio and
video technology, procedures, methodolo-
gies, facilitation, and applicable group
data. Group tasks include, but are not lim-
ited to, communication, planning, idea gen-
eration, problem solving, issue discussion,
negotiation, conflict resolution, systems
analysis and design, and collaborative
group activities such as document prepa-
ration and sharing.
EMS Concept
There are three parts to the EMS concept: group
process and outcomes, methods, and environ-
ment (Figure 2).
As a category, group process and outcomes en-
compasses several different constructs. These
include the characteristics of the group itself, the
characteristics of the task on which the group
is working, the organizational context in which
system use takes place, the process through
which the group utilizes the system, and the out-
comes resulting from system use. These ele-
ments of the larger concept can be grouped into
a research model (Figure 3) that shows the re-
lationships of these constructs to each other.
Past GDSS and CSCW research has concen-
trated on these constructs, and the model is an
attempt to synthesize past work as well as pro-
vide a model for future research within this area.
The EMS concept goes beyond the constructs
listed above; it must necessarily include meth-
ods and environments.
The second component of the larger concept is
methods. On one level, methods are the soft-
ware support provided in EMS, which can be
thought of as the discrete tools provided for.
group support. Examples include electronic brain-
storming, electronic notepads, and alternative
ranking .tools. Methods, however, are more than
just the programs that the term software implies,
because certain procedures, rules and method-
ologies are built into the software. Consider, for
example, an electronic brainstorming tool in an
EMS environment. The purpose of the tool is
to allow group members to freely generate ideas.
In order for the group to use the tool well, the
procedures that promote idea generation and
sharing must be built into the system. The meth-
odology of manual brainstorming must be trans-
ferred to the EMS tool. In some systems, a
human facilitator is also present to provide addi-
MIS Quarterly~December 1988 593

Electronic Meetings
MethodsEnvironment
Group Process and Outcomes .
Figure 2. Conceptual Model
tional methodologies as needed. Where appli-
cable, this human facilitation would also be a
part of methods.
The third component of EMS is the environments
in which the systems are used. Many people
think of GDSS in terms of a room with networked
workstations and public displays, where a single
group can convene face-to-face to hold a meet-
ing, but this is only one of many possible EMS
environments (shown later in Figure 8). Groups
can meet at the same time yet be physically dis-
persed, as is common in many CSCW systems.
In addition, group members can work together
from their own offices asynchronously. Other
EMS environments can support several differ-
ent groups, each distributed temporally and geo-
graphically. EMS moves beyond the limitations
of the meeting room to support meetings across
time and place. This is not to say that the GDSS
room has outlived its usefulness; there will
always be a need for such a facility, but EMSs
facilitate group support in many different
environments.
As Figure 2 indicates, there are areas of over-
lap between the three components. For exam-
ple, some of the procedures and methodologies
thought of as being part of methods might also
be part of group process and outcomes. Struc-
ture beyond that provided by the software might
be provided by group members as they work.
Likewise, there might be overlaps between meth-
ods and environment, and between environment
and group process and outcomes. These over-
laps should be expected since it takes all three
components to make an EMS; no such system
is developed one component at a time. As each
component is developed, it must necessarily
take aspects of the other two components into
consideration, and this produces the overlapping
areas in Figure 2. The center of the diagram,
where all three areas overlap, represents those
parts of the overall EMS that are well-integrated,
forming the heart of a well-designed EMS.
When any two of the three components has
been specified, the third component is con-
strained by the limits of the other two. There
is still some flexibility, but for the most part, the
model has only two degrees of freedom. For ex-
ample, if a particular group working on a par-
ticular task in a particular organizational context
has decided to utilize the EMS approach, and
it is further decided that the EMS environment
to support them will allow them to work in their
own offices asynchronously, then this group has
little choice of the methods they will use. The
methods will necessarily be devoid of active
human facilitation, and the task, the group, and
the physical and temporal dispersion of the
group will necessitate the use of certain tools
594 MIS Quarterly~December 1988

Electronic Meetings
Group
¯Individual Member
Characteristics
¯Group Size
¯ History
¯FormaHnformal
¯ Ongoing/One Time
¯Experience
¯Cohesiveness
¯ etc.
Task
¯ Type of Task
(eg. judgemental)
¯ Rational/Political
¯ Complexity
¯ etc.
Context
¯ Incentives and
Reward Systems
¯ Organization Culture
¯ Environment
¯ etc.
Process
¯ Degree of
Structure
¯# of Sessions
¯Anonymity
¯Leadership
¯Participation
¯Conflict
¯Non-Task Behavior
¯ etc.
Outcome
¯Satisfaction with
Process & Outcome
¯ Outcome Quality
¯Time Required
¯# of Alternatives
¯# of Comments
¯Consensus
¯ Confidence
¯ etc.
EMS
¯Presence/Absence
of EMS Tools
¯Methods Design
¯Environment Design
¯ etc.
Figure 3. A Research Model
with certain built-in procedures. If this same
group was to decide on the methods to use
rather than the particular environment to work
in, then the choice of environment would be lim-
;:$d. A more detailed discussion of each of these
three components follows.
Group Process and
Outcome
The first part of the EMS concept is group proc-
ess and outcome. As mentioned previously, this
part of the concept has been dealt with most
MIS Quarterly~December 1988 595

Electronic Meetings
in the empirical GDSS literature to date. The meth-
ods and the environments have generally been
perceived as givens, but they vary widely across
many dimensions of methods and environments,
as shall be seen in the following sections.
This section proposes a research model that can
be used as a basis for empirical studies of EMS.
The model is based on those variables and
classes of relationships important to the group
process and (~utcomes with EMS: group char-
acteristics, task, context, environment, group proc-
ess, and process outcomes. The relevant GDSS
empirical literature is then reviewed. A full review
of the empirical GDSS and CSCW literature is
beyond the scope of this article, so it is confined
to the GDSS literature. Since we see GDSS as
a part of the larger EMS concept, the implica-
tions for group process and outcomes that can
be drawn from a review of the relevant GDSS
literature also apply to EMS in general.
A Research Model
There is no commonly accepted causal model
for studying group process and outcomes in
GDSS, although several researchers have pre-
sented conceptual ideas based on theory and
observation (DeSanctis and Gallupe, 1987;
Huber, 1984; Kraemer and King, 1988). Many
studies have used the work of McGrath (1984)
on small groups as the beginning basis for such
models (Gallupe, et al., 1988; Watson, et al.,
1988; Zigurs, 1987). In general, each research
study or research group has developed a micro
model as the basis for one study or for a pro-
gram of research.
Figure 3 displays an integration of many of the
models used to conduct GDSS research. It is
necessarily incomplete, as there are far more
factors affecting meetings than can be repre-
sented in one diagram. However, it does pre-
sent some of the variables and relationships con-
sidered in past research, as well as those that
should be addressed in future research. The vari-
ables in the model are representative of those
variables studied most often in past GDSS and
computer-mediated communication research.
The classes of variables are discussed first, fol-
lowed by a discussion of the proposed relation-
ships among them.
Variables
The model has six basic sets of variables. First,
the characteristics of the group, such as group
size and group proximity (whether in one room
or distributed in several remote locations), and
past experience with the problem area, such as
group process and tools, must be considered.
The characteristics of the individual participants,
group cohesiveness and motivation, past group
history, and future relationships have also been
shown as important in studies of meetings, so
they should be considered in the study of EMS.
The exact type of task is very important to group
performance (Poole, et al., 1985), so any study
of performance must clearly define the nature
of the task performed. The second class of vari-
ables, then, deals with the task faced by the
group. One way to characterize the task is along
the "rational/irrational" dimension. Also, task com-
plexity can be measured by the number of
issues and alternatives that must be considered
and the time required to identify and assess the
issues and alternatives (Hackman, 1968; Shaw,
1932; Shaw, 1973). But categorizing task type
can be much more complex. McGrath (1984)
has developed a taxonomy of eight group tasks,
which provides a more precise method for ana-
lyzing and discussing tasks.
Third, the larger context in which the group meet-.
ing occurs (such as organizational or experimen-
tal situations) the larger environment, and indi-
vidual incentive system are important (DeSanctis
and Poole, 1987; Jessup, 1987).
Fourth, the presence or absence of an EMS,
plus the specific characteristics of EMS design
will have an impact on the group process and
outcome. There are many different types of EMS
and many different designs within each particu-
lar type; this variation may be important in ex-
plaining differences in reported outcomes.
Fifth, the nature of the group process, such as
the presence or absence of a formal or informal
group leader, the use of anonymity, the number
of meeting sessions, the degree of structure in
the group process, equality of participation, level
of conflict, and the level of non-task ("uninhibi-
ted") behavior must be considered.
Finally, there are many outcomes of a group meet-
ing that may be measured. These include the
decision / outcome quality, participant satisfaction
with the outcomes and the process, participant
confidence in the outcomes, process time re-
596 MIS Quarterly~December 1988

Electronic Meetings
quired, level of group consensus, number of com-
ments during the meeting, and the number of
alternatives or issues considered.
Relationships
Many early studies of GDSS used very simple
models to study the effects of GDSS use on meet-
ings. Independent variables such as GDSS use
were hypothesized to directly affect dependent
variables such as performance, with no inter-
vening variables. These models are straightfor-
ward and easy to use, and represent a reason-
able model of the underlying relationships. In
Figure 3, these relationships are represented by
the arrows running from the four left boxes to
the outcome box.
More recently, researchers have begun using
more complex models in an attempt to better
understand the actual relationships involved. For
example, DeSanctis and Poole (1987) propose
a model of adaptive structuration, in which out-
comes depend on the process in which the
system is used. This process is in turn depend-
ent to some extent on the group, task, context,
and environment. This is incorporated into Figure
3 by the intervening position of the process box
between the four left boxes and the outcome
box.
Group processl then, can be either a dependent
or independent variable, depending on the re-
search design. For example, many system tools
provide anonymity, while traditional manual tech-
niques do not. Therefore~ anonymity is a depend-
ent variable since it is dependent on the pres-
ence or absence of the tool. In other cases,
anonymity has been an independent variable,
since non-anonymous system tools were com-
pared to anonymous tools.
Research Findings:
Overview
It is tempting to analyze all past GDSS research
dealing with group process and outcome as a
single unified body of literature. Some may argue
there is so little past empirical GDSS research
that it makes little sense to do otherwise. How-
:.,er, after additional consideration, it becomes
vident that there are several different but re-
lated streams of empirical research. This re-
.;earch has been conducted in both the labora-
tory and in the field, and it has involved two
types of GDSS: Local Area Decision Nets
(LADN) and Decision Rooms. LADNs are char-
acterized by small group size, physically dis-
persed group members, and synchronous (or
"real time") exchange (see Figure 8). Decision
Rooms are characterized by small group size,
having the group members together in the same
room, and synchronous exchange.
Under the broader label of experimental re-
search, at least four streams of research can
be identified, as depicted in Figure 4. They com-
pare: LADNs to Decision Rooms; LADNs to no
computer support; Decision Rooms to no com-
puter support; and two different configurations
of the same Decision Room. The field research
we reviewed has been confined to the use of
Decision Rooms by real groups. In the next sec-
tion, the four streams of experimental research,
then the field work are discussed. Because work
in this area is so recent, some of the work re-
ferred to is reported in working papers and not
yet published.
Experimental studies
Local Area Decision Nets and Decision
Rooms
Three studies have compared LADN and Deci-
sion Room GDSS (Bui, et al., 1987; Gallupe and
McKeen, 1988; Jessup, et al., 1988) (Table 1).
Each used a different GDSS system and slightly
different experimental design. The Bui, et al.
(1987) study compared the use of GDSS by proxi-
mate (face-to-face) and dispersed group mem-
bers. Jessup, et al. (1988) started with the same
design, but added the dimension of anonymity.
Gallupe and McKeen (1988) compared two dif-
ferent systems -- a GDSS and a computer-
mediated conferencing (CMC) system -- in proxi-
mate and dispersed settings. In these latter two
studies, proximate groups were more satisfied
with the group process than were dispersed
groups.
Local Area Decision Nets
There has been a substantial body of work on
cross-media comparisons of computer confer-
encing (CC), led by researchers at the New
Jersey Institute of Technology (Hiltz, et al., 1986;
¯ Turoff and Hiltz, 1982) and at Carnegie Mellon
University (Kiesler, et al., 1984; Sprague, 1980).
MIS Quarterly~December 1988 597

Electronic Meetings
1. LADN vs. Decision Room
oB
\
Bn
80
8n
/
2. LADN vs. No Computer Support
3. Decision Room vs. No Computer Support
08
\ /
4. Decision Room Configuration
oB
nB
~8
\
8o
Bo
Bo
80
/
Figure 4. Current Research Streams
598 MIS Quarterly~December 1988

Electronic Meetings
Computer conferencing fits nicely into the LADN
categorization of GDSS, as shown by the groups
using these systems in experiments at the above
institutions: the groups were small, they worked
at the same time, and each group member was
isolated. Table 2 reports some of the findings
for four of the five experiments reported in Hiltz,
et al. (1986), Kiesler, et al. (1984), Siegel,
al. (1986), and Turoff and Hiltz (1982). These
experiments do not constitute all the studies
done in this area, but they are representative
of this body of work. (Full consideration of this
literature is, however, beyond the scope of this
article.) In general, the results of such experi-
ments suggest that groups using computer con-
ferencing (or LADN), in comparison to conven-
tional face-to-face groups (FTF), generate
decisions of equal quality, are less likely to reach
consensus, take longer to reach a group deci-
sion, are more likely to participate equally, and
are more likely to engage in non-task behavior
such as "flaming," although Turoff and Hiltz
(1982) found face-to-face groups more likely
engage in tension release behavior.
Table 1. Experimental GDSS Research: Decision Rooms and Disaersed Groups
COMPARING GDSS DECISION ROOMS TO LOCAL AREA DECISION NETS
Variables/ Number of Solution Decision
Studies Solutions Quality Speed Satisfaction
Bui, et al., no effect dispersed dispersed no effect
1987 groups better groups faster
Jessup, et al.,
1988
Gallupe &
McKeen,
1988
most in anonymous/
dispersed; least in
identified/proximate
no effect GDSS took
longer than
CMC; dispersed
took longer
than proximate
proximate groups
more satisfied; most
satisfied groups in
anonymous/
dispersed and
identified/proximate
no effect for GDSS;
dispersed groups
less satisfied
Table 2. Experimental GDSS Research: Local Area Decision Nets
GDSS LOCAL AREA DECISION NETS VS. NO COMPUTER SUPPORT
Variables/ Decision Time to Non-Task
Studies Quality Consensus Decision Participation Behavior
Siegel, et al., CMC* groups CMC groups CMC groups
1986, exp. 1 took longer more equal less inhibited
Siegel, et al., CMC groups CMC groups CMC groups
1986, exp. 3 took longer more equal less inhibited;
e-mail groups
Jess so
Turoff and no effect less likely no effect more tension
Hiltz, 1982, in CC groups release in FTF*
exp. 1 no reportTuroff and
Hiltz, 1982,
exp. 2
leader alone
or computer
feedback alone
more likely to
lead to consensus
*CMC = computer mediated communication; FTF = face-to-face.
MIS Quarterly~December 1988 599

Electronic Meetings
Decision Rooms Compared to No Computer
Support
The largest body of GDSS research to date is
concerned with comparing the use of a Deci-
sion Room to no computer support. In many of
these studies, groups receiving no computer sup-
port either use the same structure as the GDSS
groups, or they use no structured processes at
all, The various studies, along with some of their
findings, are listed in Table 3 and discussed
below. The most obvious generalization that can
be made from looking at Table 3 is that the re-
sults from these studies are inconsistent.
Table 3 illustrates the three most investigated
dependent variables in these 10 studies: quality
of decision, level of participation, and satisfac-
tion with the group process. The findings for
these variables are inconsistent across all 10
studies. Quality of decision was rated better in
GDSS groups than in non-GDSS groups in five
of the 10 studies, while four studies found GDSS
group decisions to be at least as good as those
made by non-GDSS groups. Use of the GDSS
had no effect on the level of participation of
group members in four of the seven studies that
reported results about participation, but produced
more even levels of participation in the other
three studies. Four of the seven studies that
measured satisfaction with the group process
found that GDSS users were no more and no
less satisfied with the process than were group
Table 3. Experimental GDSS Research: Decision Rooms
GDSS DECISON ROOMS VS. NO COMPUTER SUPPORT
Variables/ Decision
Studies Quality Consensus
Steeb and GDSS
Johnston, better
1981
Lewis, GDSS
1982 better
Ruble, no effect
1984
Gallupe, GDSS
et al., better
1988
Beauclair, no effect
1987
Watson, GDSS worse no effect
et al., than manual:
1988 better than
,nothing
Zigurs, GDSS
1987" better
A. Easton, no effect
19881 no effectG. Easton,
1988
da~enpaa,
et al.,
19885
EBB first,
workstation
2nd and
cony. last
less likely
in GDSS
Time to
Decision
GDSS
takes
longer
GDSS
takes
longer
no effect
no effect
faster in
FTF
SatisfactionSatisfaction
Participation Inhibition w/Process w/Outcome
no report increased increased
w/GDSS w/GDSS
GDSS no effect
reduces
individual
dominance
no effect reduced reduced
by GDSS by GDSS
no effect no effect
reduced
by GDSS
more even
distribution
of influence
no effect no effect
more equal
in GDSS no effect
no effect
no effect
no effect
GDSS more
satisfied
*This study, while a cross-media comparison, focused on process ~ather than outcomes.
lIn this study, structured approaches, whether automated or not, led to better quality decisions, which took longer to
make, had higher user satisfaction with outcomes and processes, and had more equally distributed participation.
:I:"EBB" stands for Electronic BlackBoard, with no other computer support. "Workstation" means a G DSS with networked
workstations, with no other computer support. "Conv." stands for conventional, meaning no computer support.
600 MIS Quarterly~December 1988

Electronic Meetings
members that did not use a GDSS. One of the
studies found higher levels of satisfaction in
GDSS groups and two found lower levels of
satisfaction.
The other four dependent variables listed in the
table have been investigated in half or fewer of
the studies. Three of the five studies that looked
at time to decision found that GDSS users took
longer to reach a decision. The other two stud-
ies found no differences. The four studies that
measured satisfaction with outcomes also had
mixed results: two found higher levels of satis-
faction, one found lower levels, and one found
no differences. The two studies investigating con-
sensus produced inconsistent findings as well:
one found no effect and the other found con-
sensus less likely among GDSS groups. And fi-
nally, the only study out of these 10 that investi-
gated "flaming" found that there were no
differences between GDSS and non-GDSS
groups in the number of uninhibited comments
they produced.
One area of possible research that has been
neglected to date is a comparison of key attrib-
utes of different GDSS Decision Rooms. Deci-
sion Rooms differ across several attributes, such
as architectural design, room configuration,
public display capabilities, and system software.
Using similar groups of subjects and a similar
task, how would two different GDSS Decision
Rooms compare on such outcome measures as
decision quality and satisfaction with the proc-
ess? This seems to be an overlooked area,
which has a great deal of promise, but which
would require cooperation among researchers
operating distinctive Decision Rooms.
Comparing Different Configurations of the
Same Decision Room
The final stream of experimental GDSS research
to be considered is concerned with the compari-
Table 4. Experimental GDSS Research: Within
son of different configurations of the same Deci-
sion Room. These studies vary one or more fea-
tures available in the Decision Room to further
the understanding of when certain features are
appropriate and when they are not. Only two
such studies have been conducted (Connolly,
et al., 1988; Jessup, et al., 1987), and they are
listed with some of their findings in Table 4. Both
studies used the University of Arizona’s
PLEXSYS system, and both studies varied the
anonymous feature of PLEXSYS (see appendi-
ces). Some groups tagged all their comments
with their names (identified) whereas other
groups sent and received comments that were
not tagged (anonymous). Both studies found that
anonymous GDSS groups generated signifi-
cantly more total comments, as well as more
critical comments.
Anonymity is only one of the many features of
PLEXSYS that could have been varied. Other
systems likewise have many features that can
be varied and tested in experimental situations.
This is an area full of potential for GDSS
researchers.
Case studies and field studies
The primary methodology currently used to evalu-
ate GDSS is experimental research. There have
been a few studies, however, that have used
the case study or field study methodology. For
the purposes of this article, a GDSS case study
involves a "real world" group using a GDSS at
the GDSS site, away from their usual operating
location. Generally, these GDSSs are located on
the premises of a university, although some are
commercial products. A field study, on the other
hand, involves the study of a "real world" group
using a GDSS at a GDSS site that is on the
premises of their usual operating location. For
clarity, we classified the Jarvenpaa, et al. (1988)
study as an experiment, even though it was con-
Decision Rooms
COMPARING DIFFERENT CONFIGURATIONS OF A SINGLE DECISION ROOM
Variables/
Studies
Jessup, et al.,
1987
Connolly,
et al.,
1988
Total
Number of
Comments
more with
anonymity
more with
anonymity
Number of
Unique
Solutions
more in
critical
groups
Number of
Critical
Comments
more with
anonymity
more with
anonymity
Number of
Supportive
Comments
no effect
more with
anonymity
Overall
Satisfaction
higher for
supportive
groups
MIS Quarterly/December 1988 601

Electronic Meetings
ducted in the field, and it will not be discussed
further in this section.
Table 5 lists five studies that utilize the case and
field study methodologies. The first four are case
studies, as defined above, and the last is a field
study. (Two other papers sometimes considered
as reports on field studies actually are not: Kull
(1982) reports on a GDSS simulation, and Ker-
sten (1985) reports on the use of a GDSS
a course environment). With the exception of the
first one, all of these studies were conducted
in the PLEXSYS environment. The case studies
were conducted at the University of Arizona, and
the field studies were conducted on-site in a
manufacturing plant.
All six studies report that "real world" users were
extremely satisfied with GDSS use. This finding
is not consistent with experimental findings on
satisfaction, where some subjects were satisfied
with GDSS use and others were not. One ex-
planation for this may be the ability of real world
users to compare GDSS use with conventional
means of accomplishing the same tasks, an abil-
ity student subjects in experimental studies may
lack. The other consistent finding, which varies
only in particulars, is that real world participants
.also judge GDSS use to be extremely effective
in helping them perform the tasks they are work-
ing on. Two of these studies. (Nunamaker, et al.,
forthcoming 1989, Vogel and Nunamaker, 1988),
for example, report vast time savings from GDSS
use over conventional means. This finding also
varies from experimental findings, where task per-
formance varies widely. An explanation could lie
in the fact that participants in case and field situ-
ations are working on solving their own prob-
lems instead of problems assigned to them by
researchers. Also, tasks dealt with in the field
are generally more complex than those dealt
with in the laboratory, and, as such, are more
illustrative of computer support benefits. Groups
:in the field tend to bring together different facets
!of domain knowledge that cumulatively yield a
compret~ensive picture of a complex area ex-
ceeding the capabilities of any individual group
member.
Implications From Group
Process and Outcome
Two generalizations become readily apparent
from reviewing the GDSS literature on group proc-
ess and outcomes: (1) very little formal empiri-
cal work has been done in the area, and (2)
many of the results from the work that has been
done are inconsistent.
So little work has been done comparing LADNs
to Decision Rooms that generalizations from the
work are not very meaningful. The same is true
of the work comparing different configurations
of the same Decision Room. These are, how-
ever, very important areas that GDSS research-
ers should actively pursue. Considerably more
research has been done on LADNs, but as these
systems move beyond simple messaging sys-
tems (e.g., e-mail), opportunities arise for re-
search into the effects of more sophisticated
LADNs on groups.
As discussed earlier, very few generalizations
can be made reliably from reviewing the experi-
mental work that compared Decision Room use
to no computer support for groups. At best
GDSS use is associated witt~ better quality deci-
sions than no GDSS use (Gallupe, et al., 1988;
Jarvenpaa, et al., 1988; Lewis, 1982; Steeb and
Johnston, 1981; Zigurs, 1987), and at worse,
there is no difference (Beauclair, 1987; Easton,
A., 1988; Easton, G.K,, 1988; Ruble, 1984). At
best, GDSS use facilitates more even levels of
participation among group members (Easton,
G.K., 1988; Lewis, 1982; Zigurs, 1987), and at
worst, there are no differences (Beauclair, 1987;
Easton, A., 1988; Gallupe, et al., 1988; Jar-
venpaa, et al., 1988). Findings on the other de-
pendent variables are either inconsistent or are
based on too few studies to mention.
It is important to observe, however, that the 10
studies that compared Decision Rooms to no com-
puter support were conducted using seven dif-
ferent GDSSs. Each GDSS facility was designed
based on a different philosophy, and the soft-
ware used in each facility also varied widely. Dif-
ferent tasks, as well as different measures of
the dependent variables, were used in the stud-
ies. There is so much variation across these stud-
ies that generalizations become problematic. A
coherent body of knowledge concerning GDSS
can be accumulated if researchers share soft-
ware designs, experimental tasks, and variable
measures in the future. In the meantime, it is
important to the accumulation of knowledge
about GDSS that researchers describe in detail
the GDSS, task, procedures, and measures they
use in their studies, This permits other research-
ers to better understand past work and to better
plan future studies. The resulting accumulation
602 MIS Quarterly/December 1988

Electronic Meetings
Table 5. GDSS Research: Case and Field Studies
CASE AND FIELD STUDIES IN GDSS RESEARCH
Observations/
Studies Satisfaction Effectiveness
Adelman, 1984 final design well supported by action taken within week
all participants after GDSS exercise
Nunamaker, et al., participants reported high levels more equal participation
1987
Vogel & Nunamaker, 1988 participants reported high levels participants said they did as much
in one morning as would have
normally taken two days
Dennis, et al., 1988 participants reported high levels meetings rated extremely effective
by management and participants
Nunamaker, participants reported high levels found manhour savings of 61%
et al., 1989 from GDSS use, compared to
unsupported sessions
of knowledge will apply to and aid in the devel-
opment of EMS.
Research findings are much more consistent in
the field studies reviewed in this article. "Real
world" users are consistently satisfied with the
group process, and they believe GDSS use to
be very effective. The implication is that there
are fundamental differences between how
GDSSs are studied in the laboratory and ap-
plied in the field, and the effects they have on
group processes and outcomes. Careful analy-
sis is necessary to isolate and understand these
differences.
One explanation for the rather mixed success
of GDSS in experimental tests, as contrasted
with its success in field studies, is the size of
the groups and the complexity of the tasks.
GDSS technology has some overhead cost. Pre-
vious experimental research focused on smaller
groups (typically three or four members) and less
complex tasks than those typically found in field
settings (with groups of seven to ten or larger).
In these cases, the overhead costs (or "process
losses") introduced by the specific GDSS system
may simply have been higher than the marginal
benefits provided to small groups addressing
less complex tasks. GDSS may prove most ap-
propriate in the support of large groups address-
ing complex tasks.
Methods
Methods are a key component of the EMS con-
cept. As discussed earlier, methods include soft-
ware and the procedures and methodologies
built into the software. Methods can also include
the efforts of the human facilitator. The purpose
of this section is to describe different types of"
methodological support that can be provided
through EMS. First, a typology of EMS methods
is described. Then the concept of an EMS toolkit
is presented. The focus of this section is on char-
acteristics of EMS methods that address the
needs of groups responsible for complex deci-
sions in which each member has a role and is
an active participant in the group’s discussions.
Methods developed for computer conferencing
or e-mail applications are not considered.
EMS methods typology
Methods developed for EMS can be classified
according to a number of different schemes. We
have identified three dimensions where they
vary: (1) whether support is provided for the fa-
cilitator only, for participants only, or for both;
(2) whether group processing is sequential
parallel; and (3) whether the methods support
single or multiple group sessions. The first two
of these dimensions are represented in the
matrix of Figure 5. For clarity, the third dimen-
sion, single or multiple session support, has
been omitted from the figure. The primary differ-
ence between single or multiple session sup-
port is the use of a knowledge base and other
technological support to facilitate integration and
use of information across sessions and between
groups.
Support for facilitators only can occur in either
a sequential processing or parallel process-
ing mode, but parallel processing for facilitators
MIS Quarterly/December 1988 603

Electronic Meetings
Processing Mode
Sequential Parallel
IT
Support
Provided
For
Facilitator
Only
Participant
Only
Facilitator
and
Participant
Figure 5. EMS Methods Typology
only is rare, and there may be few instances
where such support would be useful. The facili-
tator is the only person receiving IT support, and
parallel processing would imply having the facil-
itator run several processes simultaneously.
Much more common is sequential processing sup-
port for facilitators. This type of support is the
only kind available in single-workstation systems,
where the facilitator enters comments and solu-
tions that group members have generated
through conventional means. Groups members
generally follow the process through watching
their comments appear on a public display. Such
support is also available in multiple-workstation
systems. In such systems, the facilitator guides
the group through a complex process, such as
consolidation of ideas generated during a brain-
storming session. Group members may watch
on a public display, or video switching may be
used so they can watch on individual user
screens.
Support for participants only requires a multi-
ple-workstation EMS. Participants are allowed to
work from their individual workstations, doing
whatever they want when they want. No facilita-
tor is required for them to use the EMS. They
may work sequentially or they may work to-
gether simultaneously, in effect processing in par-
allel. In the first situation, the methodology can
be designed so that users must take turns gen-
erating comments, which are displayed on a
public screen. In the second situation, the meth-
odology is designed so that all participants can
enter comments at the same time. The methods
determine the order in which comments will be
displayed. While parallel processing is often
more efficient, depending on the number of par-
ticipants, the number and complexity of the com-
ments may soon become overwhelming unless
appropriate measures are taken to manage the
process.
Support for both a facilitator and participants
also requires a multiple-workstation EMS. Proc-
essing may be sequential or parallel. An ex-
ample of sequential processing is a Nominal
Group Technique tool, After each participant gen-
erates his or her own list of comments (which
is actually parallel processing), the facilitator con-
trols the process by which separate comments
are presented to the group. Participant support
allows each group member to choose which com-
ment to present to the g~oup at a given time.
604 MIS Quarterly~December 1988

Electronic Meetings
An example of parallel processing is brainstorm-
ing. The facilitator support again provides a
means to control the process, and participant
support allows each group member to generate
and distribute comments simultaneously.
Methods that support participants only, or both
the facilitator and the participants, have been
developed from two complementary but quite dis-
tinct perspectives or philosophies. The first,
which underlies the designs of many CSCW sys-
tems, typically sees a group as a small number
(e.g., three or four) of tightly knit co-workers with
a common sense of purpose (e.g., Stefik, et al.,
1987). Group members are seen as very coop-
erative (e.g., working together on a ~)roposal
mutually authored document). Changes can be
quickly made in terms of editing the work of
others, and authorship is generally recognized.
Essentially, there is a sense of member equal-
ity, with electronic interaction often accompanied
by a high degree of verbal exchange.
The second perspective, which is how some
GDSSs have been designed, views a group in
a task force context. In this case, the group is
larger (e.g., 12 to 24), often with subgroups. The
group typically has a common sense of purpose
and culture but members are not necessarily as
cooperative and egalitarian as those in a small
group of co-workers. The environment is often
politically charged, with many personal opinions,
agendas, and vested interests present. Tasks
are complex, with the necessity to garner input
from a variety of perspectives and member knowl-
edge domains (e.g., for corporate resource allo-
cation, planning, or negotiation). Anonymity may
be important to draw out true feelings, voting
is commonplace, and support for a facilitator as
well as participants is likely to exist. In large
groups, sequential processing can be less ef-
fective as either the opportunity for equal par-
ticipation is removed or as each person has less
time in which to contribute.
These perspectives are not mutually exclusive.
In reality, groups, whether they are ongoing, dis-
cretionary coalitions, or formed for an explicit pur-
pose, exhibit a variety of characteristics as a func-
tion of task and member characteristics. Further,
an individual member’s behavior may vary sub-
stantially from group to group or session to ses-
sion or even within a session. Equifinality is likely
to exist, i.e., there are a number of ways to sup-
port a particular group. Though methods devel-
oped from either perspective are often used for
similar situations, the EMS approach seeks to
flexibly capture aspects of both of these philo-
sophical approaches.
All of the above method types can be used to
support a single group in a single meeting. With
the addition of a knowledge base and other
types of technological support, the same types
of software can provide support for multiple
group sessions. Such methods are increasingly
emerging to facilitate integration of information
across multiple sessions and between groups.
Comprehensive communications support is pro-
vided with a variable and dynamic degree of struc-
ture based on intra- as well as inter-session at-
tributes. Support for integrating information
across sessions and between groups includes
knowledge bases with "intelligent" access
through easy-to-use interfaces. Additional
access is provided to organizational and exter-
nal information resources to support dynamic in-
tegration of relevant information.
Particular attention should be given to seamless
integration between multiple session support
methods and other organizational information
functions (e.g., teleconferencing, computer con-
ferencing, scheduling, and e-mail). These are the
types of things a focus on EMS makes possible..
The methods increasingly support i’decision
rooms without walls" in which organizational mem-
bers can be participating in group sessions with-
out the need to be continuously present in a
single room. Many of the capabilities described
in this article are only beginning to surface in
some GDSSs but will become standard in fully
evolved EMS systems.
EMS toolkit
Many early GDSSs were task-driven, as defined
by Huber (1984). They were designed to meet
the needs of one group performing one task,
and therefore addressed one, and only one, ap-
plication of meetings. As discussed in DSS lit-
erature, these systems were specific GDSS ap-
plications (Sprague, 1980). For example, one
early GDSS was designed specifically to assist
in labor-management negotiation and could not
be used for any other task (Kersten, 1985).
More recently, the need to provide a toolkit, simi-
lar to the concept of a DSS model base or tool
set (Sprague, 1980), has become apparent.
Toolkits are collections of specific tools that ad-
dress various parts of the meeting’s process. In
MIS Quarterly~December 1988 605

Electronic Meetings
the same manner as hammers and wrenches
are used for different tasks, so are the specific
tools provided in the toolkit. A toolkit can con-
ceivably include tools of each type described
above. EMS environments using toolkits are ac-
tivity driven (Huber, 1984), i.e., they have com-
ponents to support specific group activities (such
as idea generation and voting), rather than one
indivisible system to support the entire process
of one meeting application (such as decision
making or negotiation).
The key advantage provided by toolkits is flexi-
bility. This flexibility is important in three ways.
First, each tool in the toolkit will have its own
meeting dynamics. One too} in the toolkit may
support a highly structured interchange of ideas,
while another tool may provide very little struc-
ture. Groups can choose which tool they prefer.
Second, groups use many processes to achieve
their goals; they often do not proceed in a straight-
forward manner to reach their goals (Bahl and
Hunt, 1984). The tools in the toolkits can easily
be mixed and matched, and used in whatever
order the group believes is most effective to
achieve its goals. Finally, the toolkit is also suffi-
ciently flexible to enable new tools to be easily
added. The flexibility of the toolkit approach is
illustrated in Figure 6, in which each tool or
group of tools is represented as a node in a
network. Users may begin at any node and
move to any other node in any order. Some of
the tools depicted in Figure 6 are described
below.
Examples of tools in the PLEXSYS toolkit
include:
¯Session director -- guides the facilitator or
group leader in selection of the tools to be
used in a session and generates an agenda.
Default times and output reports are listed and
may be modified at the group’s.discretion.
¯Electronic brainstorming -- supports idea
generation, allowing group members to simul-
taneously and anonymously share comments
on a specific question.
¯Issue analyzer-- helps group members iden-
tify and consolidate key focus items resulting
from idea generation. Support is also provided
for integrating external information to support
identified focus items.
Voting -- provides a variety of prioritizing meth-
ods including Likert scales, rank ordering, and
multiple choice. All group members cast pri-
vate ballots. Accumulated results are
displayed.
Topic commenter -- supports idea solicita-
tion and provision of additional detail in con-
junction with a list of topics. Each topic may
",,, /
Figure 6. Network of Tool Use
606 MIS Quarterly/December 1988

Electronic Meetings
have subtopics. Participants enter, exchange,
and review information on self-selected topics.
¯Policy formation -- supports the group in de-
veloping a policy statement or mission through
iteration and group consensus.
¯Organizational infrastructure-- provides sup-
port for capturing characteristics of organiza-
tional data sets, information systems, and struc-
ture to provide a foundation for impact
analysis.
¯ Stakeholder identification and assumption
surfacing -- is used to systematically evalu-
ate the implications of a proposed policy or
plan. Stakeholder assumptions are identified,
scaled, and graphically analyzed.
¯Alternative evaluator- provides multi-
criteria decision-making support. Alternatives
can be examined under flexibly weighted cri-
teria to evaluate decision scenarios and
tradeoffs.
The choice of tool will dramatically affect the meet-
ing process and therefore the outcome of the
meeting. Likewise, selecting the best combina-
tion of tools -- the meeting agenda -- is also
crucial. For each stage in the process, the group
can select one tool from a set of possible tools,
depending on which specific group technique it
wishes to use. One of the most important activi-
ties of any group process is the premeeting plan-
ning. Key tasks during this step are setting the
objectives for the meeting, ensuring that partici-
pants understand these objectives and the roles
they will play, and designing the agenda to meet
these objectives. These activities can be sup-
ported by an agenda tool. The agenda tool may
be a simple DSS, or an expert system could
be built by using decision rules for setting meet-
ing agendas. Alternately, setting the agenda
could be sufficiently complex to require a sepa-
rate planning meeting supported by EMS.
The organization of the tools to support group
processes and tasks is facilitated in the system
Figure 7. PLEXSYS
MIS Quarterly~December 1988 607

Electronic Meetings
architecture illustrated in Figure 7. As the figure
shows, the output from the tools serves as input
to a knowledge base that provides a mechanism
for representing and storing the planning knowl-
edge using a variety of knowledge representa-
tion techniques that include semantic inheritance
networks, frames, and production rules (Apple-
.gate, et al., 1987; Mclntyre, et al., 1987). The
knowledge base approach facilitates multiple plan-
ning and decision process representations. The
representations can change dynamically as new
knowledge is added to the system. The knowl-
edge base acts as a "meeting memory" as
groups return for additional sessions and new
members or groups seek to build on the output
from previous sessions.
Impfications from methods
On one level, the EMS environment determines
the methods that can be used. For example, if
the environment involves a single workstation,
then the methods will be limited to facilitation-
only support (where one individual enters all per-
tinent information from the group session) or to
specific models or sets of models that can be
run with data collected from the group during
its meeting. Once the environment expands
beyond one workstation, to the point where each
individual in the group has access to a worksta-
tion, then the methods that can be used change.
Group members can now enter data and com-
ments simultaneously, enabling the use of tools
such as electronic brainstorming.
Methods and tool development are ongoing proc-
esses. Maintenance is critical to the success of
any EMS, but maintenance is extremely labor
intensive and therefor expensive. The develop-
ment and continuing support of EMS methods
represent a considerable investment of labor,
time and capital.
Environments
Several authors have presented alternative clas-
sifications of GDSS. DeSanctis and Gallupe
(1985) originally presented a framework using
the dispersion of group members and duration
of the decision-making session, which was later
revised (DeSanctis and Gallupe, 1987) to use
group member proximity (i.e., dispersion) and
group size. Kraemer and King (1986; 1988) de-
veloped a more detailed classification that ex-
amines the hardware, software, orgware, and
people aspects of GDSS. Burns and colleagues
(Burns, et al., 1987; Rathwell and Burns, 1985;
Thomas and Burns, 1982) introduced the con-
cept of distributed decision making, where sev-
eral groups interact and exchange information
asynchronously. Hale and Haseman (1987) used
a similar approach by categorizing GDSS on two
dimensions: local vs. distributed, and length of
the decision process, whether of limited dura-
tion or ongoing duration. Jelassi and Beauclair
(1987) categorized GDSS for small groups
three dimensions: face-to-face vs. non-face-to-
face, group member proximity, and synchronous-
asynchronous.
The next section of this article presents a cl,assi-
fication of EMS environments based on the tax-
onomies discussed above. Each EMS environ-
ment is then discussed individually, followed by
a discussion of important practical considerations
in EMS environment design. The section ends
with a series of implications from EMS
environments.
Taxonomy of environments
There are three distinctive dimensions that can
be integrated into a taxonomy of EMS environ-
ments: group size, participant location, and the
timing of the "meeting" (whether it is one or
more sessions, or a series of asynchronous
group exchanges)(Figure 8).
Group size is a relative concept. Most research-
ers would probably agree that a group of three
or four members is small, while a group of 20
or more is large. But, beyond this, general agree-
ment on what is "large" and what is "small" or
even "medium" is difficult. In this article, we con-
sider small groups to have 10 members orless
and large groups to have more than 10
members.
Group proximity refers to one "logical" group,
in the sense that all participants address the
same task. Not all participants need to be pre-
sent in the same physical location (i.e., part of
one physical group). Group proximity has three
levels that relate to group geographic dispersion.
With the first case -- multiple individual sites
-- the individual members of the group are physi-
cally separate, working in their individual offices
or workstations. In the second case -- one
group site -- all members of the group are physi-
cally together in one room. In the last case --
multiple group sites -- members of the group
meet in separate locations in subgroups, and
then these multiple subgroup meetings are in-
terconnected via EMS.
608 MIS Quarterly/December 1988

Electronic Meetings
Group
Size
Small
Large
Net
Computer
Conference Legislative
Session
Conference
EMS
Tele-
Conference/
Broadcast
Multiple
Individual
Sites
One
Group
Site
Multiple
Group
Sites
All Meet
at One
Time
Asynchronous
j "Meetings"
Time
Dispersion
Group Proximity
Figure 8. Taxonomy of Envionments
Regarding time, groups may meet synchro-
nously (i.e., at the same time) or asynchronously
(i.e., at different times). With EMS, traditional
meetings, in the sense that participants actually
meet at the same place and time, are no longer
the only choice. Electronic mail and computer
conferencing are examples of information tech-
nologies that enable individuals to move beyond
the traditional limitations of time and space, but
since much present use of these ITs support
individual work, they lie outside the EMS area
and beyond the scope of this review (e.g., Hiltz
and Turoff, 1981). Where these two ITs directly
support meetings, they can be considered EMS.
Providing examples of asynchronous work in the
group environment beyond those mentioned
above is more difficult, since few such environ-
"nents presently exist. Burns and colleagues
’,Burns, et al., 1987; Rathwell and Burns, 1985;
Fhomas and Burns, 1982) provide a few exam-
~les where geographically dispersed groups can
~eet at separate times (perhaps due to time
.,one requirements). Each group works as a unit
~_nd then transmits its results to the other groups
’or consideration in their subsequent sessions.
~ery little research has been done in the area
of asynchronous group "meetings," but there is
potential for creative applications of this tech-
nology. In the following discussions of the EMS
environments, the issue of asynchronous group
work will not be addressed explicitly.
EMS environments
There are six basic categories of EMS environ-
ments, which can be used synchronously or
asynchronously.
Decision Room
Presently, the most common form of the organ-
izational meeting is one where a small group
of participants meets together in one place at
one time. This type of meeting can be supported
MIS Quarterly~December 1988 609

Electronic Meetings
by a Decision Room environment. The Decision
Room typically contains a series of networked
computer workstations, plus wide screen com-
puter video projection screen(s) for public view-
ing of group information. As the group meets
face-to-face at the same place and time, verbal
communication is available in addition to elec-
tronic communication.
Legislative Session
A legislative session differs from a Decision
Room only in size; it accommodates a larger
group. While verbal communication is still pos-
sible with a large group, it is less effective. Either
the opportunity for equal participation of all group
members is removed, or, if equal participation
occurs, participants have far less time in which
to communicate their ideas and opinions than
they would in an equivalent small group meet-
ing. Therefore, in a legislative session, electronic
communication, and the methods to support it
become more important.
©
EMS Teleconference / Broadcast
For a large group meeting at several sites, the
EMS support is similar. However, the purpose
of the meeting might be different. For example,
organizations sometimes use teleconferencing
facilities to broadcast courses or special pres-
entations from leaders to geographically dis-
persed parts of the organization. With EMS,
these broadcasts can be supplemented with an
electronic communication channel to enable two-
way communication between the multiple group
sites.
Local Area Decision Net
A Local Area Decision Net is used to support
a small group of dispersed individuals working
at different sites (such as their offices). While
equivalent facilities to EMS teleconferences can
be provided to each individual, the cost may be
prohibitive. More likely, video and voice com-
munication channels would be omitted, leaving
the group to rely on electronic communication.
EMS Teleconference
When a small group meets in several separate
group sites at the same time, EMS teleconfer-
encing facilities could be used. This is similar
to non-EMS supported teleconferences, but with
the addition of information technology to facili-
tat~ communication. At each site, the participants
have the same technology as is available in a
Decision Room, but with the addition of a net-
work to support electronic, voice, and video com-
munication between the different sites.
Computer Conference
In the context of EMS, computer conferencing
differs only in size from a Local Area Decision
610 MIS Quarterly~December 1988

Electronic Meetings
Net. Unlike more traditional forms of computer
conferencing using a bulletin board-style inter-
face, EMS computer conferencing also includes
synchronous communication among group mem-
bers, where participants can simultaneously read
and write messages. Because computer confer-
encing involves a large group, the methods used
to manage the group interaction once again
become very important.
EMS environment design
considerations
Most systems developed to support meetings
exist in one of three environments: Local Area
Decision Net, Decision Room, or legislative ses-
sion. PLEXSYS is currently being used in the
latter two (Appendix B). The lessons learned
from the design and use of these environments
are relevant to EMS design and are discussed
in the next sections under the headings of floor
plans, public information display, workstation
design, ergonomics, and support issues
(George, et al., 1988).
Floor Plan
Floor plan considerations include group size, fa-
cilitation/leader focus, group member line of
sight, and communication issues including rec-
ognition of verbal as well as electronic exchange.
The setting should be multi-purpose and flexible
to better meet the needs of different group sizes
and task environments. Opportunities to work in
dyads on workstations should be supported. Seat-
ing should be arranged to provide opportunity
for face-to-face interchange among the partici-
pants as well as facilitation/leader focus. Group
member line of sight should permit easy view-
ing of front screen images as well as communi-
cation with other group members. Recognizing
the need for verbal and electronic communica-
tion in the room is an additional consideration
particularly in terms of technology noise
suppression.
Public Information Display
Public information display issues include front
screen type and number, information format
(e.g., text or graphics) multi-media presentation
support, teleconferencing concerns, and provi-.
sion of an electronic podium. Presentation media
should provide a wide range of support and
should not restrict decision-maker communica-
tion. Front screen projection of individual work-
station screens and consolidation group infor-
mation, audio and video recording, optical disk
technology, electronic blackboards, and over-
head projection systems all play a role in pro-
viding a full measure of presentation media sup-
port. Particular attention should be given to
aspects of presentation support that might ad-
versely affect decision-maker comfort and par-
ticipation (e.g., poor screen legibility). An elec-
tronic podium is especially useful to coordinate
multi-media presentations through a user-
friendly interface.
Workstation Design
Workstation design issues include layout of mi-
crocomputer/screen, space for ,papers/other
work, software distributed on each workstation
hard disk, and local area net (LAN) handling
voice, video, and gateways. Providing an elec-
tronic interface for each group member encour-
ages all group members to participate and en-
hances the efficiency of that participation.
Relying on a single workstation for the whole
group is not always appropriate. If multiple work-
stations are used, each should have a high
degree of local intelligence and "in residence"
software options that provide end-user help, data
capturing, and communications support capabili-
ties. As such, each group member can maintain
some independence while contributing to the
group as a whole through an interface that is
flexible and individually supportive as well as in-
tegral to the networked system. A high band-
width local area network (LAN) is necessary
maintain high levels of performance to accom-
modate network demands in transmitting text as
well as screens between individual decision
makers. Our experience has shown that users
expect to receive subsecond response for all ac-
tivities. Gateways provide group decision-
making support for situations in which the deci-
sion makers are geographically dispersed.
Ergonomics
Ergonomics issues include characteristics of the
room, such as heat, sound, lighting conditions,
and seating, where the group decision making
takes place. Ignoring the impact of the setting
may destroy the very nature of the fragile envi-
ronment in which successful group decision
making is facilitated. Aesthetics that provide a
MIS Quarterly~December 1988 611

Electronic Meetings
measure of executive appeal in terms of com-
fort and familiarity allow decision makers to
better focus on issues at hand. Carpeting, wall
coverings, and furniture appropriate to organiza-
tional conference rooms provide a setting where
decision makers can comfortably relate to com-
plex organizational questions.
Support Issues
Support issues include breakout and conference
rooms, high quality printers and copiers, gallery
seating for observers, handicap access, secu-
rity, redundancy in hardware and easy data re-
covery, observation rooms, and time stamping
of voice and data in conjunction with video
taping. Breakout rooms adjacent to the main con-
ference area or other ways to logically partition
a larger group into small discussion groups are
particularly useful in some situations, as is a
small conference room for preplanning or ses-
sion staging purposes. Fast hardcopy printout
capabilities provide additional user support. Gal-
lery seating is useful for observers as well as
facilitation helpers and executive support staff
personnel. Security and reliability are constant
concerns. Observation rooms and time stamp-
ing of audio, video, and data for subsequent play-
back and analysis assist researchers to better
understand the implications of technology as-
sisted groups.
Implications from EMS
environments
The design of the EMS environment can affect
the process and outcomes of a group meeting.
In the same way that simple design factors (such
as color) have been shown to affect human in-
formation processing, complex design factors
(such as workstation design or the design of the
public information display) can be expected to
affect meetings. While it is likely that several en-
vironment designs may be equally appropriate,
the design of the EMS environment must be com-
patible with the methods with which it will be
used. For example, the use of anonymity has
been shown to have an impact on the process
and outcome of meetings (Connolly, et al.,
1988). While anonymity is a feature provided by
EMS methods, it is rendered ineffective if par-
ticipants can observe the workstation screen(s)
of another participant(s).
In the design of EMS environments, the need
to facilitate communication among group mem-
bers is key. There are three communication chan-
nels that need to be effectively integrated, al-
though not every channel will necessarily be
provided in every environment. First, participants
have access to the electronic communication
channel via a computer workstation, which pro-
vides individual access to information. This
access is used to enter and retrieve information
to support the individual’s contribution to the
group. Public display of information in the elec-
tronic channel is also typical (either via a large
screen projection system or an on-screen public
window at the individual’s workstation), which en-
ables the group as a whole to focus on specific
pieces of information.
Second, the environment may support verbal com-
munication, either through face-to-face commu-
nication of group members in the same room,
or via a microphone/speaker voice channel for
groups members not in the same room. While
the microphone / speaker channel at first appears
to have the most critical design issues, it is im-
portant not to overlook the effects of workstation
placement, for example, in a Decision Room en-
vironment. Providing a spacing between work-
stations of 2 feet as compared to 5 feet can
affect the ease and manner of verbal communi-
cation in the group.
Third, provision for video or "sight" communica-
tion may also be useful. Once again, while the
design issues for non-face-to-face environments
may be technically challenging, it is important
not to overlook those of face-to-face environ-
ments, especially legislative sessions. Partici-
pants should be able to easily see each other
and any public information display.
In addition, certain design issues are extremely
important to the development of a GDSS Deci-
sion Room that is useful to real world groups.
As a minimum, the overall facility design, multi-
ple public screens, central file servers, and com-
munication network speed must be considered.
The overall facility design includes aesthetics,
lighting and the physical organization of the De-
cision Room. The room must be physically or-
ganized for flexibility (to accommodate groups
of various sizes) and to facilitate the use by
group members of various communication chan-
nels, from electronic to verbal. The ability to
show the screens of individual workstations on
a large-screen projector is important, as is the
612 MIS Quarterly~December 1988

Electronic Meetings
presence of more than one public screen. In ad-
dition, knowledge bases and databases, handled
by central file servers, facilitate coordination and
management of input from individual group mem-
bers and serve as "organizational" memory from
one session to another.
One current challenge in EMS research is to
move beyond Decision Rooms, legislative ses-
sion environments, and Local Area Decision
Nets into the development of EMS teleconfer-
encing and computer conferencing environ-
ments, i.e., "Decision Rooms without walls."
Such environments have the potential to reduce
the need for all participants to be physically pre-
sent at the same place for meetings, and thus
may substantially reduce travel time and costs
for geographically dispersed groups.
Huber (1988) suggests that with more produc-
tive meetings, fewer meetings are expected per
project in an EMS supported organization envi-
ronment. This may result in a decline in the total
number of meetings in organizations, or, as meet-
ings become more productive, an increase.
While these person-hour savings are dramatic,
the elapsed times from project initiation to pro-
ject completion were reduced even more sub-
stantially b from several months to several
weeks in some cases. This has even greater
implications for organizational productivity. By re-
ducing the elapsed time from project identifica-
tion to completion, organizations can be more
responsive and more productive. Problems can
be resolved faster, and market opportunities can
be analyzed and acted on before competitors
are aware of them.
Implications
What are the implications of EMS technology for
organizations? Since few organizations have im-
plemented EMS to date, the answers to this ques-
tion are still unknown. However, the implemen-
tation of PLEXSYS in a multinational corporation
provides some evidence from which we can ex-
trapolate and make predictions. The patterns
that have started to emerge are addressed in
the following three sections: organizational pro-
ductivity, meeting size, and decision
participation.
Organizational productivity
One pattern that has emerged over the past year
from our study of the day-to-day use of
PLEXSYS in a major multi-national electronics
firm (Nunamaker, et al., 1989) is that PLEXSYS-
supported meetings are more productive than
similar meetings not supported by PLEXSYS.
More than 30 PLEXSYS-supported projects
were tracked from initiation to completion. All pro-
jects required substantially fewer meetings and
fewer person-hours than budgeted to complete.
While these observations may be due to inac-
curate a priori time budgeting or measurement
error, this organization has a history of accurate
forecasting and measurement. Conversely, the
improvement may also be due to the Hawthorne
effect; meetings were simply more productive be-
cause a change -- any change b was intro-
duced. However, since projects were observed
over a one-year period, this too is less likely.
Meeting size
Some observers (e.g., Huber, 1988) have
argued that the use of GDSS technology will de-
crease the number of people involved in future
meetings. Previous non-GDSS supported re-
search generally shows that small groups are
more effective and more satisfying to belong to
(Shaw, 1981), and, therefore, the increased pro-
ductivity introduced by GDSS will increase the
strength of the forces acting to promote smaller
groups. In contrast, our experience with
PLEXSYS shows that it supports and even pro-
motes larger groups in meetings (Dennis, et al.,
1988; Nunamaker, et al., forthcoming 1989).
There are three forces acting to increase the
size of group meetings. First, one may presume
that the issue to be addressed by the group is
one that could benefit from the increased domain
knowledge and skills provided by the members
in the group; otherwise, why form a group?
Huber (1984) points out that the business envi-
ronment is becoming increasingly complex,
which increases the need for specialized domain
knowledge and skills, and thereby increases the
desired size of the group.
Secondly, Ackoff (1981) argues that it is impor-
tant for those charged with executing a plan or
implementing a decision to understand why the
plan or decision was made. The best way to
do this is to include as many of these people
as possible in the group, a~ain increasing the
desired group size.
MIS Quarterly/December 1988 613

Electronic Meetings
Finally, there are political reasons for increasing
the size of the group. By including additional par-
ticipants in the decision-making group, their sup-
port is more likely to be gained for the decision
-- or at least the blame spread! Likewise, some
organizational participants may insist on being
present in meetings to ensure that their constitu-
encies are represented (in resource allocation
groups, for example).
Prior to the introduction of IT to support meet-
ings, research -- and experiences of managers
-- demonstrated the need to constrain the size
of group meetings to increase productivity. How-
ever, since PLEXSYS has shown the potential
to increase group productivity (Nunamaker, et
al., forthcoming 1989), the size of group
meetings can be expected to increase with EMS
use.
Decision participation
As the size of group meetings increases the meet-
ings have the potential to span several hierar-
chical levels in the organization (Dennis, et al.,
1988; Nunamaker, et al., forthcoming 1989).
Indeed, this has been one of the factors in
increasing productivity at the multinational firm
described above. Bringing all hierarchical levels
involved in the decision together in one meeting
can have several advantages, from getting faster
organizational approval for decisions to
improving organizational communication (Huber,
1988).
Better organizational communication occurs be-
cause senior management is more aware of day-
to-day issues, and employees and junior man-
agement are more aware of long term issues.
As one CEO explained three months after a
GDSS-supported strategic planning meeting: "A
lot of education happened that previously hasn’t
happened during one of these things... People
walked in with narrow perceptions of the com-
pany and walked out with a CEO’s perception"
(Dennis, et al., 1988, p. 16).
Since EMS can enable more organizational
levels to be represented in group meetings, it
is expected that more organizational levels will
be involved in the decision-making process. As
a result, organizational decision making could
become more participative. However, this does
not necessarily suggest that more decisions will
be made by groups rather than by individuals;
rather, individual decision makers might be more
inclined to solicit information and opinions from
a supporting group before making decisions.
Conclusions
Information technology support for meetings is
a relatively recent focus of study in the IS field,
but it is an area of great potential and opportu-
nity. The area is sufficiently broad to merit a label
that encompasses all of its major aspects, and
we have suggested electronic meeting systems
(EMS) as that term. EMS as a concept is a com-
bination of both GDSS and CSCW, stressing the
role of information technology instead of just com-
puters, the support of meetings across time and
space instead of just in one room at one time,
and the support of various tasks instead of just
decision making.
The EMS concept has three components: group
process and outcome, methods, and environ-
ments. In looking at group process and outcome,
we presented a research model for investigat-
ing EMS. The model illustrated the variables im-
portant to group process and outcome and how
they are related to each other. We also reviewed
the studies of GDSS that are relevant to EMS
and identified four streams of experimental re-
search as well as a recent focus on field stud-
ies. There remains too much variation across
studies to make many definitive st.atements, but
using IT to support meetings does seem to lead
to better quality decisions and more equal rates
of participation among group members, More
work needs to be done in this area, especially
since some findings, such as those dealing with
group satisfaction, consistently differ between ex-
periments and field studies. In general, however,
studies in this area do point to the potential use-
fulness of IT to improve meetings.
Our typology of EMS methods, which includes
software as well as the methodologies and pro-
cedures built into the software, categorized meth-
ods according to the support they provided. Meth-
ods can be designed to support the meeting
facilitator alone, the group alone, or both to-
gether. This support can allow sequential or par-
allel processing, and it can be used for a single
session or across multiple sessions. Providing
all of this support within a single EMS can be
done through the toolkit concept, which provides
maximum flexibility for supporting meetings. The
more varied the tools in the toolkit, however, the
greater the need for a premeeting planning
session.
614 MIS Quarterly~December 1988

Electronic Meetings
Finally, under EMS environments, we introduced
a taxonomy of 12 different environments that
differ in terms of group size, time, and the prox-
imity of group members. Even though 12 envi-
ronments are possible in this taxonomy, most
current environments are limited to Decision
Rooms, legislative sessions, or Local Area De-
cision Nets. From experience, we have found
that important design considerations for the
former two environments include floor plan,
public information display, workstation design, er-
gonomics, and support issues. For any environ-
ment, the need to facilitate communication is
key, and this can be done through providing sup-
port for three communication channels: elec-
tronic, verbal and visual.
It is difficult at this time to predict with any cer-
tainty how the implementation of EMS will affect
organizations. Based on the implementation of
PLEXSYS in a multinational corporation, how-
ever, we predict that EMS will be able to im-
prove organization productivity by decreasing the
number and duration of necessary meetings. At
the same time, the number of individuals in-
volved in a particular meeting can increase with-
out affecting the productivity of the meeting since
EMS can be designed to successfully support
large groups. Finally, EMS makes it possible to
broaden the scope of a meeting to include par-
ticipants from various hierarchical levels, thereby
improving organizational communication and fa-"
cilitating faster approval for decisions.
Acknowledgements
The PLEXSYS EMS has been developed over
the past years with the dedicated efforts of Linda
Applegate, David Chappell, T. Chen, Annette
Easton, Kimlynn Gridley, Bruce Hemiter, Benn
Konsynski, Simon Leung, Ben Martz, Mark Pen-
dergast, Ed Roberts, Bill Saints, Joe Valacich,
and Lee Walker. This research was partially sup-
ported by grants from the IBM Management Of
Information Systems program, and the Social
Sciences and Humanities Research Council of
Canada.
References
Ackoff, R.L. Creating the Corporate Future, John
Wiley & Sons, New York, 1981.
Adelman, L. "Real Time Computer Support for
Decision Analysis in a Group Setting: Another
Class of Decision Support Systems," Inter-
faces (14:2), March-April 1984, pp. 75-83.
Applegate, L., Chen, T.C., Konsynski, B.R. and
Nunamaker, J.F. Jr. "Knowledge Management
in Planning," Journal of Management Infor-
mation Systems (3:4), Spring 1987, pp. 4-38.
Bahl, H.C. and Hunt, R.G. "Decision-Making
Theory and DSS Design," Data Base (15:4),
Summer 1984, pp. 10-!4.
Beauclair, R.A. "An Experimental Study of the
Effects of Specific GDSS Applications on
Small Group Decision Making," working
paper, University of Louisville, Louisville, KY,
December 1987.
Blosser, P. An Automatic System for Applica-
tion Software Generation and Portability,
Ph.D. dissertation, Department of Computer
Science, Purdue University, West Lafayette,
IN, 1976.
Burns, A., Rathwell, M.A. and Thomas, R.C. "A
Distributed Decision-Making System," Deci-
sion Support Systems (3), 1987, pp. 121-
131.
Bui, T.X. "Co-op: A Group Decision Support
System for Cooperative Multiple Criteria De-
cision Making," Lecture Notes in Computer Sci-
ence Series: 290, Springer-Verlag, Berlin,
1987.
Bui, T. and Jarke, M "A DSS for Cooperative
Multiple Criteria Group Decision Making," Pro-
ceedings of the 5th International Conference
on Information Systems, Tucson, AZ, Novem-
ber 28-30, 1984, pp. 101-113.
Bui, T., Sivasankaran, T.R., Fijol, Y. and Wood-
burg, M.A. "Identifying Organizational Oppor-
tunities for GDSS Use: Some Experimental Evi-
dence," Transactions of the Seventh Interna-
tional Conference on Decision Suppor.t
Systems, San Francisco, CA, June 1987, pp.
68-75.
Connolly, T., Jessup, L.M. and Valacich, J.S.
"Idea Generation in a GDSS: Effects of Ano-
nymity and Evaluative Tone," working paper,
University of Arizona, Tucson, AZ, 1988.
Dennis, A.R., Heminger, A.R., Nunamaker, J.F.,
Jr. and Vogel, D.R. "Bringing Automated Sup-
port to Large Groups: The Burr-Brown Expe-
rience," working paper, University of Arizona,
Tucson, AZ, August 1988.
DeSanctis, G.L. and Gallupe, R.B. "Group Deci-
sion Support Systems: A New Frontier," Data
Base (16:2), Winter 1985, pp. 3-9.
DeSanctis, G.L. and Gallupe, R.B. "A Founda-
tion for the Study of Group Decision Support
Systems," Management Science (33:5), May
1987, pp. 589-609.
DeSanctis, G.L. and Poole, M.S. "Group Deci-
MIS Quarterly/December 1988 615

Electronic Meetings
sion Making and Group Decision Support Sys-
terns: A Three Year Plan for the GDSS Re-
search Project," MIS Research Center
Working Paper WP-88-02, University of Min-
nesota, Minneapolis, MN, September 1987.
Easton, A. An Experimental Investigation of Auto-
mated versus Manual Support for Stake-
holder Identification and Assumption Surfac-
ing in Small Groups, unpublished doctoral
dissertation, University of Arizona, Tucson, AZ,
1988.
Easton, G.K. Group Decision Support Systems
vs. Face-to-Face Communication for Collabo-
rative Group Work: An Experimental Investi-
gation, unpublished doctoral dissertation, Uni-
versity of Arizona, Tucson, AZ, 1988.
Gallupe, R.B., DeSanctis, G.L. and Dickson,
G.W. "Computer-Based Support for Group
Problem Finding: An Experimental Investiga-
tion," MIS Quarterly (12:2), June 1988, pp.
277-296.
Gallupe, R.B. and McKeen, J.D. "Beyond Com-
puter-Mediated Communication: An Experimen-
tal Study into the Use of a Group Decision
Support System for Face-to-Face versus
Remote Meetings," Proceedings of the ASAC
1988 Conference, Halifax, Nova Scotia, pp.
103-116.
George, J.F., Nunamaker, J.F. Jr. and Vogel,
D.R. "Group Decision Support Systems and
Their Implications for Designers and Manag-
ers: The Arizona Experience," Transactions
of the Eighth International Conference on De-
cision Support Systems, Boston, MA, June
1988, pp. 13-25.
Gibson, D.V. and Ludl, E.J. "Executive Group
Decision Support Systems Considered at
Three Levels of Analysis," Transactions of the
Eighth International Conference on Decision
Support Systems, Boston, MA, June 1988, pp.
26-38.
Goldhaber, G. Organizational Communication,
William C. Brown, Dubuque, Iowa, 1974.
Gray, P. "Initial Observations from the Decision
Room Project," Transactions of the Third In-
ternational Conference on Decision Support
Systems, Boston, MA, June 1983, pp. 135-
138.
Gray, P., Aronofsky, J.S., Berry, N., Helmer, O.,
Kane, G.R. and Perkins, T.E. "The SMU De-
cision Room Project," Proceedings of 2nd In-
ternational Conference on Information Sys-
tems, Cambridge, MA, December 7-9, 1981,
pp. 122-129.
Hackman, J.R. "Effects of Task Characteristics
on Group Products," Journal of Experimental
Social Psychology (4), April 1968, pp. 162-
187.
Hale, D.P. and Haseman, W.D. "EECS: A Pro-
totype Distributed Executive Communication
And Support System," Hawaii International Con-
ference on System Sciences, Kona, HI, Janu-
ary 1987, pp. 557-565.
Hiltz, S.R., Johnson, K. and Turnoff, M. "Experi-
ments in Group Decision Making Communi-
cation Process and Outcome in Face-to-Face
Versus Computerized Conference," Human
Communication Research (13:2), Winter
1986, pp. 225-252.
Hiltz, S.R. and Turoff, M. "The Evolution of User
Behavior in a Computerized Conferencing
System," Communications of the ACM
(24:11), November 1981, pp. 739-751.
Huber, G.P. "Issues in the Design of Group De-
cision Support Systems," MIS Quarterly (8:3),
September 1984, pp. 195-204.
Huber, G.P. "Effects of Decision and Communi-
cation Support Technologies on Organiza-
tional Decision Processes and Structures," Pro-
ceedings of IFIP WG 8.3 Working Conference
on Organizational Decision Support Systems,
North-Holland, Amsterdam, 1988, pp. 317-
333.
Hymowitz, C. "A Survival Guide to the Office
Meeting," Wall Street Journal, June 21, 1988,
p. 35.
Ives, B. and Olson, M. "Manager or Technician?
The Nature of the Information Systems Man-
ager’s Job," MIS Quarterly (5:4), December
1981, pp. 49-62.
Jarvenpaa, S.L., Rao, V.S. and Huber, G.P.
"Computer Support for Meetings of Medium-
Sized Groups Working on Unstructured Prob-
lems: A Field Experiment," working paper, Uni-
versity of Texas, Austin, TX, January 1988.
Jelassi, M.T. and Beauclair, R.A. "An Integrated
Framework for Group Decision Support Sys-
tems Design," Systems, Objectives, Solutions
in Information and Management (13:4), April
1987, pp. 143-153.
Jessup, L.M. "Group Decision Support Systems:
A Need for Behavioral Research," Interna-
tional Journal of Small Group Research, Sep-
tember 1987, pp. 139-158.
Jessup, L.M., Connoily, T. and Galegher, J. "The
Effects of Anonymity on Group Process in Auto-
mated Group Problem Solving," working
paper, University of Arizona, Tucson, AZ,
1987.
Jessup, L.M., Tansik, D.A. and Laase, T.Do
"Group Problem Solving in an Automated En-
vironment: The Effects of Anonymity and Prox-
616 MIS Quarterly/December 1988

Electronic Meetings
imity on Group Process and Outcome with a
Group Decision Support System," Proceed-
ings of the 1988 Annual Meeting of the Acad-
emy of Management, Anaheim, CA, August
1988.
Karon, P. "PCs Enter the Conference Room,"
PC Week, November 17, 1987, pp. 51, 60.
Keen, P.W. and Scott Morton, M. Decision Sup-
port Systems: An Organization Perspective,
Addison-Wesley, Reading, MA, 1978.
Kersten, G.E. "NEGO-Group Decision Support
System," Information and Management (8:5),
May 1985, pp. 237-246.
Kiesler, S., Siegal, J. and McGuire, T.W. "Social
Psychological Aspects of Computer Mediated
Communication," American Psychologist, Oc-
tober 1984, pp. 1123-1134.
Konsynski, B. A Model of Computer-Aided Defi-
nition and Analysis of Information System Re-
quirements, Ph.D. dissertation, Department of
Computer Science, Purdue University, West
Lafayette, IN, 1976.
Konsynski, B. and Nunamaker, J.F. Jr.
"PLEXSYS: A System Development System,"
reprinted in Advanced System Development/
Feasibifity Techniques, J.D. Couger, M.A.
Colter and R.W. Knapp (eds.), John Wiley and
Sons, New York, 1982.
Kraemer, K.L. and King, J.L. "Computer-Based
Systems for Cooperative Work and Group De-
cision Making: Status of Use and Problems
in Development," Proceedings of the 1986 Con-
ference on Computer Supported Collabora-
tive Work, Austin, TX, December 1986, pp.
353-375.
Kraemer, K.L. and King, J.L. "Computer-Based
Systems for Cooperative Work," Computing
Surveys (20:2), June 1988, pp. 115-146.
Kull, D.J. "Group Decision: Can Computers
Help?" Computer Decisions, May 1982, pp.
70 +.
Lewis, F.L. Facilitator." A Micro Computer Deci-
sion Support Systems for Small Groups, un-
published doctoral dissertation, University of
Louisville, Louisville, KY, 1982.
Loy, S.L., Pracht, W.E. and Courtney, J.F. Jr.
"Effects of a Graphical Problem-Structuring
Aid on Small Group Decision Making," Pro-
ceedings of the Twentieth Annual Hawafi In-
ternational Conference on System Sciences,
Kona, HI, January 1987, pp. 566-574.
McGrath, J.E. Groups: Interaction and Perform-
ance, Prentice-Hall, Englewood Cliffs, N J,
1984.
Mc/ntyre, S., Konsynski, B, and Nunamaker, J.F.
Jr. "Automated Planning Environments: Knowl-
edge Integration and Model Scripting," Jour-
nal of Management Information Systems (2:4),
1987.
Mintzberg, H. The Nature of Managerial Work,
Harper and Row, New York, 1983.
Mosvick, R.K. "Communication Practices of Man-
agers and Technical Professionals in Four
Large-Scale High Technology Industries," pre-
sented at the National Convention of the
Speech Communication Association, 1982.
Mosvick, R.K. and Nelson, R.B. We’ve Got to
Start Meeting Like This, Scott Foresman and
Co., New York, 1987.
Nunamaker, J.F. Jr. "A Methodology for the
Design and Optimization of Information Proc-
essing Systems," in Systems Analysis Tech-
niques, J.D. Couger and R.W. Knapp (eds.),
John Wiley and Sons, New York, 1974.
Nunamaker, J.F. Jr., Konsynski, B.R., Ho, T. and
Singer, C. "Computer Aided Analysis and
Design of Information Systems," Communica-
tions of the ACM, December 1976.
Nunamaker, J.F. Jr., Applegate, L.M. and
Konsynski, B.R. "Facilitating Group Creativity
with GDSS," Journal of Management Informa-
tion Systems (3:4), Spring 1987, pp. 5-19.
Nunamaker, J.F. Jr., Applegate, L.M. and
Konsynski, B.R. "Computer-Aided Delibera-
tion: Model Management and Group Decision
Support," Journal of Operations Research, No-
vember-December 1988a.
Nunamaker, J.F. Jr., Konsynski, B.R., Chen, M.,
Vinze, A.S., Chen, I.I. and Heltne, M.M. "Knowl-
edge-based Systems Support for Information
Centers, Journal of Management Information
Systems (5:1), Summer 1988b, pp. 4-24.
Nunamaker, J.F. Jr., Grohowski, R., Heminger,
A., Martz, B. and Vogel, D.R. "GDSS Experi-
ence at a Corporate Site," forthcoming in Pro-
ceedings of the Twenty-Second Annual
Hawaii International Conference on System
Sciences, Kona-Kailua, HI, forthcoming Janu-
ary 1989.
P0ole, M.S., Siebold, D.R. and McPhee, R.D.
"Group Decision-Making as a Structurational
Process," Quarterly Journal of Speech (7:1),
February 1985, pp. 74-102.
Rathwell, R.A. and Burns, A. "Information Sys-
tems Support for Group Planning and Decision-
Making Activities," MIS Quarterly (9:3), Sep-
tember 1985, pp. 255-271.
Richman, L.S. "Software Catches the Team
Spirit," Fortune (115:12), June 8, 1987.
Ruble, M.R. An Empirical Test of a Decision Sup-
port System in a Group Decision Making En-
vironment, unpublished doctoral dissertation,
MIS Quarterly~December 1988 617

Electronic Meetings
Arizona State University, Tempe, AZ, 1984.
Shaw, M.E. "A .Comparison of Individuals and
Small Groups in the Rational Solution of Com-
puter Problems," American Journal of Psychol-
ogy (44:3), July 1932, pp. 491-504.
Shaw, M.E. "Scaling Group Tasks: A Method
for Dimensional Analysis," JSAS Catalog of
Selected Documents in Psychology (3:8),
1973.
Shaw, M.E. Group Dynamics: The Psychology
of Small Group Behavior (3rd edition),
McGraw-Hill, New York, 1981.
Siegel, J., Dubrovsky, V., Kiesler, S. and
McGuire, T. "Group Processes in Computer
Mediated Communication," Organizational Be-
havior and Human Decision Process (37:2),
April 1986, pp. 157-187.
Sprague, R.H. "A Framework for the Develop-
ment of Decision Support Systems," MIS Quar-
terly (4:4), December 1980, pp. 1-26.
Steeb, R. and Johnston, S.C. "A Computer-
Based Interactive System for Group Decision-
making," IEEE Transactions on Systems, Man,
and Cybernetics (SMC-11:8), August 1981,
pp. 544-552.
Stefik, M., F(~ster, G., Bobrow, D.G., Khan, K.,
Lanning, S. and Suchman, L. "Beyond the
Chalkboard: Computer Support for Collabora-
tion and Problem Solving in Meetings," Com-
munications of the ACM (30:1), January 1987,
pp. 33-47.
Teichroew, D. and Hershey, E.A. "PSL/PSA: A
Computer-Aided Technique for Structured
Documentation and Analysis of Information
Processing Systems," in Advanced System De-
velopment/ Feasibifity Techniques, J.D.
Couger, M.A. Colter and R.W. Knapp (eds.),
John Wiley and Sons, New York, 1982.
Teichroew, D., Hershey, E.A. and Yamamoto,
Y. "The PSL/PSA Approach to Computer-
Aided Analysis and Documentation," in Ad-
vanced System Development/Feasibility Tech-
niques, J.D. Couger, M.A. Colter and R.W.
Knapp (eds.), John Wiley and Sons, New
York, 1982.
Teichroew, D. and Sayani, H. "Automation of Sys-
tems Building," in Systems Analysis Tech-
niques, J.D. Couger and R.W. Knapp (eds.),
John Wiley and Sons, New York, t974.
Thomas, R.C. and Burns, A. "A Case for Dis-
tributed Decision Making Systems, The Com-
puter Journal (25:1), January 1982, pp. 148-
152.
Tubbs, S.L. A Systems Approach to Small
Group Interaction, Addison-Wesley, Reading,
MA, 1984.
Turoff, M. and Hiltz, S.R. "Computer Support for
Group Versus Individual Decisions," IEEE
Transactions on Communications (30:1), Janu-
ary 1982, pp. 82-91.
Vogel, D.R. "The Impact of ’Messy’ Data on
Group Decision Making," Proceedings of the
Twenty-First Annual Hawaii International Con-
ference on Systems Science, Kona, HI, 1988.
Vogel, D.R. and Nunamaker, J.F. Jr. "Health Serv-
ice Group Use of Automated Planning Sup-
port," Administrative Radiology, September
1988.
Vogel, D.R., Nunamaker, J.F. Jr., Applegate,
L.M. and Konsynski, B.R. "Group Decision Sup-
port Systems: Determinants of Success," Trans-
actions of the Seventh International Confer-
ence on Decision Support Systems, San
Francisco, CA, June 1987, pp. 118-128.
Vogel, D.R., Nunamaker, J.F. Jr., George, J.F.
and Dennis, A.R. "Group Decision Support Sys-
tems: Evolution and Status at the University
of Arizona," Proceedings of IFIP WG 8.3 Work-
ing Conference on Organizational Decision Sup-
port Systems, North-Holland, Amsterdam,
1988, pp. 287-304.
Watson, R., DeSanctis, G. and Poole, M.S.
"Using a GDSS to Facilitate Group Consen-
sus: Some Intended and Unintended Conse-
quences," MIS Quarterly (12:3), September
1988, pp. 463°478.
Zigurs, I. The Effect of Computer Based Sup-
port on Influence Attempts and Patterns in
Small Group Decision-Making, unpublished
doctoral dissertation, University of Minnesota,
Minneapolis, MN, 1987.
618 MIS Quarterly~December 1988

Electronic Meetings
About
Alan R. Dennis is a doctoral student in MIS at
the University of Arizona. He received a bache-
lor of computer science degree from Acadia Uni-
versity and an MBA from Queen’s University in
Kingston, Ontario, and was a winner of the
AACSB National Doctoral Fellowship. Prior to en-
tering the Arizona doctoral program, he spent
three years as a faculty member of the Queen’s
University School of Business, and has pub-
lished several articles and book chapters. His
current research interests include electronic meet-
ing systems, decision support systems, and busi-
ness graphics.
Joey F. George is assistant professor of MIS
at the University of Arizona. He earned his bache-
lor’s degree at Stanford University in 1979 and
his Ph.D. in management at the University of
California at Irvine in 1986. His research focuses
on how computing affects work. His current re-
search investigates the effects of electronic meet-
ing systems (EMS) use on the group work proc-
ess and its outcomes, EMS use in non-American
cultures, the effects of computing on decision
authority in organizations, and the effects of ex-
tensive computerization in work groups.
Leonard M. Jessup is a doctoral candidate in
organizational behavior at the University of Ari-
zona, with a minor in management information
systems. He is a member of the Academy of
Management and the Institute of Management
Sciences. He received his B.A. in information
and communication studies, and his M.B.A. from
California State University, Chico. He has
worked as a community relations director in the
health care industry and as a sales engineer in
the Authors
the electronics industry. His research interests
include small group research, process analysis,
and automated group problem solving.
Jay F. Nunamaker, Jr., is head of the Depart-
ment of Management Information Systems and
is a professor of management information sys-
tems (MIS) and computer science at the Univer-
sity of Arizona. He received a Ph.D. from Case
Institute of Technology in systems engineering
and operations research. He was an associate
professor of computer science and industrial ad-
ministration at Purdue University. Dr. Nunamaker
joined the faculty of the University of Arizona
in 1974 to develop the MIS program. Author of
numerous papers on electronic meeting sys-
tems, the automation of software construction,
performance evaluation of computer systems,
and decision support systems for systems analy-
sis and design, he has lectured throughout
Europe, Russia, Asia, and South America. Dr.
Nunamaker is chairman of the Association for
Computing Machinery (ACM) Curriculum Com-
mittee on Information ,.£ystems.
Douglas R. Vogel is assistant professor of MIS
at the University of Arizona. He received his M.S.
in computer science from U.C.L.A. in 1972 and
his Ph.D. in MIS from the University of Minne-
sota in 1986, where he was also research coor-
dinator for the MIS Research Center. His cur-
rent research interests bridge the business and
academic communities in addressing questions
of the impact of management information sys-
tems on aspects of interpersonal communica-
tion, group decision making, and organizational
productivity.
MIS Quarterly/December 1988 619

Electronic Meetings
Appendix A
The History of Information Technology
to Support Meetings at the University of Arizona
The Systems Development Process
Each information technology system that has been developed to support meetings grew out of a unique
development project. An examination of each project would reveal different starting points for research.
An understanding of the historical starting points helps users and developers better understand the
system’s current state and underlying design. The systems developed under the PLEXSYS project
at the University of Arizona are no different. The purpose of this appendix is to provide the necessary
background information on the PLEXSYS project.
The underlying concept for PLEXSYS had its beginning in 1965 with the development of Problem
Statement Language/Problem Statement Analyzer (PSL/PSA) as part of the ISDOS (Information System
Design and Optimization System) project at Case Institute of Technology (Teichroew and Hershey,
1982). Nunamaker was involved in the project that led to PSL/PSA from its inception. The PSL/PSA
process started with the assumption that the requirements were known, or the individual or group
responsible for the systems building project was capable of stating the requirements. There was no
emphasis on developing an organizational consensus on the "correct" set of requirements, because
at the time, it was assumed that the systems analyst was in charge and would be able to satisfactorily
define the systems requirements. The emphasis on involving the user in requirements analysis was
not to develop for another ten years.
The collective wisdom of the ISDOS project at that time decided that it was more important to develop
methods to reduce the time to build a system, starting with "the assumption of correct requirements"
as a given. The rationale was that the "correct requirements" are not constant; they change with
changes in the organization. The users themselves change with respect to what they think they need
to do their job. The basic objective was to reduce the time from the initial statement of requirements
until the target system was operational. Automation or computer support was envisioned for each
task in the systems life cycle.
From this conceptual framework, developed by five doctoral students under the direction of Professor
Daniel Teichroew at Case Institute of Technology, evolved a number of software tools for automating
the systems building process. This approach utilizing computer support for the systems building proc-
ess resulted in PSL/PSA in 1965 (Teichroew and Hershey, 1982) and later in PLEXSYS
~ (Konsynski
and Nunamaker, 1982). In 1965-1968, three activities shaped the development of PSL/PSA and even-
tually the development of PLEXSYS: (1) The first version of the problem statement language and
problem statement analyzer was developed by Nunamaker (1974) as input to a computer-aided sys-
tems analysis and design software package called SODA (Systems Optimization and Design Algo-
rithm); (2) the prototype for the problem statement language was developed by John Paul Tremblay;
(3) the prototype for the problem statement analyzer was developed by Paul Stephan. These three
developments led to the PSL/PSA version, which was used by well over 100 organizations for docu-
menting and analyzing the set of requirements for an information system (Teichroew, et al., 1982).
PSL/PSA is a tool for describing requirements of a system, recording the descriptions in machine-
processable form, and storing them in a database (Figure la). With the PSL/PSA approach, data
expressed in a formal language called PSL. As PSL statements are entered into the database, PSA
analyzes the statements for correctness, completeness, and consistency with data and information
already present in the database. PSA then produces a set of reports that represent the combined
views of the many analysts or problem definers working on the requirements. These reports describe
PLEXSYS is derived from the word "plexus," which is defined by Webster’s as "an interwoven combination of parts or
elements in a structure or system." The "sys" in PLEXSYS is short for system.
620 MIS Quarterly~December 1988

Electronic Meetings
Commands in
Command Language
Statements Reports and
in Problem Messages
Statement °
Language ¯ Consistency
¯ Feasibility
(PSL)
PSA
DataBase
Dictionary
Figure la. PSL/PSA
the inputs, outputs and system flow along with system structure, data structure, data derivation, size,
volume and systems dynamics.
Next, Pat Blosser (1976) and Benn Konsynski (1976), doctoral students at Purdue University in
early 1970s, added procedural definitions to PSL to facilitate automatic code generation from PSL/PSA.
This served to facilitate code generation and moved the systems specification process further from
the user.
Nunamaker and Konsynski moved on to Arizona in 1974. During the process of using SODA/PSL,
SODA/PSA and ADS/PSA (Accurately Defined Systems) (early prototypes of PSL/PSA) on a large
project for the U.S. Navy, a change took place in their thinking (Nunamaker, et al., 1976). There
were problems in depending on end users to utilize a formal language for requirements specifications.
The end users at the Navy would not write their specifications in a PSL/PSA-like system, so an ac-
counting firm was hired to work with the end users and write the specifications in the language. Insights
gained from the deficiencies in this solution led to the development of the PLEXSYS concept. The
idea was to develop a phase that came before the use of PSL/PSA, i.e., develop software to assist
the users with the determination of requirements (Konsynski, 1976; Nunamaker, et al., 1988a). This
phase would help developers determine what was needed in addition to the software in order to de-
velop systems that would be used by the end users of the information system.
In many of the organizations Nunamaker and Konsynski worked with, the user group was represented
by a steering committee or task force consisting of 10-20 people. It became clear in 1979 that a
special meeting room was needed for the task force to use, or for the user group to meet to address
the information requirements of an organization. The function of the room would be to display the
system flows, data structures and information requirements on a large screen projection system and
permit each user seated at a workstation to interact with the set of requirements and the proposed
design of the system. The PLEXSYS-84 system, which was an extension of the PSL/PSA/ISDOS
MIS Quarterly~December 1988 621

Electronic Meetings
project, was a workbench~workstation environment for the system development team. A collection of
integrated tools, procedures, transformations, and models were available to the systems developer
to analyze and design systems. It was expected that PLEXSYS would shorten the life cycle of devel-
opment by facilitating a fast implementation of a prototype system. It was recognized that the design
process could not be completely automated and that PLEXSYS would be a computer-aided support
system with databases, knowledge bases, model management and inquiry facilities.
2
Appendix B
Arizona’s Decision Room Facilities
1st Facility u PlexCenter
Construction of the first computer-assisted group meeting facility at the University of Arizona began
in 1984. The facility, which opened in March 1985, was conceived as a meeting room for end users,
systems analysts, systems designers and project leaders to review and analyze system specifications
and designs. As usage of the system progressed over the first 18 months of operation we found
that the software was valuable in planning efforts of all types, not just information systems planning.
In fact, the usage of the room shifted from requirements and design review to initial discussion of
issues and problems. The participants in each session became the group responsible for decision
making regarding the organization’s goals and objectives relative to the task. The system was built
for one particular audience but was found to be useful in a broader context.
The first facility, called the PlexCenter, houses a large U-shaped conference table with 16 computer
workstations (Figure lb). Each workstation is recessed for line-of-sight considerations and to facilitate.
interaction among participants when appropriate. A BARCO large screen (lOft.) projection system can
display screens of individual PCs. In addition, a video switcher facilitates the movement of screen
images from PC to PC or downloads the public screen (facilitator’s) display to each workstation. The
facility includes four breakout rooms, also equipped with PCs, for small group discussion. PLEXSYS
software consists of a large number of tools, including tools for brainstorming, issue analysis, voting,
stakeholder identification, assumption surfacing, and recording what happens during a meeting. The
facilitator’s station provides access to and control over the group support tools. The facilitator helps
the group get the most out of the GDSS process by both guiding the meeting and running the software.
The interfaces have been set up so the user can understand the screens that appear even if they
have not seen a particular screen previously. The system, written in Turbo PASCAL, uses pop-up
menus, cursor selection from menus, and keyboard instructions to communicate with the user.
Based on insights gained from the operation of PlexCenter, it was decided to build a second facility.
The success of using PLEXSYS in small and medium groups suggested that larger groups might
benefit even more, so the second facility was designed to accommodate large groups. Over the first
eighteen months of operation, it was observed that satisfaction of the group with the use of the system
increased with group size. This led to the desire to build a larger facility to test the hypothesis that
"satisfaction using the system increases with group size." In addition, building a new facility provided
an opportunity to improve the facility design, to develop a new systems architecture, and to take advan-
tage of recent technological developments.
As far as we know, based on the literature and usage of software tools, the first set of CASE (computer-aided software
engineering) tools, namely PSL/PSA, came from Case Institute of Technology in 1965-68. At that time only
Teichroew’s group of Case, which later moved to the University of Michigan, was concerned with computer-aided
support of the systems development process. It is ironic that 15 years after the development of PSL/PSA, this type of
tool came to be known as CASE tools. Here, we recognize Teichroew as the originator of CASE tools.
622 MIS Quarterly/December 1988

Electronic Meetings
Break-Out
I
¯ Room I
Break-Out|
Room /
Break-Out
Room
White Board Wall Mounted Projectio’n Screen White Board
Facilitator Consol~
and Network
IFile Serve~" ~_
Barco ~-~
Projector~___J
Workstations
Storage \
\
Control
Room
Break Area
Figure lb. PlexCenter
2nd Facility
The second Arizona facility designed to support group work with information technology was opened
in November, 1987. The room was designed to accommodate 24 workstations with space for two
people per workstation (Figure 2b). In addition, gallery seating for 18 observers was included in the
back of the room. The room has a distinct legislative feel to it but it also facilitates talk across the
room, if appropriate. The 24 workstations house IBM PS/2 model 50s with high resolution color moni-
tors. The room is equipped with 38 audio pick-up microphones and six video cameras with stereo
audio capability. In addition to the two large screen displays, a high resolution video projector with
a remote control unit displays computer (analog and TTL) and NTSC video signals. This system per-
mits display of laser disc, transparencies, videotapes, 35mm slides and Videoshow 160, a computer
graphics presentation system with special effects.
A separate control room was built to house TV monitors, audio mixers, and video editing equipment
for monitoring and processing session recordings. The capability exists to capture the computer inputs
from all participants as well as audio and video recording of a session. Years later, a replay of a key
corporate discussion or decision could be reproduced. This capability would provide tremendous in-
sight for changes to corporate strategy, planning, etc., in the future.
Future Plans
result of our experiences (Nunamaker, et al., 1988a) has led us to consider the next phase
development of information technology to support meetings, which is to distribute some of the
MIS Quarterly~December 1988 623

Electronic Meetings
Projection Room
Conference
Room
Figure 2b. Decision Room for Larger Groups
functions of the collaborative meeting room to a participant’s office. It is not necessary to bring every-
one together in the electronic meeting room for each task. We are also planning to support groups
distributed around the country and the world.
In the near future, we will integrate PLEXSYS software tools with a videoconferencing system in order
to test the concept of distributed meeting room facilities. The first step is to connect our two GDSS
facilities with electronic and video links.
We envision a facility in which the scenarios are the same as those in our first and second rooms,
but participants are located thousands of miles away. Imagine a facility in which your group is sitting
in the center of a circular room. The walls of the room are covered with screens of the participants
from around the world. Each local group would find themselves in the center of all participants.
624 MIS Quarterly/December 1988