THROUGH GDSS DESIGN AND IMPLEMENTATION*
GDSS are "socio-technical packages" comprised of hardware, software, groupware, and people (Kraemer & King, 1988). The hardware such as workstations, computer network, and private and public display screens adds new communication channels to the traditional meeting space. Software such as database management systems, spreadsheets, and word processors add support for general information processing. Groupware is software that supports the collective work of the group such as judgment and policy-making. People like facilitators and technicians are needed to help make sure the GDSS runs effectively and to help integrate the GDSS into a groupís work. Each aspect contributes to the makeup of GDSS as a communication medium. GDSS groupware design, the facilitators, and a common implementation strategy will be analyzed here in order to understand better the solution GDSS intervention poses.
Genres of Organizational Communication
Yates and Orlikowski (1992, p. 301) define a genre of organizational communication as "a typified communicative action invoked in response to a recurrent situation." Organizational media are invented and redesigned better to suit organizational exigencies and even to change genres of organizational communication. While intimately related, there is a difference between a genre of organizational communication (e.g., memos and interviews) and the media (e.g., electronic mail, paper, and telephones) people use to communicate. Memos, for instance, have been developed to manage the demands of sharing and recording information among organizational members. The memo circa 1900, of course, is quite different than the memo of 2000. Yates and Orlikowski point out that genres of organizational communication evolve over time with changes in the exigencies organizations encounter and the availability of new media for communicating. While Yates and Orlikowski stress evolutionary adaptation of medium and organizational context, a genre also articulates organizational routines and expectations about who communicates, how they communicate, and for what purpose they communicate (Fairclough, 1995).
Meetings are a genre of organizational communication central to the structure and culture of organizations (Schwartzman, 1989). Many meetings involve deliberation, which is the activity aimed at resolving differences of opinion about the most prudent course of action to take in uncertain circumstances (Walton, 1992). It happens, for instance, when plans and policies are developed. A great deal of attention has been given to the improvement of meetings, and deliberation in particular (Frey, 1995; Schwartzman, 1989). The history of GDSS, for example, is a story about the pursuit of a new medium that would help organizations realize better deliberation when meeting (Aakhus, 1997). From the genre perspective, the GDSS medium is a complex mixture of tools, practices, and normative expectations intended to meet the exigencies collective choices and conflict represent for organizations and their members. GDSS intervention presents a view about how organizational members should give voice to their opinions and how organizational members should relate to the problems they experience.
GDSS is a medium designed for process intervention. That is, GDSS is introduced to aid the resolution of organizational choices and conflicts. The promise of GDSS intervention lies in its tools to aid interaction not in providing substantive expertise or treatment for individual incompetence. GDSS groupware applications, for instance, provide decision-makers tools to manage scarce discussion resources like turns, attention, and memory (Nunamaker et al., 1991; , Poole & DeSanctis 1992). There are tools to help groups overcome the lack of critical comments, aggressive interruptions, or inadequate problem analysis (DeSanctis & Gallupe, 1987; Kraemer & King, 1988; Nunamaker, Dennis, Valacich, Vogel, & George, 1991; Poole and DeSanctis, 1992; Poole, DeSanctis, Kirsch, and Jackson, 1995). GDSS facilitators help plan meetings, run the technology, and shape the discussion that takes place outside the GDSS (Bostrom, Anson, & Clawson, 1993). GDSS intervention is intended to help organizational members communicate and reason together more effectively. However, is the aid simply a neutral channel through which communication flows, or does GDSS reflect an official view and nascent theory about how choices and conflicts ought to be handled?
Prior research and commentary on GDSS suggests that it is not neutral in its design but that GDSS design prefers a particular type of deliberative communication for handling choices and conflicts. A democratic disposition in GDSS groupware has been noted by numerous scholars (Nunamaker et al., 1991; Poole & DeSanctis, 1992; Poole et al., 1995; Sharrock & Button, 1997). Others have criticized the assumptions about groups and organizations in GDSS groupware design (Allen, 1993; Grudin, 1994; Kraemer & King, 1988; Lyytinen, Maarenen, & Knuuttila, 1994; de Vreede & Bruijn, in press). The introduction of GDSS groupware does not produce uniform effects on organizational communication or decision-making (Seibold, Heller, & Contractor, 1994). Recent research also shows that GDSS are more useful when information gathering is important and less useful when synthesis is required (Bikson, 1996; de Vreede & Bruijn, in press). These findings suggest that GDSS designs presume particular forms of deliberation that may not match organizational members' expected or habitual forms for handling choice and conflict. There is increasing interest in the theories of mediation implicit in GDSS (Ngwenyama & Lyytinen, 1997; Poole & DeSanctis, 1992), yet there is much to learn about the solution GDSS intervention presents. Research on argumentation provides a means to understand what, if any, theory of intervention is available in GDSS for handling choices and conflicts.
The conduct of deliberation varies as does the rationality of deliberative processes. Contemporary theories of argumentation show how rationality is located in the practices and procedures for deliberating and thus varies as people institutionalize means for handling choice and conflicts (van Eemeren, Grootendorst, Jackson, & Jacobs, 1993; Toulmin, 1972; Walton, 1992; Willard, 1996). Walton (1992), for instance, shows that choices and conflicts can be handled through different types of "dialogue games" such as planning, negotiation, or inquiry. Each dialogue involves a different goal and set of rules from which flow different expectations about participant roles, message exchange, and standards for good arguments. Jacobs, Aakhus, Aldrich, and Schultz (1993; Aakhus & Jacobs, 1999) in a similar vein have shown how dispute mediators enact different rational models to create different forms of dispute mediation such as bargaining, therapy, or critical discussion.
Kochman (1981), in studying black and white interaction in classrooms, has shown how people construct different versions of the deliberation genre and the different rationality of participation associated with those genres. Deliberation is constructed as a "discussion" where participants engage in a contest of ideas and use reason rather than emotion to debate each other. It is also constructed as "argument" which is a contest of ideas and individuals where participants use affect and dynamic opposition to debate each other. The different versions of deliberation depend on different beliefs about the use of communication to manage differences of opinion and are manifested in the role turns, argument, arguers, authority, opposition, and silence play in deliberative interaction. The differences are summarized in Table 1.
Table 1: Genres of Deliberation
How the deliberation is constructed, according to Kochman (1981) and contemporary theories of argumentation, has consequences for how parties participate in deliberation and what deliberation produces. The truth-values of ideas in the argument version of deliberation, for instance, are established through the advocacy and defense of the merits of the positions held by opponents. In the discussion version, however, participants avoid advocacy by taking an objective stance to maintain an open mind. The truth-values of ideas, instead, are discovered through discussion that gives clear expression to ideas and through certification processes external to the interaction (e.g., experts) for verifying the ideas. The argument version succeeds in producing the best ideas when the best content and style dominate. If some participants have good style and poor content, however, participants with good content and poor style will bow out of the deliberation. This preference for style over content hurts the development of good ideas. The discussion version succeeds in allowing the greatest possible participation since one's participation is a right independent of their style. In discussion, dominant participants can be overruled; however, there is no good way to make participants express their own opinions. Thus, participants can invoke a "right" to remain silent which disables the capacity to test and validate ideas through discussion.
In the remaining sections, GDSS intervention will be examined as a form of organizational media designed to promote an appropriate form of deliberation for handling choices and conflicts.
Constructing Deliberative Communication
GDSS Groupware Design
Groupware is defined by Peter and Trudy Johnson-Lenz as the "intentional group processes and procedures to achieve a specific purpose plus software tools designed to support and facilitate the groupís work" (Hiltz & Turoff, 1992, p. 69). GDSS groupware is a small percentage of a nearly 6 billion dollar groupware industry with approximately 250 million users (Price Waterhouse, 1997). Nearly all GDSS groupware has emerged from academic development groups (Jessup & Valacich, 1993; de Vreede & Bruijn, in press). A review of GDSS groupware that, at one time or another, has been commercially available and used by professional facilitators and organizations reveals a common design that points to a more general theory of GDSS intervention.
The GDSS groupware products reported here are a sample of a broader population of software products that continue to evolve. This breadth and variety presented a problem for developing a comprehensive pool of products to analyze. Therefore, products were selected because the products grew out of different development environments that presumably reflect rival cases of design. The generalizations about GDSS intervention are built inductively from this naturally occurring set of contrasts. Descriptions of these different GDSS groupware packages rely on the authorís use of the tools supplemented by the academic and industry literature available about these products. Two products represent the mainstream of GDSS groupware development, MeetingWorks (1998) and GroupSystems (1998). The author has observed and used both of these commercial products. GroupSystems has a long, well documented development history, it has been the object of considerable GDSS research, and it has met with considerable commercial success relative to other GDSS groupware (Jessup & Valacich, 1993; Trauth & Jessup, 1999; de Vreede & Bruijin, in press).
Two other products, Council (1998) and k.net (1998), were also examined through demonstration versions of the product. These products were selected because the products lie outside the mainstream and thus present rival cases of GDSS groupware design. Council is groupware built for Apple McIntosh systems and k.net is a pc based product. k.net was developed to improve on the GroupSystems approach by offering improved applications enabled to operate over the internet. A final product, Software-aided Meeting Management (SAMM) is included here, even though the author has never used it, because of its prominence in GDSS research and its successful commercial use (Jessup & Valacich, 1993; Poole et al., 1995) . SAMM differs from the other products because it was designed to give all users full and equal access to the GDSS groupware. The other GDSS groupware products leave special software controls to a group leader or facilitator. Descriptions of SAMM are taken from (Poole & DeSanctis, 1992; Poole et al., 1995). So, SAMM presents another rival case of GDSS design .
GDSS Groupware products typically include a suite of applications that groups can use to collect, organize, and evaluate participant contributions. Table 2 identifies three common functions designed into GDSS groupware and lists the applications found in GDSS groupware that correspond to these three common functions . Collection functions allow participants to submit messages and to read and comment on the messages submitted by others. Organization functions allow participants to move messages previously submitted into various categories or outlines. Evaluation functions allow participants to rate or rank messages or categories of messages previously submitted. These systems also include an agenda application for organizing the flow of activity among applications as well as other supporting applications intended to improve work-group productivity.
Table 2. GDSS Applications by Purpose
GDSS product design treats deliberation much like the way the students described by Kochman (1981) treated deliberation as discussion not debate. This treatment is most evident in the way the tools enable participants to be separated from their positions and expressions. The applications guarantee every participant a turn independent of what they have to say. The possibility of anonymity provides cover for those who may otherwise remain silent about their point of view. So, through anonymity and unlimited turns, participants gain the benefits of remaining silent while still contributing. The separation of participants from their expressions promotes a contest of ideas where everyone is given maximal opportunity for participation.
While GDSS intervention does not prevent participants from engaging in deliberation as argument (a contest of ideas and individuals), that form of deliberation would have to take place in spite of the tools which promote deliberation as discussion. This shaping of the form of deliberation is evident in the way GDSS groupware design first separates participants from their expressions and then provides applications to recombined participant expressions into collective statements. Resolution is fostered by the meta-information organization and evaluation tools provide. In other words, categories and votes build and make available standpoints independent of any individual, and it is from these standpoints that the group can judge contributions. Collective opinions provide a functional substitute to experts for external verification and testing of ideas. GDSS groupware is not designed to obtain commitment through dynamic opposition and struggle. Struggle and opposition are removed from individuals into abstract categories and vote summaries that provide a means to discover the collective will of the group without imposing on individuals or forcing individuals to publicly commit to their positions.
The design pattern in GDSS groupware applications points to an official view about how choice and conflict are handled. The tools designed into the groupware as represented in the applications summarized in Table 1 are means for orchestrating choices and conflicts into sequential phases that funnel the aspects of choice or conflict toward a settlement (Aakhus, 1999). The fundamental problem GDSS groupware design attempts to solve is misunderstanding caused by the inability of participants to express their opinions and beliefs. This approach emphasizes the discovery of ideas by removing barriers that the social aspects of choices and conflicts represent.
Divergence-Convergence.The sequential or funneling orientation for orchestrating interaction is also evident in a common strategy for implementing GDSS described in the literature on GDSS. The strategy, commonly known as "divergence-convergence," consists of three basic moves (Bikson, 1996; Niederman & DeSanctis, 1995; Trauth & Jessup, 1999). First, the groupware is used to help the participants generate a massive pool of messages relevant to the issue at hand. It is not unusual for groups to produce hundreds of messages in just a few minutes. Second, the participants independently and collectively scan the mass of messages to identify redundant or duplicate answers to the problem. The purpose is to consolidate the mass of messages into a short list of candidate claims. Third, the participants determine and finally confirm an answer to the choice they face by evaluating the short list. Nunamaker et al. (1991, p. 43), in their classic article on GDSS, describe this common sequence of use that they suggest occurs independent of task and groups:
Meetings typically begin with participants generating ideas (e.g., "How can we double our sales over the next four years?"). As they type their comments, the results are integrated and displayed on the large screens at the front of the room as well as being available on each workstation. Everyone can see the comments of others, but without knowing who contributed what. Participants can build on each otherís ideas independent of any positive or negative bias about who contributed themóideas are evaluated on their own merits, rather than on the basis of who contributed them. These ideas are then organized into a list of key issues (e.g., "Stronger ties with customers"), which the group can prioritize into a short list. Next, the group could generate ideas for action plans to accomplish the important issues, followed by more idea organization and prioritization, and so on. The result of the meeting is typically a large volume of input and ideas, and a group consensus for further action.The divergence-convergence strategy is aimed at generating as much input as possible and then engaging the group in organizing and evaluating that input until a consensus emerges around a smaller set of solutions to the problem at hand (Trauth & Jessup, 1999). This strategy is one well-understood means for implementing GDSS, as one experienced GDSS facilitator told the author during an interview:
In all the processes I view, no matter what, what you produce with the group, it has to go through the three stages. You, you've got to get um, get any deliverable, any list or text. Youíve got to get people to talk, try and brainstorm or, you know, generate a list of potential items, potential deliverables. They need to generate, have some discussion. And then get them to agree on at least what that course set ought to be. And then either prioritize them or at least check with the group and say, "Is this a valid set or not?"The "divergence-convergence" strategy parallels the funneling found in GDSS groupware design. This is evident in the way it emphasizes maximal participation in resolving the choice or conflict while channeling participation to produce a final collective conclusion. This strategy reflects a broad understanding about the appropriate implementation of GDSS groupware. Each tool, however, must be made accessible and usable in order for the GDSS groupware to work. GDSS facilitators develop practices to render the GDSS usable.
Facilitation Practices.Two accounts about the effective use of GDSS groupware, selected from a corpus of interviews with GDSS facilitators, will be used to illustrate how the effective use of GDSS groupware design and divergence-convergence is understood . These accounts were selected because the facilitators describe how to implement groupware applications to handle common exigencies facilitators encounter. The accounts are not presented here to prove the generality or frequency of these practices but to index a more general orientation about the appropriate handling of choices and conflicts through GDSS intervention. The accounts were selected also because they portray a use of GDSS applications that is faithful to both groupware design and the divergence-convergence strategy. The accounts are far from deviant cases and thus help to articulate further the GDSS intervention rationale. Both facilitators work in corporate settings facilitating strategic planning and information system design meetings. Both facilitators at the time of the interview were relatively new to the facilitation role, but neither was a novice. They both use GroupSystems and have received similar training. The first account is about using the categorizing tool and the second about using the voting tool found in GroupSystems. Each account is previewed with a description of the groupware application referred to in the facilitatorís account. Both accounts help show the details involved in accomplishing divergence-convergence.
"Paring down" is a facilitation practice that helps move a group from divergent activity to convergent activity and thus foster decision-making progress. When paring down, the facilitator uses the categorizing tool to recombine and eliminate contributions in order to build a common representation of the choice or conflict. The facilitator engages the participants in a joint activity of list management, that is, how to build the shortest possible list that incorporates all the opinions expressed by the group members. The list forms the basis for further deliberation. The GDSS application will be described first followed by the facilitator's account of use.
The categorizing tool in GroupSystemsô is illustrated in Figure 1. This tool is designed to help decision-makers bring order to the volume of contributions that they make in collection activities like brainstorming. It is not unusual for participants in a GDSS session to produce hundreds of messages in just a few minutes. The GDSS groupware allows participants or the facilitator to copy contributions from the prior activity and paste these into the categorizing tool, or to make new entries. The categorizing tool allows participants to make headings that are used to categorize contributions. The tool also allows participants to make categories of categories, or "buckets," to further organize the contributions from prior activities. Categorizing is essentially an activity of making lists of items and then reducing that list of items. One way that list building is accomplished is through the practice of paring down described in excerpt 1 below.
Figure 1: "Categorizer" Application
The facilitator's account in excerpt 1 explains how the categorizing tool is used in paring down. The facilitator is a lead facilitator for a major North American financial institution. She specializes in strategic planning and information system design meetings. Her account was made while she discussed her plans for a forthcoming meeting that involves about 20 participants who will develop design requirements for a new information system. The meeting goal is to determine the most important functions to be built into the proposed information system for the firm. The meeting will take place in a conference room outfitted with 15 laptop computers networked and loaded with GroupSystemsô. Typical of GDSS intervention, the facilitator expects participants to use both oral and computer-mediated discussion in the course of the meeting.
The facilitatorís account begins with her reflection on a classic situation faced by facilitators. The GDSS groupware helps collect contributions but does not sort or make sense of the contributions for the participants. So, after collecting contributions, groups need help to make sense of their contributions. There are typically too many contributions for the group to work through. It is a situation the facilitator has faced before and one she anticipates for the upcoming meeting. In the excerpt 1, the facilitator describes a past situation where she used the categorizing tool to capture ideas from the participants and then, through paring down, uses the categorizing tool to shape the contributions into a usable basis for further deliberation. Implicit in the facilitator's explanation of paring down is a tacit theory of how deliberation should be constructed.
Excerpt (1)Paring down is a practice that renders the categorizer application useful to the participants. Paring down is presented by this facilitator as a technique for managing message volume that helps the group move towards a decision. The messages posted are scanned and clarified to "eliminate duplicates." Similar ideas are incorporated into more general statements so that the pool of messages is reorganized into a manageable list of ideas for further discussion. As such, the practice has the effect of expanding participant opportunities to engage each other and build shared understanding of their decision through the close analysis of their contributed messages. Participants gain greater control of the micro-features of their choice by having greater access to what others have said. While paring down renders the application useful, it also constructs the deliberation among the participants more as discussion and less as argument.
Paring down treats deliberation as list management. Paring down downplays deliberation as the clash of individual opinions and interests. Instead, it encourages participants to engage each other in building lists of categories that faithfully represent the non-redundant ideas in the pool of contributions. Resolution of opinions is achieved indirectly through a search for a set of categories that accounts for all the contributions in the pool. The facilitator helps the group retain all non-duplicate contributions thus decreasing both the list size and the volume of messages. Thus, the list stands as a common ground from which the participants can move forward in finding an agreeable outcome to their choice or conflict. Paring down actively shapes the nature of the disagreement by generating the fundamental categories for the decision.
Categorizing the pool of contributions is important to constructing deliberation as a funneling interaction. Paring down creates a set of agreements about how to represent all that has been said prior to categorizing. Thus, it creates grounds for further interaction. The practice does not aim to eliminate contributions but to find ways to include all contributions. In the facilitators example of paring down, her saying that something is a duplicate implies a basic agreement between two contributions without exploring the different positions or points of view in the expression. Also, keeping a contribution in the groupware system allows the idea to stand, thus promoting the equivalence of opinions with little regard to what is being claimed. Paring down does not prevent comparison, evaluation, and testing of individual opinions but it makes it easier to avoid these aspects of resolving choices and conflicts. The facilitator displays this preference by offering discussion about deleting or rewording an item as secondary options on which to proceed.
The clash of opinions is treated more as an obstacle to resolving the larger choice or conflict that the group faces rather than a means to resolution in paring down. This approach is evident in the hypothetical question the facilitator poses in her account about where a contribution fits. The answer posed by the facilitator involves asserting her neutral stance and offering procedural choices to the group. It is noteworthy how the facilitatorís description of paring down downplays the possibility for substantive disagreement and yet suggests the possibility that the facilitator could be instrumental in fostering the clash of opinions. While paring down could lead to more pointed argument and more persuasive arguments, it is conducted to create categories into which expressions are dumped. The potential for a clash over expressed opinions is overshadowed by the move to find abstract categories that will "hold" everyoneís opinion. Thus, the ground for deliberation is shifted to argument over appropriate categories and whether a contribution belongs in that category and away from the substantive implications of what a participant expressed in a contribution.
The "drive to consensus" is a facilitation practice that treats deliberation as the reduction of individual points of view into a collective point of view. In the drive to consensus, the evaluation applications are used to encourage decision progress by building points of agreement through successive voting. The GDSS application will be described first followed by the facilitatorís account of use.
Voting applications allow participants to perform various types of numerical ratings and rankings of items on a list. These applications also return summary statistics of votes such as the number of votes, the range, mean, and standard deviation for each idea evaluated. Examples of GroupSystemsô voting applications are illustrated in Figures 2 and 3. These tools are designed to help decision-makers bring closure to the choices they face. Figure 2 shows how votes are collected using a scale of 1 to 10.
Figure 2: Voting Application
Figure 3 shows how the voting tool provides summary statistics about the vote.
Figure 3: Vote Display Application
The facilitator in excerpt 2 below provides an account about using voting tools that display the practice of driving to consensus. The facilitator is a primary facilitator for a firm specializing in GDSS facilitation. He specializes in strategic planning meetings for businesses. His account addresses another classic problem faced in decision-making. That is, how to promote discussion and move a group toward a decision when choices become progressively more difficult.
While voting is often understood as a means to render a final choice, the voting tools in GDSS are often used as means to make sense out of the unexpressed premises and sentiments of the group and to validate the direction a deliberation is taking. Voting tools provide a quick statistical summary that can enhance the ability of participants to interpret where the group stands on the issues they face in their decision. It is this use of voting tools in GDSS intervention that is addressed in the following account.
In excerpt 2, the facilitator explains a standard situation where a group attempts to determine the root causes of a problem. He describes how a list of alternatives is reduced through successive votes conducted to reveal where true consensus for an alternative or set of alternatives exists. In this account he shows how vote statistics, such as standard deviation and vote clustering, are used to shape the direction and grounds of the deliberation.
Excerpt (2)The drive to consensus handles differences of opinion through the cycling of votes. As in this facilitator's description, the voting tool is used to find the points of scattered, incoherent opinion that provide the grounds for further discussion aimed at producing a collective standpoint. Deliberation is conducted to shift opinion away from murky, fuzzy dispersed opinions into ever-smaller sets of alternatives around which opinions are tightly clustered. Polling stimulates decision progress by "uncovering" the points of disagreement and thus aids the progression from problem formulation to final solution.
Clash and struggle are transformed into abstractions captured in lists and vote tallies separate from individuals. Participants should not be forced into public commitments but should be allowed privacy and the choice as to whether to explain their reasons for a holding a position. Individuals are not publicly accountable for their positions. The purpose of deliberation is to close points of disagreement rather than open them up. In addition to finding points of agreement, deliberation is also structured to motivate commitment to the emerging outcome. That is, the drive to consensus also identifies the alternatives that are the least problematic for the most participants.
The drive to consensus idealizes the rational standard in deliberation as popular opinion. The emergence of alternatives, for instance, is driven by a search for popular items as indicated by a low standard deviation. When asked how he guides discussion about disagreement, the facilitator asserted his neutrality and explained that the "beauty of the tool" is the ability to cycle through votes efficiently. Moreover, the facilitator encourages deliberation to take place over points where dispersion appears to exist not over points where agreement appears to exist. The facilitatorís account shows an orientation directed towards finding agreement for the sake of agreement and treating consensus as a problem of lining up votes for alternatives.
Deliberation as "Buy-In"
This analysis began by considering how GDSS is a medium designed to help organizations and their members realize meetings where choices and conflicts are more effectively handled. As such, GDSS intervention is a mixture of tools, practices, and normative expectations about handling the exigencies of organizational choice and conflict. The question addressed is whether GDSS intervention is a neutral medium or represents an official view and nascent theory about handling choice and conflict. The analysis suggests that GDSS intervention does promote a particular view that will here be called "buy-in." This term refers to deliberation conducted (1) to ensure maximum opportunity to express opinions and (2) to formulate popular opinion as a means to arbitrate differences and foster progress toward resolution. This view of how to construct deliberation is evident in GDSS groupware, the divergence-convergence strategy, and facilitation practices.
Buy-in idealizes a form of deliberation that, in Kochman's (1981) terms, more closely resembles discussion rather than argument. The format and rationality of buy-in is captured in a British Broadcasting Corporation documentary investigating the effective use of decision-support systems in organizations. The documentary featured GroupSystems as one new-wave solution to the age-old problems of effective decision-making in organizations. After the use of GroupSystems was demonstrated in a financial services firm, a cameo of the firms's CEO appeared where he explained why GDSS intervention is effective (British Broadcasting Corporation, 1991):
I think managers who fear certain kinds of issues arising in an environment in which they cannot exercise control will ah perhaps feel discomforted by this kind of thing. My belief, my philosophy is such that Iíd much rather have these issues surface so that I and the organization together can deal with them, rather than have the issues repressed where they continue to do their damage. . . . There is a greater awareness on the part of the affected people as to why the decision seems to be tending the way it is uh and how it actually concluded. All of which means that when we implement that decision there is a much greater degree of buy-in because there has been this greater degree of participation. So the the system allows for this honest, open broad input and at the same time is consensus building.This colloquial use of buy-in captures the motive or purpose of buy-in, which is achieving broad input and consensus. Buy-in idealizes the deliberation genre as a sequence of activity that funnels choice and conflict toward a conclusion and as a model of rationality that both explains and justifies the approach. The format and rationality of buy-in will be addressed here to outline this nascent theory of intervention.
The GDSS groupware design, intervention strategy, and practices analyzed here show how the deliberation genre is idealized as a form of interaction that funnels differences of opinion and points of view toward a consensual conclusion. GDSS intervention resembles sequential models of decision-making interaction, such as Simons' (1960) "choice" model or Deweyís (1910/1991) "reflective sequence." Sequential models treat the resolution of choices and conflicts as a "cognitive process that can be decomposed into a sequence of simple, programmable steps," which can be organized to converge on a decision by moving from a problem formulation phase to a final choice phase (Langley, Mintzberg, Pitcher, Posada, & Saint-Macary, 1995). Sequential models are discussed and prescribed in the communication and decision-making literature because they are believed to maximize information search, the testing of assumptions, and the production of the most rational answer to the problem addressed (Gouran, Hirokawa, Julian, & Leatham, 1993; Langley et al., 1995). Indeed, GDSS have been criticized as too rational in this regard (Allen, 1993; Grudin, 1994). The procedures and practices of GDSS intervention analyzed here, however, suggest that the deliberation format fostered in GDSS intervention promotes a particular type of rational model for resolving choices and conflicts.
There are assumptions about intervention legitimacy and efficacy reflected in the GDSS intervention practices and procedures analyzed earlier. Jacobs, Aakhus, Aldrich, and Schultz (1993; Aakhus & Jacobs, 1999) develop three models of rationality of dispute resolution by identifying what divorce mediators treat as the source of conflict, the optimal solution, the principles of resolution, the process of resolution, and the mode of resolution. The rational model of buy-in is reconstructed here using these five categories. The source of conflict from the buy-in perspective is misunderstanding more than conflict of rights or interests. The optimal solution to a conflict is a proposal that does not impose on the autonomous positions of the participants. The principle of resolution is whether a proposal can be justified against popular opinion. The process of resolution is to collect, organize, and evaluate individual comments in cycles until a proposal emerges. Finally, the mode of resolution is a plan of action that projects the collective sentiment of the group. The combination of facilitator actions and GDSS design establish the popular opinion of the participants as the rational standard for judgment in decision-making. This emphasis on acceptance of majority domination, combined with the impression that all participants have their say in deliberation, is the essence of the buy-in rationality.
Buy-in shows intelligence in its capacity to build common ground where none exists and to construct a means by which collectives can judge ideas. It demonstrates this intelligence in a way that allows individuals to be separated from their expressions. This separation not only provides a way for individuals to protect themselves but it also creates freedom for participants to float expressions that can then be tested by the group. It may be that the evolution of buy-in as a form of intervention through GDSS is a selective adaptation for successfully creating agreements under contemporary conditions of choice and conflict in organizations. On the other hand, buy-in promotes a form of ad populum argument where resolution is achieved by the test of public opinion rather than drawing out the substantive implication of claims. This implicit understanding promotes a kind of argumentative relativity enhanced by the capacity to construct group beliefs about decisions through procedures that aggregate participant opinions. The form of consensus promoted is one that maintains the possibility that all the participants can have their epistemic way and does not challenge participants to reason across the social, political, and epistemic boundaries in an organizational choice or conflict.
The buy-in view of intervention reconstructed here is more illustrative than demonstrative and more analytic than conclusive. In many ways, the present analysis raises more questions than it answers. That buy-in is an operative GDSS intervention rationale, the logic of which is hard to deny based on the authorís extensive experience in the field. However, the pervasiveness of the logic in the culture of GDSS intervention practice is assumed not demonstrated here. The status of buy-in as an explanation of GDSS intervention is open to criticism, refinement, and rejection. It could be that buy-in is a mythology of GDSS intervention practice used retrospectively to make GDSS appear as a legitimate and effective form of intervention. Thus, the "real" theories of intervention practitioners put into play have yet to be identified. It could also be that buy-in is an evolutionary step in the development of GDSS intervention practice. Thus, once practitioners and users get beyond initial uses they may develop more nuances to their intervention practice. GDSS intervention seems to have its place and time but more work needs to be done in this area. Whether buy-in is the only construction of deliberation possible in GDSS is not explored here. It is worth exploring whether there are other models for constructing deliberative interaction. If there are some, then a more complete contingency approach will be necessary. If there are none, then there is a need for normative theorizing that opens up views of deliberation in the practice of GDSS intervention. These are issues that can be pursued in future research.
Whether or how buy-in may dominate GDSS intervention is not adequately addressed here but there are reasons to suspect that such a view may dominate how GDSS intervention is understood in the research and practitioner literature, if not how it is practiced. There are two basic questions pursued in prior research on GDSS. First, does GDSS enhance group interaction and reduce process losses involved in group work (Connelly, Jessup, & Valacich, 1990; Dennis, Valacich, & Nunamaker, 1990; DeSanctis & Gallupe, 1987; Nunamaker et al., 1991; Vogel, Nunamaker, Martz, Grohowski, & McGoff, 1989). Second, does the faithful adoption of the GDSS by users result in an improved decision-making process and outcomes such as higher decision consensus and satisfaction (Anson, Bostrom, & Wynne, 1995; , Poole & DeSanctis, 1992; Poole et al., 1994; Poole & Holmes, 1995; Poole et al., 1991). While these two lines of research build from two different orientations, the research is geared toward buy-in in its concerns about effective GDSS intervention. For instance, the effectiveness research attempts to show whether there is an increase the capacity for information exchange (Clawson, Bostrom, & Anson, 1993; Dickson, Lee-Partridge, Limayem, & DeSanctis, 1996), a reduction in social and authoritative pressure (Anson, Bostrom, & Wynne, 1995), an improvement in decision consensus and satisfaction (Dennis & Gallupe, 1993; Ngwenyama, Bryson, & Moboluren, 1996; Poole & Holmes, 1995) and an increase in good ideas (Brashers, Adkins, & Meyers, 1995; Brashers, Adkins, Meyers, & Mittlemen, 1995; Connelly, Jessup, & Valacich, 1990).
To put buy-in into perspective and go beyond it will require going beyond outcome measures that presume deliberation is a harmonious discussion for dealing with choices and conflict. The challenge is most notable in the way openness is conceptualized as privacy and resolution as abstraction and popular opinion in both research and design. This trend, however, is not surprising. GDSS were largely designed, after all, to support decision-making in hierarchical organizations. The procedures designed into GDSS intervention were also borrowed wholesale from ideas about choice and conflict that emerged during the salad days of the human relations movement (Aakhus, 1997).
The move beyond buy-in will be challenging because the model is likely to be embedded in researcher, designer, and user expectations about good deliberation, thus making it hard to see how or why intervention models should change. Developing alternatives requires that researchers, designers, and users entertain alternative expectations about how organizational choice and conflict could be handled. These expectations should include contemporary views of argumentation. Making such alternatives real, however, will be a struggle equivalent to organizational research overcoming the grip of the human relations school on its research agenda (Perrow, 1986). The present analysis is one step in the direction of opening up alternative views of deliberation and the management of choice and conflict through technology because it shows how a theory of deliberation is embedded in the technical procedures and implementation practices designed to improve communication.
Author's Note: Earlier versions of this manuscript appeared as:
Aakhus, M. (1997). Settlement on the electronic frontier: The use of group decision support systems in argumentation management. In J. Klump (Ed.), Proceedings of the Tenth NCA/AFA Conference on Argumentation: Argument in a time of Change - Definitions, Frameworks, and Critiques (pp. 132-137). Annandale, VA: Speech Communication Association.
. The author is hesitant to make any strong claims about SAMM relative to the other GDSS groupware because of his limited exposure to the product. Given descriptions of SAMM in the literature, it appears to represent the same theory of GDSS intervention as the other products.
. The information about the GDSS products used in Table 2 was gathered from discussion with designers, interviews with professional facilitators who use GDSS, attendance at two professional facilitation conferences, the author's use of the product, and researching the websites for the GDSS product. These were the most popular products at the time of the research between 1993 and 1996. The categories of functions are taken from Simon's (1960) model of decision-making.
. The data reported below are taken from a corpus of interviews with 40 facilitation professionals who either implement or design groupware products. The interviews ranged from 1 to 2 hours in length. The professionals interviewed primarily intervened on strategic planning and information system development meetings to foster discourse quality and decision-making progress in organizational decisions.
Aakhus, M. (1997). The communication logics of computer-supported facilitative interventions: A study of the community of practice and social technologies surrounding the use of group decision support systems in process facilitation. Unpublished doctoral dissertation, University of Arizona, Tucson.
Aakhus, M. (1999). Reconstruction games: Assessing the resources for managing collective argumentation in groupware technology. In F. H. van Eemeren, R. Grootendorst, J. A. Blair, and C. A. Willard (Eds.), Proceedings of the Fourth Interntional Conference on Argumentation (pp. 1-7). Amsterdam: International Centre for the Study of Argumentation.
Aakhus, M. & Jacobs, S. (1999). The
relationship between ideals and practice in divorce mediation. Manuscript
submitted for publication.
Bostrom, R., Anson, R., & Clawson, V. (1993). Group facilitation and group support systems. In L. Jessup and J. Valacich (Eds.), Group Support Systems: New Perspectives (pp. 146-168). New York: MacMillan Publishing.
Brashers, D., Adkins, M., & Meyers, R. (1995). Argumentation computer mediated group decision making. In L. Frey (ed.), Group communication in context: Studies of natural groups (pp. 263-282). Hillsdale, NJ: Lawrence Erlbaum.
Brashers, D., Adkins,
M., Meyers, R., & Mittleman, D. (1995). The facilitation of argumentation
in computer mediated group decision-making interactions. In F. van Eemeren,
R. Grootendorst, J. A. Blair, C. A. Willard (eds.), Proceedings of the
Third ISSA Conference on Argumentation (pp. 606-621). Amsterdam: International
Centre for the Study of Argumentation.
Dickson, G., Lee-Partridge, J., Limayem, M., & DeSanctis, G. (1996). Facilitating computer-supported meetings: A cumulative analysis in a multiple-criteria task environment. Group Decision and Negotiation, 5(1), 51-72.
Gouran, D., Hirokawa, R., Julian, K., & Leatham,
G. (1993). The evolution and current status of the functional perspective
on communication in decision-making problem-solving groups. In S. Deetz
(ed.), Communication Yearbook 16 (pp. 573-600).
Jacobs, S., Aakhus, M., Aldrich, A., & Schultz, N. (1993, November). The functions of argumentation in models of conflict resolution. Presented at the Speech Communication Association of America annual convention, Miami, FL.
k.net [Computer Software]. (1998). Austin, TX: Milagro, Inc. Available: http://milagro.austin.tx.us/k.net/building.html
Lyytinen, K., Maaranen, P., & Knuuttila, J. (1994). Groups are not always the same: An analysis of group behaviors in electronic meeting systems. Computer Supported Cooperative Work (CSCW), 2, 261-284.
Ngwenyama, O., & Lyytinen, K. (1997). Groupware environments as action constitutive resources: A social action framework for analyzing groupware technologies. Computer Supported Cooperative Work: The Journal of Collaborative Computing, 6, 71 - 93.
Poole, M., & DeSanctis, G. (1990). Understanding the use of group decision support systems: The theory of adaptive structuration. In J. Fulk, & C. Steinfeld (Eds.), Organizations and communication technology (pp. 173 - 193). Newbury Park: Sage.
Poole, M., DeSanctis, G., Kirsch, & Jackson, M. (1995). Group decision support systems as facilitators of quality team efforts. In L. Frey (Ed.), Innovations in Group Facilitation (pp. 299-322). Cresskill, NJ: Hampton Press.
Seibold, D., Heller, M., & Contractor, N. (1994). Group decision support systems (GDSS): Review, taxonomy, and research agenda. In B. Kovacic (Ed.), New approaches to organizational communication (pp. 143-168). Albany, NY: State University of New York Press.
Sharrock, W., & Button, G. (1997).
On the relevance of Habermas' theory of communicative action for CSCW.
Supported Cooperative Work (CSCW): The Journal of Collaborative Computing,
Trauth, E., & Jessup, L. (1999). Understanding computer-mediated discussions: Positivist and interpretive analyses of group support system use. Management Information Systems Quarterly [online] 83 pages. http://www.misq.org/archivist/forthcoming/accepted.html [October, 1999].
de Vreede, G., & Bruijn, H. (in press). Exploring
the boundaries of successful GSS application: Supporting inter-organizational
Copyright 2000 Communication Institute for Online Scholarship, Inc.
This file may not be publicly distributed or reproduced without written permission of the Communication Institute for Online Scholarship, P.O. Box 57, Rotterdam Jct., NY 12150 USA (phone: 518-887-2443).