Prior Studies of Computer Conferencing

The existing literature of computer conferencing can be usefully described within this four element grammar of mediators, characteristics, effects, and practices. The usefulness of breaking up the literature in this way is two fold. First, most current literature is dominated by an orientation to one or another of these elements. Indeed, the bulk of this literature, beginning with Hiltz and Turoff (1978), is largely oriented to either projecting or documenting the effects of computer conferencing on those that use the medium. Second, whole communities of research have oriented themselves, and with good reason, to one or another of these elements. The literature of the computer conferencing development community, for instance, is composed almost entirely of descriptions of mediators: computer networks, clusters of terminals, computer conferencing software functions, software human factors, and proposed standards for data interchange.

The first major text concerned with computer conferencing, Hiltz and Turoff's The Network Nation (1978) is perhaps the most generalized. Although the book is dominated by an interest in the effects and prospects for computer conferencing, whole chapters are devoted to descriptions of computer conferencing software functions and the user's interface to that software. Other chapters explore the characteristics of computer conferencing and the emergence of rules, roles, and norms among computer conference participants. This diversity of content separates the Hiltz and Turoff text from the rest of the current computer conferencing literature. It may well account, in part, for its continuing power.

Mediators

The literature of computer conferencing mediators is generally restricted to the computer conferencing software development community. This literature is generally concerned with the features of computer conferencing system software, the nature of user interfaces, and the manner in which systems have been adapted to computer networks and user requirements. This literature begins with Bush's (1945) As We May Think, which imagines a prototype for what would, twenty years later, become the computer terminal. Written well before the first general purpose computer was built, this work of imagination describes the possible workings of a practical text database accessed through an interactive computing system. It can be regarded primarily as an attempt describe a set of mediators that form a projected medium of communication.

Most subsequent mediator-oriented literature is much more strongly oriented to the reality of what can be done today. Flavin, Williford, and Barzilai (1986), when describing "Computer Conferencing Data Structures in the GRANDiose System", are fairly typical of this literature. They first explore the kinds of software functions and structures needed in order to properly support the ways in which people want to use computer conferencing. They then describe a computer conferencing software implementation that implements these functions and structures. Other elements of the same software are described in Flavin and Williford (1985) and Williford, Santero, and Flavin (1987).

Palme (1984b, 1984c, 1985a, 1985b, 1987) represents a second tradition of mediator-oriented computer conferencing literature. These and other papers by Palme are oriented to documenting the computer conferencing software he and his associates have built and setting standards for electronic message interchange. Palme (1988) documents a variety of key filtering mediators which are sometimes incorporated into computer conferencing systems, including:

• conferences, the built-in division of conference facility content into distinct topics,
• tree-structured sub-conferences that organize material within topics,
• access control filters that control access to conferencing facilities and topics. Such facilities can be used to create open conferences, closed conferences, and conferences which allocate different levels of access to different people,
• moderators who have the power to delete messages and reorganize conference entries,
• preview filters that allow appends to be reviewed in advance of distribution,
• conversations ... a series of messages chained together by "replyto" or other reference links,
• information-retrieval facilities that allow the reviewing of previous entries,
• selection filters that scan incoming messages and select messages for viewing based on selection criteria.

Note, in this, that the mediators that most concern the designers of computer conferencing software are rarely computers, terminals, or networks. The mediators that matter are the ones used in software to coordinate these physical elements with the objectives of the user. The implementation of mediators in software makes computer conferencing and other computer media particularly flexible. Modifications of these software mediators can be used to reshape the workings and appearance of a medium to meet user needs, to provide rapid filtering of information that, on other media, would typically be handled by people.

Chess and Cowlishaw (1987) overview the mediators underlying the TOOLS software and IBMPC computer conferencing facility that are the focus of this study. TOOLS differs from most other computer conferencing software in its orientation to distributed computer conferencing over wide area networks. Its implementation of "shadows", distributed duplicates of a master copy of a computer conferencing facility, allows IBMPC participants all over the world to view IBMPC in exactly the same way they might view it if they actually used the same computer system the main computer conference is located on. Its implementation of "peers" allows a particular conference (called a forum in IBM) to be shared between several computer conferencing facilities and maintained identically on each.

Characteristics

Although there is no distinct literature of computer conferencing characteristics, there are discussions of the characteristics of the medium embedded in other literature. Mediator oriented literature (Chess and Cowlishaw, 1987) frequently discusses characteristics in the course of discussing the effects of mediators. Effects oriented literature (Turoff and Hiltz, 1987) frequently discuss characteristics in the course of describing the medium. For Chess and Cowlishaw (1987), computer conferencing offers "fast asynchronous retained communication". Turoff and Hiltz (1987) summarize the characteristics of computer conferencing as follows:

• written form
• asynchronous
• signaling and organization make it possible to synchronize asynchronously-submitted contributions into a meaningful transcript.
• shared memory, made possible through the long term preservation of a conference transcript.
• structuring of the protocols of the communication process is made possible by software that organizes interaction.
• immediacy. Communication is conducted at the individual's convenience.
• human roles can be created as a function of the software.
• integration of the communication process with other activities, including data base access and other computer based activities.

Effects

If impacts are, as Hewes and Planalp (1987) suggest, the property most commonly linked to the definition of communication, it should come as no surprise that the earliest studies of computer conferencing as a medium of communication were strongly oriented to identifying the prospective impacts of the new medium. Hiltz and Turoff (1978) and Johansen, Vallee, and Spengler (1979), the texts that introduced computer conferencing and computer mediated communication to the field of communication as important prospective objects of study, are both heavily oriented toward identifying the prospective effects of the medium on individuals, organizations, and societies. This concern is easy to understand. A medium of communication that changes the environment of its users for the better is more likely to be successful than one that offers little that has not been offered by prior media.

The computer conferencing literatures orientation to effects begins with Hiltz and Turoff's (1978) The Network Nation. At the time this text was released, the world of computer conferencing was very small. Not many people had even heard of computer conferencing. Only a fraction of those that were familiar with the idea had ever participated in a computer conference. Serious users of any computer media could still be counted in the thousands, and only a handful of them had made serious use of computer conferencing. Most computer conferences remained small, involving no more than a few dozen people, and only a small number of developers and researchers used the conferencing continuously for more than a few months.

Hiltz and Turoff (1978)

Hiltz and Turoff concede all this in their preface, which dates the first computer conference to 1970 and limits the sum of computer conferencing experience to "tens of organizations and a few thousand people" (p. xxix). Yet from what is clearly a minimal set of data, the authors attempt to frame a grand vision of what computer conferencing is destined to be, starting with the following assertions (p.xxix-xxx):

• Computer Conferencing will be a prominent form of communications in most organizations by the mid-1980's.
• By the mid-1990's, it will be as widely used as the telephone today.
• It will offer a home recreational use that will make significant inroads into TV viewing patterns.
• It will have dramatic psychological and sociological impacts on various group communication objectives and processes.
• It will be cheaper than mails or long distance telephone voice communications.
• It will offer major opportunities to disadvantaged groups in the society to acquire the skills and social ties they need to become full members of the society.
• It will have dramatic impacts on the degree of centralization or decentralization in organizations.
• It will become a fundamental mechanism for individuals to form groups having common concerns, interests or purposes.
• It will facilitate working at home for a large percentage of the work force during at least half of their normal work week.
• It will have a dramatic impact upon the formation of political and special interest groups.
• It will open the doors to new and unique types of services.
• It will indirectly allow for sizable amounts of energy conversation through the substitution of communication for travel.
• It will dramatically alter the nature of social science research concerned with the study of human systems and human communications processes.
• It will facilitate a richness and variability of human groupings and relationships almost impossible to comprehend.

These projections emphasize the potentially potential positive effects of computer conferencing. All lead to desirable ends. Some, especially the projected impacts of computer conferencing on the handicapped and disadvantaged, would, if realized, be outstanding contributions to the future of everyone. The authors are not blind to the problems a new computer conferencing medium can bring, however. Indeed, they express concern that poor decisions in the regulation of such media may (p. xxx):

• Further widen the gap between small and larger institutions in terms of concentration of wealth and power.
• Further sharpen the distinction between the well-to-do, the educated, and the disadvantaged.
• Further the decay of the central city.
• Allow for comprehensive forms of censorship and regulation of information exchange, while destroying the press economically.
• Be regulated out of existence or emerge as a shadow of the potential inherent to the technology.

Eleven years later, it is difficult to entirely reject any but the first of the positive assertions. Late in the 1980's, Computer Conferencing is not a prominent form of communications in most organizations. It appears unlikely, moreover, to become as widely used as the telephone in the next decade. Computer Conferencing is a prominent form of communication in a growing number of organizations, however, and this study will show, among other things, that in at least one organization, computer conferencing and electronic mail may each be more widely used than the telephone. Hence it is still quite possible that computer conferencing will fulfill the promise of these projections, albeit some years later than Hiltz and Turoff estimated in 1978.

Turoff (1988) ratifies this sentiment:

it should be made clear that my enthusiasm has not dampened, nor do I feel pessimistic about some of my earlier predictions actually occurring. My time scale has considerably shifted over the years, but the basic driving forces of the potentials and benefits that these systems offer will ultimately overcome current impediments.

Hiltz and Turoff (1978) might be described as a projective ethnography, built from a mix of literature review, participant observation in small scale computer conferences, informed guessing, and fictional vignettes. Their literature review includes a range of topics that bear on computer conferencing, including discussions from sociology, social psychology, small group communication, and ethnomethodology on such diverse topics as "Face" (Goffman, 1967), interaction process analysis (Bales, 1950), status (Steiner, 1972), group size (Shaw, 1960; Wicker, 1969), risky shift (Teger and Pruitt, 1967) and learning. The informed guessing takes advantage of the expertise of a number of early participants in computer conferencing and uses a sophisticated projective survey technique called Delphi. The participant observation is drawn from experimental computer conferences conducted using computer conferencing software and facilities developed and administered by Turoff and his associates.

Other effect-oriented hypotheses put forward by Hiltz and Turoff (1978) include:

• The cue-emitting capabilities ... in computer conferencing are both more limited than and different from those in the well-rehearsed face-to-face situation. ... (p.87)
• It seems likely that organization of discussion and problem solving through computerized conferencing might enable a large, diverse group to tap the resources of all the members without losing the ability to communicate freely or experience the negative effects of large group size.(p.114)
• Computer conferencing may cut down the likelihood of imprudent or risky decisions being made.(p. 116)
• greater participant comfort and satisfaction with a low-immediacy and low-intimacy mode, such as computerized conferencing, for some kinds of communication tasks and for heterogeneous groups with hostile factions.(p. 117)
• A CC system might decrease the amount of time necessary for department heads to agree among themselves on common actions. It can also help to spread the "presence" and influence of a leader.(p.141)
• computerized conferencing might encourage the humanistic approach to be carried to the extreme, where group task accomplishment is replaced by group therapy. The heightened sense of real communication with others can become addictive in nature and seem an end in itself.(p. 143)
• We believe that computerized conferencing would allow complex team projects to exist where the members are geographically dispersed.(p. 143)
• Computerized conferencing can facilitate the decentralization of information exchange and decision making. The Home Office might become simply a supplier of services to relatively autonomous units of the organization. (p. 144)
• <Computer Conferencing>> can lend support to individual initiative and judgmental approaches by widening the range of intuition that can be brought to bear in a coherent manner on a given issue. (p. 162)
• Operational systems have demonstrated that managers in dispersed locations in an organization can communicate more fully and efficiently with CC systems than without them. Many potential applications appear promising in the managerial area, including work forces, crisis management, budget allocation processes, and the replacement of single-purpose word processing systems with multipurpose conferencing systems. (p. 162)
• Computer Conferencing appears to be ideally suited to increasing the speed and amount of communication and to supporting ... geographically dispersed scientific research communities. (p.214)

Johansen, Vallee, and Spengler (1979)

Electronic Meetings: Technical Alternatives and Social Choices (Johansen, Vallee, and Spengler, 1979) is entirely effects oriented. This text takes a broad perspective on media, examining teleconferencing and videoconferencing in as much detail as computer conferencing. As a result, its relevant content, restricted to a review of previous literature and listing of the strengths and weaknesses of videoconferences, audioconferences, face-to-face meetings and computer conferences, is much more limited that Hiltz and Turoff (1978). Here, reorganized and restated for clarity, is a sampling of some of the more important strengths and weaknesses of computer conferencing, as enumerated in the Johansen, Vallee and Spengler text:

• The value of a written record
  • Users have reported an ability to deal with larger amounts of information more efficiently, but information overload can occur if the amount of information gets out of hand.
  • Computer conferencing allows more time for reflection on the topic of conversation, but problems can take longer to solve in written mode.
  • Participants can obtain more deliberate answers to technical questions, backed up by written facts and with less delay.
  • In synchronous computer conferencing, messages are not sequential and multiple topic threads can appear; information overload can thus result. It can also be difficult to focus discussion.
  • Handwritten messages are more persuasive than video or face-to-face communications.
  • Typewritten communications are much less verbose than voice channels for solving the same simple problems.
  • Written negotiations take more time, are more rigid, and are more susceptible to developing intransigent positions.
• Computer Conferencing in the work-place
  • Access to computer conferencing can make working hours more flexible.
  • Computer conferencing can provide a continuous link among disseminated researchers.
  • Computer conferencing appears particularly useful in coordinating technical projects.
  • With computer conferencing, as many as 50 people can work together on a project.
  • Computer conferencing is well suited for communication involving the deaf, the handicapped, and homebound persons.
• Computer Conferencing and Interpersonal Skills
  • Computer conferencing can support self-presentation and emotional subtleties, but there is often a lack of interpersonal feedback; those who perceive the need for immediate feedback might thus be frustrated.
  • Computer conferencing is perceived as satisfactory for exchanging information, asking questions, exchanging opinions or orders, staying in touch, and generating ideas, but is perceived as unsatisfactory for bargaining, resolving disagreements, persuasion, and getting to know someone.
  • Regularity of individual participation is sometimes difficult to enforce in computer conferencing.
  • Computer conferencing can enhance candor of opinions.
• Computer Conferencing and Group Skills
  • Computer conferencing can be used to enhance crisis resolution.
  • Computer conferencing can be used to aggregate group judgement.
  • Greater equality in group participation can be facilitated by the use of computer conferencing, especially in synchronous situations.
  • Computer conferencing demands strong leadership.
• Characteristics of computer conferencing
  • Computer expertise is not a prerequisite to effective use of computer conferencing, but participants must learn new skills to use computer conferencing (e.g. how to send a message).
  • Computer conferencing is vulnerable to poor human/machine interface with both computer networks and computer terminals.
  • Participants can learn to use computer conferencing quickly, but training of new users is very important.
  • Typing ability is not a prerequisite to participation in computer conferencing.

Like Hiltz and Turoff (1978), this text is seminal. There is almost no literature to draw on. Hence, although other supporting studies are also cited, 34 of the 49 statements Johansen, Vallee, and Spengler make about computer conferencing come from research of the authors. Hiltz or Turoff are fully responsible for another four. Ten of the remaining eleven conclusions have no basis in research on computer conferencing, but come from studies of written communication. Less than a third of the results are replicated outside of the the research programs of Vallee or Turoff's associates.

The Science of Informed Guessing

Kerr and Hiltz (1982) provide what is probably the most complete distillation of the prospective impacts of computer media, including computer conferencing. Using the Delphi survey method, their book forms a number of of hypotheses concerning the probable impact of computer conferencing on individuals, organizations, and societies. Delphi is a multi-pass survey technique in which respondents are asked questions and then given an opportunity to modify their answers based on a review of the answers of all respondents. The technique, which is usually conducted over a period of weeks or months, treats each pass in the same way. Hence, for each question, the participant is given a set of possible answers and asked to modify or add to those answers. The answers distributed in the first pass, of course, are generated by the people constructing the survey, in this case on the basis of a review of relevant literature.

Seventy-nine impacts are hypothesized as a result of this multi-part survey. Some of these impacts are trivial (for instance, two of the individual cognitive impacts tell us that "It requires new skills" and "It discriminates in favor of the literate"), but some offer fairly interesting hypotheses. According to this study, computer conferencing may:

• expand "effective scope"; the number of alternatives, pertinent stimuli, awareness of social and cultural horizons. (p.95)
• allow users are able to deal with larger amounts of information more efficiently. (p. 95)
• increase the possibility of information overload. (p. 95)
• have the potential for addiction. (p. 104)
• increase the number and strength of support systems: kin, friends, the availability of professional help. (p. 104)
• support self-presentation and emotional subtleties. (p. 104)
• blur the distinctions between work and leisure. (p. 112)
• create the opportunity to be "in the center of the action" without regard to geography. (p. 112)
• create group resources as individuals join on the basis of verbal output rather than traditional credentials. (p. 122)
• create opportunities to develop communities of interest rather than those based on geography, discipline, etc. and redefine of the meaning of "local." (p. 122)
• increase lateral network linkages within and between organizations. (p. 132)
• make research communities more open. (p.132)
• increase the effective size of working groups. (p. 133)
• make the emergence of a leader different and less likely. (p. 133)
• kinship ties resolidify to counter residential mobility. (p.134)
• accelerate the shift from time-binding (traditional, religious) to space binding (political, pragmatic, instrumental) societies (p. 156)
• increase shared metaphors for people in many walks of life. (p. 157)
• decrease the importance of geographically defined communities as sources of identification, with increasing attention to shared interest. (p. 157)
• increase the potential for democratic capitalism. (p. 158)
• increase the potential for centralization of power. (p. 158)

Computer conferencing is, for this book, still more a potential than a reality. The authors and their experts are familiar with computer conferencing, but the community experts are few (18), and the electronic communities they are familiar with (via studies and/or experience) are both small and, in general, short lived. Hence while the authors feel that they may have a reasonable handle on the individual and group hypotheses, they regard the societal implications as "informed guesses" (p. 159). The development of computer conferencing, for the purposes of this book, remains immature. Hence the effects hypothesized are not seen as knowns, but as the basis for future research.

Subsequent Literature oriented to effects

All of these texts are critical to our current understanding of what computer conferencing is and may yet be, but none are entirely satisfying. Informed guesses are the single major constituent of the studies. Where there is real data, it is still based on very small computer conferences (usually involving "only relatively small numbers of users" (Kerr and Hiltz, 1982, p. 93); rarely more than a dozen people), usually of fairly brief duration (a few weeks or months from beginning to end), and usually with participants who either have little real investment in what occurs (most often student experimental subjects) or who have too much investment in what occurs (developers of computer conferencing software).

These characteristics are common in the literature of computer conferencing. Babitz et al (1983), Baird and Borer (1987), Baldwin et al (1978), Bamford (1980), Crickman and Kochen (1979), Dodd (1983), Elton and Carey (1983), Ferguson (1976), Guillaume (1980), Hahn (1987), Hiltz (1983), Johansen and DeGrasse (1979), Kaplan (1987), Lipinski, Adler, and Shoemaker (1982), Lucas (1978), Maude et al (1984), Mills (1983), Nisenholtz, Morphos, and Mason (1983), Oeffinger (1987), Paisley, (1983), Palme (1984a), Peiper, (1984), Peiper and Williams (1982), Rice (1982), Rice and Paisley (1982), Rice and Case (1983), Saunders and Heyl (1988), Smith (1988), Taylor and Naffah (1982), and U.S. Department of Commerce (1984) all make their observations from limited duration computer conferences involving small numbers of participants.

Hiltz et al (1981), Morrell (1983), Keisler, Siegel and McGuire (1984), Thorn and Connolly (1987), Smilowitz, Compton, and Flint (1988), and Matheson and Zanna (1988) all observe limited duration experimental trials involving small numbers of participants. Pullinger (1986) observes a specific computer conference over the course of 4 years, but the conference remains small to the end, and the results make it apparent that "critical mass" was never achieved.

The best of the rest of the literature, including Bosomworth and Foulger (1982), Catchings (1985), Cross and English (1983), Daney (1985), Dizard (1982), Ducey (1983), Englebart (1963, 1973; 1982a, 1982b), Foulger (1981 a, 1981b, 1981c, 1981e, 1981g, 1981h, 1982c, 1982d, 1982i), Haight and Rubinyi (1983), Jordan (1984), Lancaster (1982), Marvin and Winther (1983), McLuhan and Powers (1981), Pyne (1988), Mortensen (1987), Pfaffenberger (1986), Saporito (1987), D. Schiller (1982), and H. Schiller (1978), Turoff and Hiltz (1987) and Tucker (1983), is usually little more than informed guesswork.

Keisler, Siegel, and McGuire (1984) summarize the situation eloquently as they review the literature of conferencing:

The functions and impact of computer-mediated communication are poorly understood. Critical information (such as who uses it for what purposes) is lacking, and social psychological significance is controversial ... (p. 1123)

It can be claimed that the scholarly literature surrounding the effects of computer conferencing has barely moved since 1982. The reason for this lack of movement, in all probability, is that there has been little real data to advance with. Moody (1987) estimates that "in 1985, probably fewer than 10 companies had in-house conferencing capabilities". Of the vendor sources of computer conferencing timesharing services Moody cites, moreover, only one, CompuServe, had anything resembling large, ongoing computer conferences at that time.

Recent Literature

Recent growth in the use of computer conferencing has started to change this situation, however:

• Foo et al (1988) reports, in a study of the use of computer conferencing among HIV/AIDS researchers over a two year period, that the medium appears to be having an important impact in increasing communication among researchers and facilitating the exchange of information.
• Rice and Love (1987), in a content analysis of 338 messages contributed to a single computer conference over a six week period, find a high level of socio-emotional content. Nearly 30% of sentences sampled involved such content.
• The broad distribution and growing use of the Minitel system in France is increasingly demonstrating that computer conferencing and other computer media can be successfully provided to a mass market. Marchand (1988) reports extremely rapid growth in both usage and available services between 1984 and 1988.
• Adrianson and Helmquist (1988) were able to identify 370 highly experienced ongoing users of the COM computer conferencing facility. A survey of this group indicates that the most satisfactory applications of computer conferencing for this group included giving and receiving information, exchanging opinions, and generating ideas. The least important applications included decision making, bargaining, getting to know people, persuasion, and resolving disagreements.
• Hiltz (1988) was able to track 500 actual users of 4 computer computer mediated communications systems for four months. The results of this study indicate substantial productivity increases for users during this period.
• Murray (1986), in a survey of a large number of experienced users of computer media in IBM, that 86% of respondents spent at least 10% of their working time engaged in computer mediated communication and 28% of respondents spent more than 30% of their working time on CMC.

These studies, which generally entail fairly large numbers of experienced computer conferencing users interacting over a significant period of time, remain isolated instances, however. Based on the existing literature, the effects of computer conferencing remain more guesswork than clear cut result.

Practices

A small but growing literature explores the practices associated with computer conferencing and computer mediated communication. Observations concerning practices starts, once again, with Hiltz and Turoff (1978), who offer the following practice-related hypotheses:

• Computerized conferencing as a social and psychological process differs markedly from other modes of communication, such as face-to-face meetings, telephone, or letter writing. (p. 76)
• In computerized conferencing, a new system of "conventions and procedural rules" emerges among participants, who must master them as well as the mechanics of the system before they feel comfortable and accomplished on the medium. (p. 82)
• What may seem an inadequate set of cues in computerized conferencing for novice users may later be overcome by participants learning how to substitute for missing kinds of cues. (p. 88)
• With time, it can be expected that ... a kind of collective "subculture" will develop, ... as well as some shared norms and understanding about etiquette and level of participation, such that the observed behavior will be much more regular ... than has occurred ... thus far. (p. 90)
• Participants in a computerized conference are physically alone. While they learn to concentrate more and more on sending and receiving typed cues, they almost immediately drop the usual "face-work" or maintenance of "front" that would occur in a face-to-face meeting. (p. 94)
• In computerized conferencing, there will be less of a tendency for a single dominant individual to emerge, and that this contrast in degree of dominance will increase as the group size increases. (p. 107)
• In computerized conferencing, there are more likely to be multiple leaders, each specializing in, and deferred to for expertise in, a particular aspect of the problem. (p. 107)
• Knowing that computer-mediated communication seems "cold" or "machinelike" or "inhuman," most new users seem to compensate heavily for this by engaging in very strong efforts to be warm, friendly, and personal. (p. 111)
• The "chilling effect" of rank on disagreement with a bad idea or decision advocated by a high-status person will not be operative in computerized conferencing. (p.114)
• Lack of a "critical mass" within a conference or group can doom it to failure. This critical mass is related to both the number of active participants and the number of geographical locations. The minimum seems to lie somewhere between 8 and 12 active participants in three of more geographical areas. Below this critical mass, there are not likely to be enough new messages or conference comments entered, so that there would not always be new items to receive or respond to when a member signed on. Above the minimum size and dispersion, enough activity and controversy is generated to motivate members to sign on frequently and to participate actively in the exchanges. (p.122-123)
• A new user is usually passive, receiving much more than he or she sends relative to experienced users. A new user is more likely to send messages than engage in conference activity.(p.123)

This is a diverse set of hypotheses. It suggests that users of computer conferencing will construct rules, roles, norms, and styles of interaction. It suggests that these constructions will differ from analogous constructions in such media as face-to-face communication, telephone, and letter writing. It suggests that these rules will emerge in reaction to the problems they encounter in using the medium, including inadequate nonverbal cues and disjoint interaction patterns. It suggests that these rules will cover such things as how often people participant and how they should behave. It suggests that one result of these mutually negotiated rules will be the development of an on-line community.

Writing Style

These hypotheses have been extended in subsequent literature. Spitzer (1986), for instance, documents a variety of writing conventions that seem to have evolved specifically to aid communication in the computer conferencing environment. Most of these conventions involve the clever use of punctuation to indicate affect. Levinson (1986) extends this with the extension that McLuhan's "holographic" writing style has much in common with the style of contributions to computer conferences. These contributions tend to be short, between 20-60 lines, and wildly intermingled with other contributions on different topics.

Sherblom (1988) observes some generic variations in the use of signatures in electronic mail. A content analysis was performed on 157 electronic mail files received over the course of several months by a middle manager in the computer services department of a large organization. It was found that subordinate employees were significantly more likely to sign their electronic mail (write their name at the bottom) than were superiors. Olsen (1989) refers to these variations as "genre conventions".

Etiquette

Smith (1988), based on a limited duration observation of student use of a computer conferencing system, suggests that there are few rules associated with computer conferencing, particular in comparison with other media. The author suggests that a stronger CMC etiquette might increase the educational value of computer conferencing. This suggestion responds, in part, to results obtained by Kiesler, Siegel, and McGuire (1984), who found that computer mediated communication behavior was substantially more uninhibited than face-to-face communication.

Uninhibited, in this case, reflects a high frequency of remarks containing swearing, insults, name calling, and hostile comments. This characteristic uninhibited behavior has been labeled "flaming". Kiesler, Siegel, and McGuire report that this uninhibited behavior is quite general, and is observed among students, non-students, and working adults. There are no reports in the literature countering this observation at this writing. Smith's (1988) suggestion of rules of etiquette points up a possible restriction on the generalizability of Kiesler et al's results, all of which were obtained in short duration experimental groups. There may not have been sufficient opportunity for rules to evolve in these short meetings.

Learning Plateaus

Turoff (1989) suggests that individuals and groups pass through at least four plateaus as they learn to use computer conferencing:

• learning the mechanics of the system.
• learning how to be an effective communicator in the electronic written medium. This entails, among other things, developing a style that can express socio-emotional content using paralingistic cues.
• understanding the processes and techniques that foster a particular communication objective. This is where groups establish rules and roles.
• understanding the system, in function and use, sufficiently well so as to begin to suggest how the medium can be improved. There is no evidence that this stage ends. It is here where users become a part of the design process.

One notes that all but the first of these objectives involve mastering practice. The first entails understanding the mediators which allow you to enter and control the system. The last provides for making changes to the software mediators; providing the essential mechanism required to power the circular process of media evolution.

The current study

If there is, from the perspective of this study, a recurring fault in prior studies of computer conferencing to date, it is in their time scale and scope:

• Both Hiltz and Turoff (1978) and Smith (1988) suggest that new users of computer conferencing do not have a feel for the etiquette of computer conferencing. If correct, it seems clear that a reasonable study of computer conferencing must entail adequate time for this etiquette to be realized.
• Turoff (1989) suggests that there are four distinct stages to the process of mastering computer conferencing, with the user able to competently negotiate etiquette at the fourth stage. If correct, it seems clear that an understanding of what computer conferencing may become depends on studying computer conferencing participants that have reached this fourth stage.
• This fourth stage frequently entails a change in the structure of mediators that, in effect, reconstructs the system. If the formulation is correct, understanding this iteration, its effects on the medium, and its effect on users, will require study of an ongoing and developing computer conferencing facility that has a stable core group of users for an extended period of time.

There is no study in the current literature of computer conferencing that meets these specifications. It is hoped that this study, which has observed a single ongoing and developing computer conferencing facility that has had a stable (if expanding) core group of users for six years (eight years if one counts archival data), will help to fill the void.