Justification for qualitative inquiry

The difficulties in producing a reputable quantitative study of MUD phenomena are formidable, unless one resigns oneself to convenience samples based on replies to surveys posted to various MUD-related Usenet newsgroups. One might improve this sampling technique by getting a list of all known MUDs, randomly sample from this list a set of all MUDs to study, and get a player list of each of the chosen MUDs. Then, one might randomly sample from the list of chosen players and conduct a previously pilot-tested survey (Schwartz, 1995).

Though the aforementioned quantitative method may seem airtight on its face, there are onerous difficulties at nearly every stage of such a procedure. To name a few, there is the problem of defining what is meant by "all known MUDs"; if one narrowed the definition to "all publicly advertised MUDs", this still fails to address the fact that MUDs come and go, week to week. In fact, public MUD lists often range from 350 to 650 MUDs listed. Another difficulty arises when asked to make a list of all players on a single MUD. For instance, should every person who ever logged in to the MUD be included, even if they only spent a few scant minutes in the environment? Also, it is possible that numerous characters are played by the same person-- how many times will such a person be counted?

Given the statistical hazards of such a research scheme, the present research has been guided by the methods of participant observation and informed by Philipsen's tradition of ethnographic research on speech communities.

      [The researcher should] specify a phenomenon of interest,
      link that phenomenon conceptually to the process of communication,
      and specify a framework for describing that phenomenon in its 
      particularity in any given social field, and that the descriptive
      framework itself will be subject to revision contingent upon 
      the results of the field work.
                                         (Philipsen, 1977, p. 48)

Indeed, the researcher's fieldwork and experiences with MUD communication, coupled with numerous informal interviews and a review of both popular and technical literature, provided a rich tapestry of communicative data upon which to reflect and evaluate in light of previous work on nonverbal communication.

Reliability

LeCompte and Goetz (1982) posit several hindrances to the reliability (the degree to which another researcher would make similar discoveries in a speech community) of an ethnographic study. The most important of these are researcher status position, choice of informants, social conditions, and analytic constructs and methods. Each of these potential snags will be addressed below.

If a researcher spends more time within certain subgroups of a speech community than others, s/he may gain a less than complete understanding of the interrelations of those various subgroups. Likewise, if a researcher belongs to a certain class of humanity (e.g., black female academic), she may be received by the speech community in a different way than were she to be some other class. In short, a researcher's role in the speech community has an effect on the phenomena to be studied.

One source of concern for reliability in this study was the fact that the researcher attained a position of authority and recognition on one of the three MUDs being observed, Law Arch of Paradox II. As such, there was some consideration of the possibility that the researcher's words and deeds might be altering or even creating the phenomena being observed. However, part of the position attained was the ability to make oneself completely invisible to all other participants. While invisible, no significant differences in the behavior of the other participants were noted. In addition, while this position was noteworthy on one of the three MUDs studied, the researcher held no such position on the other two; again, no significant differences in the behaviors of the other participants was noted. The ease with which the author could define his own social status while assuming multiple identities was an affordance perhaps peculiar to MUDs and their ilk.

An additional reliability concern for researchers is whether one's informants are appropriate spokespersons for the community under scrutiny. Clearly, if the chosen spokesperson is merely a peripheral member of the research group, their reports are less than likely to be indicative of the thoughts and feelings of the group's core members. However, the researcher's ample time spent in the field effectively nullifies the possibility of such a poor choice of informants.

LeCompte and Goetz also state that the social conditions in which a researcher collects qualitative data can have effects upon that data. Whether the information is collected in formal individual interviews or as part of a casual conversation in a bar (Stout, 1995) can have important effects upon what sorts of information the informant is willing to part with. In this study, the primary mode of data collection was simply observation of public behavior, but occasionally E-mail (or intra-MUD mail) would be exchanged (which tended to be more verbose and specific), and sometimes directed MUD conversations or interviews would be utilized. This variety of communicative contexts ensured that the phenomena observed were not simply spawned of the particular context from which the data were culled.

A final means by which qualitative researchers may benefit their research design, thus ensuring that future researchers are able to make meaningful comparisons with their own work, is to be careful to explicitly state and define their analytical constructs and make clear their descriptive framework. In so doing, the researcher irons out any ambiguities which future ethnographers might have perceived in making comparisons, hence, improving reliability. In this thesis, care has been taken to explicate each of the nonverbal forms and functions which shall be examined as applied to MUD interactions. In addition, a section below shall make clear the specific methods whereby the research was carried out.

Notably, one of the means which LeCompte and Goetz suggest as an excellent reliability improving tactic in ethnographic research is to mechanically record as much data as possible (LeCompte & Goetz, 1982, p. 43). For this thesis, through screen copying and various logging mechanisms, this strategy was inherent in every step of the research.

Validity

Validity was further assured in this study through a variety of means. First, as it purports to examine nonverbal communication on that class of text-based virtual realities referred to as adventure MUDs, a variety of such MUDs were studied; The three MUDs in question were different in theme, size of participant group, history, and conventions of syntax and programming. As LeCompte and Goetz (1982) state, "[external validity] is addressed to an extent by multisite ethnographic designs," (p. 51).

The fact that the 2000 hours of observations took place over more than two full years was an additional validity-ensuring detail. This study is by no means a depiction of a moment frozen in time, but rather a continued effort at analytic observation of a specific genre of communicative environment.

LeCompte and Goertz caution against "research exhaustion", in which the researcher "goes native", and loses the ability to discern peculiarities of the culture being studied. Given the thousands of hours the author spent in the field, this was a concern at times, and it was only through regular meetings with his thesis director (who maintained a careful naivete with regards to the modalities of MUD interaction) that these pitfalls were avoided.

Validity was further bolstered by the long term observations of the MUDs examined. Over the course of nearly three years, the populations of the three MUDs was "recycled" every five to seven months. In other words, the group of participants who frequented each MUD, while retaining some core members, changed periodically, as players lost Internet access, stopped MUDding altogether, or simply moved on to explore new MUDs. This mortality effect, while noted, had no significant effect on the observed behaviors on any of the three MUDs studied.

Finally, notes were kept throughout the research process. This was often easily effected, as an intra-MUD E-mail system existed on all three MUDs being studied, and afforded the researcher a convenient method of information storage and retrieval. In addition, the multi-tasking facilities of modern personal computers made it especially simple to record examples, nuances, reflections, and commentary.

Participants

Though the majority of MUDders reside in the United States, the broad spanning arms of the Internet assure that MUD participants come from all over the world. Rosenberg (1992) lists the countries that the players on his favorite MUD, WolfMOO, hail from: Canada, Ireland, England, Germany, Netherlands, Italy, Switzerland, Denmark, Russia, Iraq, Finland, Hong Kong, Indonesia, Malaysia, Singapore, New Zealand, Israel, Australia, Sweden, Brazil, and the United States. Informants from Ancient Anguish, one of the MUDs studied, live in such places as British Columbia, Holland, Mexico, Nova Scotia, Australia, Spain, Italy, Argentina, and the United States.

So, while it is possible to generalize about the make up of the MUDding population (and more importantly, the make up of the participants in this study), it is not possible to put forth definitive figures. Nevertheless, one common thread that certainly runs through all of the participants on the MUDs studied is that they have Internet access and they have voluntarily connected to an adventure MUD.

Data Collection

Others' analyses were gleaned from both popular literature as well as scholarly sources (within disciplines such as communication, computer science, English, psychology, and cultural studies; these sources are reviewed in Chapter 1). Most sources were found in non-print media, i.e., they were downloaded from various personal and informational FTP (File Transfer Protocol) and WWW (World Wide Web) sites on the Internet.

Four Usenet newsgroups on the Internet were monitored closely: rec.games.mud.admin, rec.games.mud.lp, rec.games.mud.announce, and rec.games.mud.misc. Given the subject of this inquiry, it was fortunate that these newsgroups existed, and that they were so active. Surveys regarding MUDders' attitudes about the reality of their virtual experiences, MUD romance, and MUD communication were posted from time to time by other researchers, and MUDders would often make their replies public, which became an additional source of qualitative data. When particularly interesting material appeared, the researcher E-mailed the individual directly. In such an E-mail, the purpose of the study was revealed, and a few questions about the content of the individual's posting would be asked. Some responses were surprisingly candid and detailed, and provided an invaluable window on MUD relationships. Other material found among Usenet postings included songs about MUDding, poems about MUDding, and debates on how seriously one should take MUD interaction (see appendices). Needless to say, these artifacts found among Usenet newsgroups became an important source of qualitative data, above and beyond the actual fieldwork.

Clearly, though, with over 2000 hours of time spent in MUD environments, it was the field work that the researcher was most able to draw upon in his investigations. Entry was first gained in September of 1993, which began a learning process that would continue to develop for two and a half years. This first MUD experience, on the public adventure MUD called Ancient Anguish, is described at length in Masterson's (1995) ethnographic work. A more succinct depiction shall be provided here.

One benefit of beginning field work in MUD environments is the ease of entry. While it may take quite a while for a reporter on culture to gain the trust of a group of Neo-Nazi Skinheads (to such an extent as they engage in all their typical behaviors in the presence of the researcher), such hindrances do not molest the MUD researcher. Indeed, all that is required to become an active member of a MUD culture is a computer with a modem, time, and modest typing skills.

It is true that MUD environments are baffling at first to the completely uninitiated, but this bewilderment soon gives way to compelling fascination and curiosity. One of the first realizations that one makes, as one becomes enculturated in a MUD, is that the stream of messages on one's screen is being produced by real people all over the world. A common reaction to this realization is to wander around the MUD asking people where they are, who they are, how they found out about MUDs, etc.. Such social activity brings about the rapid enculturation of newcomers to the virtual environment, for friendly, curious, unprompted communication among participants is generally encouraged and even expected.

The researcher spent time on three separate adventure-style MUDs throughout the two years of the study, one strictly medieval in theme (Ancient Anguish), one with a multitude of acceptable themes (Paradox II), and one based on a popular science fiction theme (TrekMUSE). At least 200 hours were spent in each environment, during which the researcher familiarized himself with the conventions of the environments and effectively became a part of each culture.

As a participant observer, the researcher was careful to maintain some distance from the phenomena being observed, especially during the second year of the study, as the project's direction solidified. In other words, care was taken not to be completely engaged by the events that transpired in the virtual environment. In so doing, it became possible to have first hand experience with the phenomena in question, without abandoning the primary jobs of the researcher, that of observation and analysis.