Abstract:
Text based computer mediated communication (CMC) has recently been the focus of many ethnographic studies by social scientists. In my own research of cybersex, I followed the lead of these researchers and utilised ethnographic methods but encountered several significant difficulties. These difficulties include the lack of parameters for users of text based virtual environments, the necessity of online interviews rather than face to face ones, and the frequent misinterpretations that occur due to the narrow bandwidth of text based CMC.

Keywords: Cybersex, Ethnographic Methodology, Virtual Communities, Computers and Society, Internet - Social Uses

In recent years academic researchers have written extensively about computer mediated communication (CMC). A significant amount of this research has looked at the ways in which people use text based CMC to chat with each other in real time on the Internet and on socially oriented online services such as America Online (AOL) and CompuServe. In these studies, researchers have found that text based virtual environments (chat rooms, IRC chat channels, and MUDs) are places where users can experiment with identity and gender (re)construction (Reid 1991, 1994; Turkle 1995), form new friendships (Baym, 1996), and join together with other users in the building of virtual communities (Rheingold 1991, 1995; Lichty 1994). Most of the existing social scientific research of the online world has been ethnographic. Given the prevalence of ethnographic methodology in the study of social phenomenon in text based virtual environments, it is surprising that it's use in cyberspace has yet to be analysed in any great detail.

In this article I will discuss the use of ethnographic field research methods in cyberspace, focusing more specifically on my own ethnographic research of the online sexual practices of the inhabitants of AOL chat rooms. Below I will briefly describe this research to give readers an idea of the focus and findings of the study which I undertook. Following this, I will discuss several of the problems encountered by researchers of cyberspace. These include the unavailability of data on the parameters of the population and the necessary dependence upon online interviews and questionnaires rather than face to face interviews. I will also discuss one of the dangers faced by social scientists who do investigations entirely within the constraints of a text only medium, the ease with which misinterpretations can be made there.

The locational focus of my study was online chat rooms, virtual rooms where multiple users of an online service (in this case AOL) can "chat" by sending each other both public and private text messages. In these chat rooms, people are able to meet and chat with others who share similar interests but who may be geographically located hundreds or even thousands of miles away. Just as in other places where people get together socially, some of those who meet in online chat rooms later choose to have sexual relationships with each other. In the case of those who meet in online chat rooms, these sexual relationships are most often acted out online in the form of cybersex chat.

In my research, I have identified two forms of cybersex chat that occur in AOL chat rooms. The first form is computer mediated interactive masturbation in real time. In this form of cybersex, users type instructions and descriptions of what they are "doing" to each other and to themselves while masturbating. The second form of cybersex chat is the computer mediated telling of interactive sexual stories (in real time) with the intent of arousal. These sexual stories are sometimes based on reality, sometimes on pure fantasy, either way they are almost always very detailed and highly erotic. (Hamman, 1996) These forms of cybersex, according to Nguyen and Alexander, are often satisfying enough that they can "evoke physical orgasm" for many of the participants. (Nguyen & Alexander, 1996, 116)

In my research of cybersex chat, I decided to use an ethnographic approach, following the precedent of some other important studies of cyberspace (Reid 1991, 1994; Turkle 1995; Baym 1996). Baym suggests that, "Rather than focusing on building predictive models of CMC [Computer Mediated Communication], more naturalistic, ethnographic, and microanalytic research should be done to refine our understanding of both influences and outcomes." (Baym, 1996, 161)

Ethnography is defined by Marshall as "the acts of both observing directly the behaviour of a social group and producing a written description thereof." (Marshall, 1994, 158) In ethnography the "description of cultures becomes the primary goal... the search for universal laws is downplayed in favour of detailed accounts of the concrete experience of life within a particular culture and the beliefs and social rules that are used as resources within it." (Hammersley & Atkinson, 1995, 10)

Researchers who have used ethnographic methods in cyberspace have been confronted with several problems that are different from the ones they are likely to encounter in research off-line. These problems, which will be discussed below are: locating the parameters of the population of study, whether or not to depend on online interviews, and the frequent misinterpretations caused by the absence of physical cues and gestures in text based virtual environments.

We know that the use of the Internet and online services such as AOL is growing rapidly, and that the language of these virtual inhabitants is almost always English. We also know that the use of these technologies requires not just literacy, but computer literacy. On AOL, there are methods of obtaining data on the number of people using a specific chat room and of determining the total number of chat rooms at a given point in time. On AOL, there is also a way to access a "profile", a personal biography stating characteristics such as age and gender as well as listing hobbies and other interests, for chat room participants who wish to make their personal details public1

.

Unfortunately, our data on the parameters of the population of online chat room users is limited to the above. We don't know the age, race, or gender of chat room users unless they make that infor mation available to us. We don't know how many people, over an extended period of time, use online chat rooms. There is no data telling us how long each individual user spends engaged in online chat and we don't know at which times they are likely to come and go. In fact most of the demographic information that we do have about users of online chat rooms is self-reported and unverifiable. Online services I have contacted are unwilling to supply academic researchers with demographic data since data of this type is a closely guarded trade secret which could be used by competitors if it were made public.

An important question faced by researchers of online chat rooms is whether to depend solely upon online interviews and observations in the gathering of data. Sherry Turkle notes this problem, stating that, "virtual reality poses a new methodological challenge for the researcher: what to make of online interviews and indeed, whether and how to use them." (Turkle, 1995, 324) Turkle chooses not to use online interviews in her research unless she has additionally met that person in real life. However, Turkle maintains that her choice not to use online interviews is based on the focus of her work, not for fear of fundamental flaw in the method. (Ibid.) Others suggest that there are certain advantages to interviewing people in their own environment. According to Hammersley and Atkinson, "interviewing them [respondents] on their own territory... is the best strategy. It allows them to relax much more than they would in less familiar surroundings." (Hammersley & Atkinson, 1995, 150 brackets added) Online interviews provide the researcher of cyberspace the opportunity to observe people in their own surroundings. This can itself provide important data, such as giving the researcher an idea of the technical prowess of the respondent. It also, as Hammersley and Atkinson note above, allows the respondent to feel more at ease during the course of the interview. In fact, most of my respondents admit that they would not talk with me about cybersex (and the other issues it brings up such as solitary masturbation) if I were to interview them face to face. In contrast to this, in the online interviews that I completed, I found that nearly all respondents were almost immediately willing to speak about very intimate details of their sex lives2

.

Another problem with face to face interviews is locating suitable respondents away from the location where cybersex chat occurs. I did attempt to use other interviewing strategies before settling on the use of online interviews. These included attending a computer user group meeting and speaking with people at cyber-cafes, both of which failed because I was unable to locate users who reported that they had previously engaged in cybersex. A similar problem with attempting to undertake face to face interviews with members of this population is that, due to the geographically disperse nature of computer networks, respondents may be physically located far from the researcher. To locate people who have cybersex in online chat rooms, and in order for them to feel comfortable enough to speak about such activities, online interviews are a necessity.

Stone makes an important distinction between face to face communication and computer mediated communication (CMC), explaining that "Reality is wide-bandwidth, because people who communicate face to face in real time use multiple modes simultaneously - speech, gestures, facial expression, the entire gamut of semiotics... Computer conferencing is narrow-bandwidth, because communication is restricted to lines of text on a screen." (Stone, 1995, 93) In narrow-bandwidth computer mediated communications, information important for understanding is missing, making " ferocious misunderstandings over simple textual utterances" frequent. (Stone, 1995, 175)

The ease with which misinterpretations of language can occur in text based CMC is of methodological concern as well as of concern to the users of chat rooms who must confront it every time they go online.

Late one night, after hours spent interviewing people online, I decided to venture into my favourite chat room on AOL. I acted in the room not as a researcher, but as a participant looking for some good conversation before logging off for the night. I typed back and forth with some friends, feeling more awake as I did. As my friends began to log off, I decided that I would try to do one or two more interviews.

I asked aloud in the chat room if anyone wanted to talk with me about their experiences with "online intimacy and online relationships" and received replies from two women. I already knew Annie, one of these respondents, from having chatted with her about music several times previously. Since I did not feel comfortable interviewing someone who I was socially involved with, I focused my attention on Rebecca, the other respondent.

Rebecca is a third year history student at a well known and respected university on the Eastern Seaboard of the United States. We chatted in a friendly manner for a few minutes prior to the start of the formal interview and I quickly found her to be intelligent, well spoken, and friendly. I asked Rebecca to wait for a minute while I typed her a message about anonymity and what to expect in the interview. "Wonderful!!!!!!!!!! : )" was Rebecca's reply, betraying her enthusiasm to be interviewed by me. I asked myself why a person would be so excited to be interviewed and came to the conclusion that perhaps she had other ideas in mind. I decided that I would not ask Rebecca specifically about cybersex, concentrating instead on other issues of online communication. This way, I hoped, I cold gain valuable data without the risk of giving the enthusiastic respondent the idea that I might engage in some participant observation of cybersex with her as my partner3

.In order for me to proceed with an interview or to speak with anyone online other than informally, I first electronically send them an agreement which they must read. It states that respondents will remain anonymous in everything I write, and that I will not have cybersex with them at any time before, during, or after the interview. The part of the agreement about anonymity helps to relieve any fears of exposure the respondent may have. The second part of the agreement, stating that I will not have cybersex with respondents at any time, helps to set the tone of the intervi ew. It makes it clear to the respondent that I am asking the questions I do for research purposes, not for my own arousal. The message I sent to Rebecca included the passage "Since I was in that room just having fun and not researching, I can't really ask you about cybersex without it seeming funny, so I won't". This was added due to my fears that we could have ended up off task, and possibly in ethically dangerous territory, if we were to discuss cybersex. Her reply, unbroken by me, went like this:

Rebecca: yea
Rebecca: But what about cyber??
Rebecca: I don't mind answering any questions.
Rebecca: : )

It was late at night and I was, in hindsight, too tired to practice good judgement in difficult situations. When Rebecca, who seemed overly excited to be interviewed by me, said "But what about cyber ??" I thought that I knew exactly what she was asking. I questioned whethe r she had really meant it, feeling very flattered by even the suggestion, since, after all I did find Rebecca's personality to be very attractive. My mind raced with the few details that I did know about Rebecca, and I found myself wanting to have a romantic conversation with her, possibly leading to something more.

Fighting the very real urge to forget about the interview in order to interact with her on a more social level, I managed to inform Rebecca that she had shocked me by asking me to have cybersex. I explained that I had to keep the line between my researcher self and my pleasure self separate and that I was afraid that my objectivity and my research ethics could be severely compromised if we continued with the interview at that time. She responded, saying only that she could tell me "yucky stuff too". I became more confused about Rebecca's intentions and asked her if she wanted to tell me "yucky stuff" for purposes related to my research or if she wanted to tell me about it for other reasons known only to her. Then she asked me if I was really doing research at all. This went back and forth for a while and we didn't seem to be getting anywhere, so I explained more clearly that I was indeed doing research on cybersex, but that I did not like the idea of continuing an interview with her since she had so abruptly asked me to have cybersex with her.

There was no typing for about 30 seconds and I began to wonder if my computer had "froze" or if Rebecca had logged off of AOL. When text finally did arrive on my screen, I learned that Rebecca had innocently asked "But what about cyber?" because she wanted to know when I would be asking her about her cybersex experiences. Rebecca explained by saying, "I merely said that I would help you out with your interviewing. I WASN'T asking you to cyber...." . Rebecca was not flirting with me4 , she was trying to be helpful.

For a combination of reasons, I had misinterpreted her words as a sexual proposition. Some of this misinterpretation can be blamed upon my being tired at the time of the interview. But it was the play of the narrow-bandwidth of computer mediated communications upon my own desires that was the main reason for my misinterpretation of Rebecca. I am confident that, had I been able to see Rebecca's body language as she spoke, I would have known that she was trying to tell me that she was willing to talk about cybersex, not suggesting that we have cybersex. The narrow-bandwidth of the computer medium caused me to make gross misinterpretations of Rebecca's words and intention s because important information was missing from our communication. I had replaced the missing information with my own idealised version of events and meanings, leading me to believe that Rebecca had propositioned me.

This experience is a revealing one. It demonstrates how I carried certain expectations, needs, and desires into the interview. This often occurs in more traditional face to face research as well, the difference being that in text only cyberspaces, there are no visual or audible cues to protect against misinterpretations as there are in the physical world. Researchers working in text based virtual environments need to pay close attention to their own subjectivity or the narrow-bandwidth effect will cause them to completely misconstrue an interaction like was done in the experience I tell of above. On other occasions5 , I had similar misunderstandings with people as I interviewed them online, all of which would have been easily avoided in a face to face setting.

There is already a tradition of ethnographic research in text based virtual environments such as online chat rooms. Doing online interviews and observations, rather than face to face ones, is necessary when researching cybersex chat for a number of reasons. First, it allows respondents to remain in a place with which they are familiar and where they are confident that their true identity will remain unknown. Second, and more importantly, it is the only way that researchers with limited time or resources can gain access to an online population for study, especially considering that the parameters of this population are relatively unknown. Despite it's necessity in many cases, completing ones research entirely on the basis of data obtained online is problematic. Gross misinterpretations of language can occur quite easily within the limited cues narrow bandwidth environment. The effect of a researcher's own desires upon interaction with respondents becomes stronger, not weaker, in a cues filtered out environment such as online chat rooms. To avoid misinterpretations like the one I speak of above, researchers who conduct online interviews should always seek to confirm the meaning of what respondents say. Online interviews may be a good way to elicit sensitive data and to gain access to online research populations, but researchers must be keenly aware that misinterpretations of language are frequent in the narrow bandwidth of text based cyberspaces.

Baym, Nancy K. "The Emergence of Community in Computer-Mediated Communication." in Cultures of Internet: Virtual Spaces, Real Histories, Living Bodies. Rob Shields (Ed.). London: Sage, 1996. 138 - 163.

Lichty, Tom. The Official America Online for Macintosh Tour Guide. Research Triangle Park, North Carolina: Ventana Press, 1994.

Hamman, Robin. Cyborgasms: Cybersex Amongst Multiple-Selves and Cyborgs in the Narrow-Bandwidth Space of America Online Chat Rooms. Unpublished Dissertation, University of Essex, 1996.
URL: http://www.socio.demon.co.uk/home.html

Hammersley, Martyn & Paul Atkinson. Ethnography: Principles in Practice. London: Routledge, 1995 (Second Ed.)

Marshall, Gordon (Ed.) The Concise Oxford Dictionary of Sociology. Oxford: Oxford University Press, 1994.

Nguyen, Dan Thu and Jon Alexander. "The Coming of Cyberspacetime and the End of the Polity." in Cultures of Internet: Virtual Spaces, Real Histories, Living Bodies. Rob Shields (Ed.). London: Sage, 1996. 99 - 124.

Reid, Elizabeth M. "Electropolis: Communication and Community On Internet Relay Chat." 1991 Honours Thesis, University of Melbourne.
URL: http://www.ee.mu.oz.au/papers/emr/index.html

-----. "Cultural Formations in Text-Based Virtual Realities." 1994 Masters Thesis, University of Melbourne.
URL: http://www.ee.mu.oz.au/papers/emr/index.html

Rheingold, Howard. Virtual Reality. New York: Summit Books, 1991.

-----. The Virtual Community: Surfing the Internet. London: Minerva, 1995.

Shields, Rob. (Ed.) Cultures of Internet: Virtual Spaces, Real Histories, Living Bodies. London: Sage, 1996.

Stone, Allucquere Rosanne. The War of Desire and Technology at the Close of the Mechanical Age. London: MIT Press, 1995.

Turkle, Sherry. Life on the Screen: Identity in the Age of the Internet. New York: Simon and Schuster, 1995.

1 I've found that mo st people using AOL's chat rooms do have a "profile" and that other users often look at this before entering int o a conversation, and especially before having cybersex. Some users also exchange electronic photographs of themselves with their cybersex partners and online friends.
2 Researchers of off-line sexuality may wish to investigate the use of online interviews in their studies. It is my feeling that the anonymity of online interviews would allow respondents to be more candid than they are in face to face interviews, although this remains to be tested.
3 I do not feel that participant observation of cybersex is ethical.
4 Later I realised that Rebecca, a heterosexual woman, had wrongly assumed that I was also a heterosexual female (I am, in fact, a heterosexual male). Rebecca had also made errors in her interpretations of me due to the narrow bandwidth of cyberspace.
5 Race, gender, and age were misread or wrongly assumed on a number of occasion during this and subsequent research.