Ethics of Online Research: Considerations for the Study of Computer Mediated Forms of Interaction Dennis Waskul By Dennis Waskul Computer mediated communication, the Internet, and related services, represent far more than simply a medium of communication. Cyberspace is more than a neutral delivery system--more than just a channel for conveying information from one location to another. Online communication is simultaneously a medium and an environment in and of itself. It is a context where the medium (computers, modems, and other technology) and the social environment (the setting or context) become situatedly intertwining elements of communication. This unique form of social interaction transforms social spheres into new social environments with new patterns of action, feeling and belief. This complex techno-social phenomena has brought people into contact with others in new ways, where new forms of interaction in conjunction with new contacts between people have culminated in virtual communities at an amazingly rapid rate. Not surprisingly, this phenomena has not gone unnoticed by academia--particularly in the fields of social and behavioral sciences. An unknown number of researchers have and continue to conduct research on the net. As a result, a significant amount of research has already been published, considerably more is in the works, and there is every reason to believe that net-related academic research will grow and expand exponentially in the years to come. Academic research on the Internet (and related services) offers great potential for expanding understandings in a wide variety of academic fields. Undoubtably, such research will present conceptual, theoretical and methodological challenges--the resolution of which represents the stuff of academic development. However, academic advancement is not always s ynonymous with the best interests of participants. Thus, ethical issues surrounding such research must be carefully considered and not left as a matter of post-facto consideration. Responsible and ethical research is not a matter of codes, policy or procedure. Rather, responsible and ethical research centers on a commitment to protect the participants of ones study from potential harm. Such a commitment necessitates a keen eye on the impact of research on the individuals involved, the context of the situation, the interpersonal dynamics of the group, and the implicit, as well as explicit, structure of group interaction. In most cases this commitment entails the acknowledgement that ones research will become a part of the phenomena, and therefore theresearcher must carefully consider how this will impact the experiencesof participants. Such deliberations provide a foreground for making informed ethical decisions. Ethical decisions that adhere to the commitment to participants, placing the interests and experiences of the participant over and above the interests of the researcher and the aims of research. Ethical codes, policy, and procedure systemize ethical considerations, providing guidelines for the researcher--they are not, however, safe havens to hide from the responsibility of a sound ethical choice. Obviously, online research is not exempt from this ethical commitment. In fact, the very newness of virtual communities and online interaction as a topic of academic research should encourage a heightened awareness of ethical issues (including the possibility that codes of ethical conduct may not be entirely applicable, and may need revision). This paper will explore ethical considerations of online research, identifying a variety of ethical issues that emerge from the dynamic form of interaction that defines computer mediated communication, concluding with suggestions for an ethical approach to the study of online interaction. Ethical Considerations and the Form of Computer Mediated Interaction Online communication is a unique form of interaction that qualifies participants as a special population (King 1995). It is a form of interaction situated within an electronic-computer medium that has changed the significance of time, space and physical location as interaction variables. Indeed, it is the form of online interaction that has transformed social spheres into new environments, with new patterns of social interaction, belief, and feeling. Furthermore, characteristics of this unique form lend a special character to the problem of ethics. For this reason, ethical issues will be identified through discussion centered on a hand full of ideal-typical elements that define the form of online interaction. Privately-Public and Publically-Private Form of Online Interaction: Online interaction defies locality. It is neither public, nor private--rather, it is both. In other words, a defining element of the form of online interaction is that it is situated in a context that is publicly-private and privately-public. Electronic media (as a whole) override boundaries and definitions of the situation that emerge in physical settings. Whether we communicate via telephone, television broadcast, or computer-modem, the situation we are in is only marginally related to our physical location (Meyrowitz 1985). Consider, for example, the common scenario of communicating with another person on the telephone while comfortably resting in ones living room. In this case, the situation we are physically in (the living room) is not synonymous with the situation we are socially in (communicating on the telephone). The social situation challenges the definition of the situation that the physical qualities of the living room alone suggest. Electronic media in general (and computer mediated communication technologies in particular) serve to breakdown the segregation of situations by eroding the time, space and location variables that previously kept them apart (for a full discussion see Meyrowitz 1985). By changing the boundaries of social situations, electronic media do not simply give us quicker or more thorough access to events and behaviors. They give us, instead, new events, behaviors, and situations (Meyrowitz 1985 pp. 45). Unique to these new social situations are the means by which electronic media thrust public issues into personal realms, as they simultaneously thrust personal issues into public realms (Meyrowitz 1985). Electronic media (particularly television and computer mediated communications) allow users to engage public issues, and live events from around the globe, all within the privacy of ones home or workplace. Likewise, personal issues are increasingly cast into public realms (take for example day-time talk shows where it seems no issues are too private for the public realm). As a result, distinctions between public and private spheres become conceptually and experientially blurred, as the medium increasingly embodies neither, but both. The blurring of public versus private experience is particularly characteristic of online communication, and the source of much of the ethical controversy of online research. What is public and private is not always clear, in conception, experience, label, or substance. This highlights the fact that the over-simplified public-vs-private dichotomy of domains of cyberspace refers to accessibility, not the experience of participants. It must be emphasized that labels of public and private are metaphors applied to the context of cyberspace. Cyberspace defies locality. That is, the social reality of cyberspace is not confined to the geographically fixated computer networking technologies that sustain it. For this reason we can only metaphorically speak of cyberspace as a place dichotomized into private and public domains. Public and private metaphors are applicable only so long as we agree that dimensions of cyberspace are like physical places. Hence, public and private metaphorically dichotomizing labels will always breakdown when we attempt to apply them to the experiences of participants. Like any other metaphor, these labels help to clarify some aspects of the phenomena, yet systematically distort our conceptualizations if taken literally. This seems to be the case with cyberspace--because the medium is publicly accessible, some persons seem inclined to define the activities that comprise it as equally public. This is an ethically dangerous mis-conception. Like other forms of electronic media, online interaction allows users topublicly interact from the privacy of a home or workplace. As King (1995) notes, such a situation promotes a perceived sense of privacy. The ability to interact with others from the privacy of ones home or workplace shapes the resulting definition of the online situation. Alone at ones computer, one may extend this situation into interactions online, making it easy for participants to define the situation as private and interact accordingly. More importantly, however, even in situations recognized as public, participants can (and do) engage in private forms of interaction. For this reason, the public-vs-private dichotomy that is frequently used to label domains of cyberspace does not provide a sound basis for making responsible ethical decisions that protect the interests of participants. It is misleading to conclude that online environments labeled or otherwise described as public is necessarily representative of the social definition of that environment. The declaration of some domains of cyberspace as public encourages researchers to perceive such interactions as fair game (within the constraints of copyright laws). Apparently, it is easy for some researchers to legitimize the exploitation of a public domain through simply declaring participants as having a false sense of privacy, or in denial of the public nature oftheir communications. Or worse yet, fail to acknowledge participants altogether, legitimizing the ethicality of their actions through appeals tolaw and/or ethical theory. Such a perspective is intellectually barbaric, and clearly unethical. Simply stated, researchers do not have the right to redefine the contextof the research situation (the experience of the participants) to meet their own vested interests and research aims. To do so is to bring potential harm to the very phenomena that one intends to understand. This is not a vague ethical issue! Removed from the context of online interaction, the issue is more clear. For example, consider the following hypothetical situation: You are seated on a bench in a public park with a group of close personal friends. Small talk quickly dissipates into issues of a more serious nature. In the course of confiding personal and private issues to your friends, you turn your head to discover someone tape recording the discussion. Outraged you confront this person, who proceeds to explain some ambiguous research project, and attempts to justify the act by citing the public context of your discussion. . .after all, this is a public park. Would you not feel that your privacy was grossly violated? Would you not be outraged? Would you demand to know what institution, or agency sponsored and/or approved of the research? Would you be at all swayed by the fair game-public context legitimization? How would an IRB board, or a Human Subjects Committee handle such a proposed research project? Would you react any differently if you hadn't discovered the covert researcher, but found your exact remarks in published text? More importantly, how is this imaginary scenario different from the hypothetical example proposed by King (1995), or the ethical controversy posed by Finn's (1994) research in Computer-Based Self-Help Groups for Sexual Abuse Survivors? The public context of interaction does not preclude the emergence of private interactions. Awareness of this distinction is critical to the maintenance of ethical online research. Similarly, we could compare the nature of this ethical dilemma to the infamous ethical controversy that surrounds Laud Humphreys. The ethical issues raised in Tearoom Trade (1970), are similar and informative to the ethical issues that surround the publicly-private nature of online interaction, and should be instructive to actual and potential online researcher. Humphreys posed as a loitering voyeur and lookout, waiting in public restrooms for opportunities to study brief sexual encounters between anonymous men. In the course the study, Humpherys noted license plate numbers and a year later tracked participants to there home where he posed as a market researcher for the purpose of conducting interviews that provide more information than was possible during covert observation of their homosexual encounters. All events in Humphreys study occurred in a public location (note, one with high access--a public restroom). Yet this study continues to illustrate controversial ethical issues of the violation of participants privacy. Clearly Humpreys methodologies are ethically questionable due to his use of deceit in the course of interviewing these anonymous men. But other unresolved ethical issues remain concerning the ethicality of observing the activities of these men for the purpose of academic research, in spite of the publicality of the location. Synonymous violations of this sort are not acceptable. The key point can be simply stated: The public context of a situation is not a sufficient license to grossly invade the privacy of others. What about this issue is ethically unclear? In sum, researchers must consider the socially defined sense of privacy as interpreted by the individuals participating, as well as the group in which the interaction occurs. The emphasis of this point is a consideration of online interaction as being more than multiple individuals communicating. This is not simply a matter of privacy rights. Appeals to the rights of individuals (or more importantly the lack of) naively and/or conveniently ignore the broader social context of the and the form of online interaction. The form of online interaction is made up of a setting, group, and definitions of the situation. Perceived levels of privacy represent a vital element of the phenomena researchers intend to understand and therefore is an element that ethical researchers must be committed to respect and protect. The Anonymous Form of Online Interaction: Most participants may choose their level of anonymity, allowing for degrees of selectivity in the personal information they report. This is to say that online interaction is an anonymous (or at least quasi-anonymous) form of interaction. The anonymity of online interaction, on one hand protects participants from exposure to other participants in unwanted ways. On the other hand, the anonymity of interaction also allows users the option of presenting a self that is virtually unlimited in form and content. Anonymity (particularly in the faceless environment of electronic space) allows participants to engage others in a self-game where actual social identities become optional elements in the usually non-verifiable presentation of an online virtual social identity (for a discussion of actual and virtual social identities see Goffman 1963). Participants can (and do) choose to modify presentations of self to meetthe demands of the situation. These choices are shaped by the context of the online situation. Thus, the range of possible selves one might present are as limitless as the possible situations in which one presents aself. In sum, selves that emerge online--cyberselves--are not only highly situated (grounded in the fiber of the context), but they present the potential for extreme fluidity. Anonymity is the means by which participants both shield themselves fromunwanted activities of other participants, and flexible and fluid onlineselves emerge. In either case, most users exhibit, value, and will protect at least some degree of anonymity. Since the online world is marked by a certain degree of anonymity, it is easy for researchers to assume that the ethical standards of assuring the anonymity of participants is hard-wired into the context, and thus only requires an alteration of the aliases of participants. The reasoning here is deceptively simple: ifparticipants are already anonymous, only name changes are necessary to protect anonymity. This perspective ignores the embedded importance of the online environment so critical to the form of online interaction. Morespecifically, such a viewpoint (1) confuses the nature of anonymity as characteristic of the individual rather than a situated social condition,(2) confuses online anonymity with the ethical standards of anonymity, and (3) is neglectful and potentially harmful to the participants experience. First, it is important to note that anonymity is a situated social condition. That is, the experience of anonymity is a performance and perception of one's interactions with others, situated in time and space. To experience anonymity is to be selective in the discloser of personal information, and to exert control over ones participation in the location, situations and timing of interaction. This is to say that anonymity is a condition mediated between people in a given situation, location and time. As such, online anonymity is mediated between others in the course of interaction as a variable element of a participant's definition of the situation. Therefore, anonymity will be more-or-less practiced, valued, and protected by users. The emergent and variable nature of anonymity as a mediated quality of online interaction adheres to the observation that there exists numerous degrees of anonymity, suggesting that the anonymity that emerges between people is linked to the participants motivation and nature of interaction. Generally, the degree of anonymity that one is likely to encounter in say, academic or professional oriented virtual communities, will differ from the degree of anonymity between those who gather for the purposes discussing private and personal experiences (issues such as sexual or physical abuse, alcoholism, or depression), and further anonymity is likely to be observed between those that gather for, say, the purpose of mutual sexual arousal or consumption of pornography. Indeed, degrees of anonymity are as varied as the reasons that people engage one another online, and the forums of such interactions. It is important that researchers perceive the anonymity of online interaction not as a given characteristic of the individual, but as a variable and emergent element that exists between people in a given social context. Such a realization highlights the importance of contextual and interactive elements in the construction and maintenance of anonymity, thereby emphasizing the researchers commitment to protect not only the anonymity of the individual, but the context of anonymity itself. Second, the anonymity of online interaction is not the same as the anonymity mandated by ethical standards. Online anonymity represents a series of choices made by participants in the course of interaction. The ethical standard of anonymity is not a choice at all, rather, it is a commitment to the participants. Because some users choose to remain anonymous online is not a valid justification for providing anything short of the highest standards of anonymity that one would expect of any research. Like any other research scenario, all identifying characteristics of person and place should be removed from the research in the course of data collection. This most certainly includes not just names and demographic data, but statements that could potentially reveal the online location of the individual. Lastly, ignorance of the nature of anonymity online is neglectful and potentially damaging to the participants and their experiences of online interaction. To directly identify an anonymous alias has obvious damaging implications, and is clearly irresponsible. Likewise, as King (1995) notes, to directly identify the location of ones research can potentially damage the nature of the interactions that occur. Yet, it should also be stated that researchers must be keenly aware of the importance of anonymity to the user, acknowledging that some users are likely to value and protect there anonymity far more than others. This awareness must recognize that the mere soliciting of participation in a study may lead the user to feel that their anonymity is potentially threatened, thereby potentially damaging the experience of the user. In summary, anonymity is an emergent and variable definition of the online situation. Anonymity is a central element to the form of online interaction, yet it is neither a universal, nor a characteristic of the individual. Because the form of online interaction entails some degree of anonymity does not imply that the researcher is safeguarded from the ethical commitment to providing anonymity to participants and the context of their interactions. The Multiple and Simultaneous Form of Online Interaction: Cyberspace is a literal smorgasbord of potential interaction. Participants belly-up to the buffet of a vast and widely diverse selection of forums and potential others to engage for interaction. Moreover, like a communication orgy, the technology of online interaction allows users the potential to simultaneously converse with numerous others, through a multiplicity of channels. The multiple and simultaneous form of interaction reflects the fluidity of online associations. Like a finicky eater at the social buffet, some participants come and go, with more or less consistency, as they wander from one location to the next. Sampling little bits of conversations, that appeal to the palette of ones taste. Meanwhile, other participants feast on their favorite selection, displaying some degree of commitment to a given online location and/or forum of interaction. In either case, most online associations, and virtual communities are loosely and informally organized at best (in fact, anything more represents a serious organizational challenge). As the metaphors indicate, the multiple and simultaneous revolving doors of access invite participants to browse a variety of locations as they surf from one location to the next. As a result, isolating the participants in any given online location can be difficult, if not all together impossible. Such situation presents serious problems to the actual and potential researcher, particularly as one attempts to obtain some degree of informed consent. The ethical dilemma of obtaining informed consent from online participants is similar to the dilemma of obtaining informed consent in a wide variety of research situated in busy public places. Can, for example, a researcher obtain informed consent from participants in a study on a city bus, a popular bar, or a county fair? Should a researcher bother to obtain such consent? From who should the researcher obtain consent--those who are responsible for the public area, those who participate, orboth? Such questions remain largely unresolved, and essentially rest onthe researchers personal ethical stance in relation to the nature of thestudy and the practical implications of obtaining informed consent. Online research presents a similar dilemma, and any honest debate over the ethics of informed online consent is likely to reveal extreme differences between researchers. Regardless of ones ethical stance, the reality of the multiple and simultaneous form of online interaction often renders obtaining informed consent a practical impossibility. The most ethically cautious of researchers, ones who insist on obtaining informed consent from all participants in any given study, will find this scenario particularly daunting. How, for example, is one to obtain informed consent from individuals in an onlinegroup that is constantly in flux? Who do you approach to obtain informed consent to conduct research on the group? How is one to keep the fluidmembership of the group informed? It is not likely that researchers will agree on these ethical questions, let alone offer logistical solutions to the practical problems they present. As a result, a cautiously variable stance concerning informed consent should be sought. Researchers must carefully consider the totality of the online research scenario, and the decision to obtain informed consent should follow accordingly. Such considerations should include (1) careful thought as to the implications of the nature of informed consent in the specifics of the research context, that is (2) considerate of sensitive persons and groups. First, researchers need to consider the nature of informed consent itself. Informed consent represents a two-part agreement between the researcher and the participant. First, researchers have informed participants as to the true nature, aims and intent of the study, and secondly, participants consent to the researchers use of the information they provide. Informed consent, by its very nature builds from the premise that researchers provide accurate information to the participants of the study. Thus, informed consent cannot be attained in deceptive research, and therefore deceptive research always stands on questionable ethical grounds. The faceless context of online interaction does not allow participants the luxury of observing the non-obtrusive researcher. Participants cannot see the researcher taking field notes, there are no tape-recorders, or any other outward clues that a perceptive participant can use to identify a researcher. The faceless context of online interaction demands that one communicate personal characteristics relevant to the situation. In the context of online interaction, if a researcher does not inform participants of his/her research role, then does this imply deception? In other words, by not stating one's intentions and aims, isn't theresearcher deceiving the participant by playing off his/her ignorance--ignorance of something that the participant could not possibly know anything about? Anything less than informed consent, seems to implies a degreeof deceitful research and thus invokes all the ethical baggage that accompanies it. Secondly, the publicly-private nature of online interaction renders the research scenario sufficiently different from research in public places to warrant careful consideration of the specific applications of the ideals of informed consent--particularly in attempts to study online participants and forums that emerge for deviant, emotionally charged, self-help, and other deeply personal purposes. Such motives for interaction should be perceived as sensitive and fragile. That is, researchers would be wise to assume that some (if not many) of the participants would not engage in such interactions outside the context of cyberspace, and the collectivities that emerge around sensitive topics should be viewed and approached as select-closed groups--regardless of whether this adheres to the reality of access (access to the online group is not synonymous with membership with the group). These groups should be perceived as relatively fragile, and susceptible to damage, even from the simple attempt to obtain informed consent, and unbridled participant observation can be disastrous to participants and their online experiences. In summary, ethical online research most often necessitates some degree of informed consent--anything less would imply deceit. However, the multiple and simultaneous form of online interaction raises serious logistical problems in the effort to obtain informed consent. Furthermore, the publically-private and anonymous elements of online interaction support the development of sensitive virtual groups, to whom the mere solicitation of participation in ones research can be potentially damaging to their experience. We recommend that researchers strive to obtain some degree of informed consent whenever possible. Furthermore, online research--particularly with sensitive groups-- should proceed through key informants. To avoid ethical problems, researchers should obtain a small number of key informants from the online environment of study. Thoroughly debrief key informants, obtaining informed consent. Perhaps with time, the emergence of a good rapport, and a degree of trust other willing participants can be obtained. Faceless, Non-Oral, and Computer-Mediated Form of Online Interaction Online interaction is entirely textual, appearing cold and sterile to outsiders. Furthermore, the computer-mediated medium makes it easy (even for a novice) to access virtual groups, obtain, and save online data. In fact, unlike many other research contexts the recorded data from online studies is frequently identical to the actual experiencestudied. That is, most often one does not have to transform the data tomake it useful for research purposes (for example, you don't have to transcribe online interviews). It is important to note that the textual data that one obtains is not frozen interaction, rather they are theremains of interactions that once transpired. Like a photo, they are limited in dimension and represent only a part of the total picture. Bothreal-time and delayed interaction are processes that cannot be captured in text alone, in spite of its textual base. Each of these issues illustrate how easy it can be to objectify the online interactions that one encounters. Although a certain degree of objectification will occur in any research scenario, in this case the utter absence of flesh-and-blood can potentially result in extreme objectifications that completely lose sight of the subjective meaning that participants associate with their interactions. Such objectifications will always impair a researchers ethical judgment. For this issue, there are no hard-and-fast rules. All that can be recommended is a close association between researcher and participant. Or, in the words of one of our mentors you should get close enough to smellyour participants. Such a rough ideal is helpful online. We must remain close to participants, intimately involved with the research scenario, and if possible a participant in the experience. By doing so, the subjective nature of the interactions, participants involved, and the very process of online interaction may check-and-balance the easily objectified textual remnants of interaction. Lastly, the capabilities of the computer itself raise ethical considerations. Computer technologies make it possible to perform unethical actions from distant locations at increasingly faster rates. Computers allow researchers to collect vast amounts of data, raising questions as to how much data we should collect (all we need, or all we can get?). There can be no doubt that computers make it possible to perform activities that were previously too difficult (or impossible) using manual methods. Each of these issues illustrates how computer technologies alters the scope, range and nature of potential unethical activities. Conclusions The codification of online research ethics has merit. However, such considerations must acknowledge and incorporate elements that uniquely characterize the medium, environment and interactions that define cyberspace. As King (1995) indicates, ethical codes of research must be interpreted and modified to fit the online environment. Unfortunately, we have only begun to gain the knowledge, insights and perspectives necessary to make sound ethical decisions, let alone codify them. In this paper we have identified several ideal-typical elements of the form of online interaction--each raises unique ethical issues. Without doubt the ethicality of online research will continue to be an issue for on-going debate, and we hope that this paper has illustrated one very important point: Ethical considerations for online research should not solely stem from ethical codes of conduct. Rather, ethical considerations should entail an interplay between codes of conduct and an intimate understanding of the nature of the online environment. Required in the efforts to construct guidelines of ethical online research is a renewed ethical commitment to the participants--one in which our concern extends far beyond the experience of the individual, and one that places the interests of participants above and beyond the aims and goals of the researcher. This commitment should always precede codified ethical guidelines, and thus represents where we should begin in the daunting task of defining the ethics of online research. Notes 1. And sociologically, the online environment is more than the sum of it's parts--more than the sum of the communications and computer technologies that comprise it. 2. Note the use of the word participant, as opposed to subject. A subject is a person who is treated or experimented on by others, the word also means to be under domination or control, to be open or exposed. Participant, on the other hand, means to take or have part with others. These two words imply a completely different research relationship--they are not synonymous. 3. In cyberspace any potential self is possible, and anyone can presentthemselves as being anything. It is important to note that in cyberspace there is no such thing as a body, at least not in the sense that we inhabit our body (Penny 1994). That is, we live in our bodies, and our bodies both act on the world, and register the action of the world upon us (Penny 1994). Such an entity is not possible in cyberspace. All that exists are fleeting electronic images associated with a screen name orlog-on ID (other fleeting electronic images). In this sense, cyberselves are disembodied. This alone would suggest that cyberselves are always divergent from selves that emerge in a participant's everyday life--regardless of intentions. That is, some cyberselves are intentionally different from selves that define participants in everyday life, othercyberselves are attempts at a true reflection of who the personis. Regardless, a self is not a fixed entity that one carries around(like an appendice of the body), rather it is a process that emerges in the course of social interaction. Further more, because the online environment does not include a veritable, fixed, physical entity that represents the person (a body), no single meaning can be affixed to the person. Hence, situational elements are particularly central to the emergence, ma intenance, and transformation of the cyberself. Cyberselves are always situated, disembodied social entities of a knife-edge present, and therefore always represent some degree of departure from the selves thatpeople present in everyday life.