INTRODUCTION
The rapid growth of the Internet has been accompanied by a growth in the number and types of virtual environments supporting computer-mediated communication. This was soon followed by interest in using these virtual environments for research purposes: the recruitment of research participants, the conduct of research, and the study of virtual environments. Early research using virtual environments raised a number of ethical issues and debates. As early as 1996 a forum in the The Information Society (volume 12, issue 2) was devoted to ethical issues in conducting social science research online. The debate has continued with more recent collaborative attempts to develop guidelines for ethical research online (Ess & AoIR ethics working committee, 2002; Frankel & Siang, 1999). In this article we explore contemporary ethical issues associated with conducting research online.
BACKGROUND
The basic principles of ethical research with humans are integrity, respect, beneficence, and justice (National Health & Medical Research Council, 2006). Based on these principles many professional associations provide ethical guidelines, or codes, for the conduct of research. Guidelines and legislation vary across disciplines and across countries. However, these codes have typically been developed for use in off-line settings, prior to consideration of research being conducted online1. While these codes contain guiding principles for research generally, the translation of these principles into actions for conducting research in virtual environments is open to interpretation. The process of translating ethical guidelines into ethical practice online involves a deliberation of the options available to the researcher and the likely impact on research participants, their communities, and the research process. Central concerns in this process are maintaining respect for individuals, their online identities, and the ownership of words.
PUBLIC Vs. PRIVATE SPACE
Research online can take place within a range of virtual environments that vary in terms of purpose, synchronicity, access, number of users, and norms. A major issue in developing ethical research procedures for use within a particular virtual environment is determining whether the setting represents a private or public “space.” Various attempts have been made to distinguish between the public and the private in virtual environments (see, e.g., Lessig, 1995), but little agreement has been reached. There are currently no clear guidelines for researchers on what constitutes private vs. public space in virtual environments, yet the distinction is important, as it affects the rights of participants to be advised of the research and to give or withhold their informed consent.
The defining of public vs. private space cannot be reduced to the single dimension of accessibility to the virtual environment. Interactions that occur within publicly accessible virtual environments may be perceived by participants to be private. Newsgroups can be accessed without restriction, yet newsgroup postings can be, and frequently are, high in self-disclosure and are perceived by many users to be private (Witmer, 1997). Similarly, support groups on sensitive issues may be conducted in publicly accessible sites with participants adhering to norms of confidentiality and privacy (Elgesem, 2002).
Some ethical codes exempt naturalistic observations and archival research from requiring informed consent where no harm or distress is likely to come to those researched and where their confidentiality is protected. It has been argued that the decision not to inform members of online groups about research conducted on the group has the advantage of the research being ” unobtrusive” (Langer & Beckman, 2005). Others, while acknowledging the benefits of naturalistic observation, regard this approach as placing researchers in a position “little better than spies” (Bakard-jieva & Feenberg, 2001, p. 234). King (1996) highlighted the potential for psychological harm to members of online groups where research is conducted and published without the prior knowledge and informed consent of participants.
Where there has been the expectation of privacy within a group (however misinformed that expectation may be) the individual may feel violated upon hearing of, or reading, the results of that research.
Where the presumption is made that online communication occurs in public space simply because it is accessible without restriction, an anomaly may result in how research participants are treated in equivalent settings in online and off-line research. For example, research on support groups off-line requires the informed consent of research participants, while similar research online may occur without the knowledge or informed consent of the participants, on the grounds that all postings are public documents (see, e.g., Salem, Bogat, & Reid’s 1997 study of a depression support group). Despite the inequities this raises, in a recent review of psychological research conducted online (Skitka & Sargis, 2006) it was noted that “most Institutional Review Boards are concluding that online postings represent the public domain and that researchers do not need to obtain informed consent to use this material” (p. 549).
Table 1 summarizes possible dimensions against which the public/private nature of a virtual environment can be assessed. Virtual environments where all dimensions fall on the left-hand side of the continua may be deemed as public environments for research purposes and subject to guidelines for research in public settings. We recommend virtual environments where all dimensions are on the right be deemed private environments, requiring informed consent from research participants. The difficulty arises with the majority of settings that do not fall clearly into public or private spaces. Researchers do not have the right to define virtual environments as public or private to meet their own research needs (Waskul & Douglass, 1996). Rather, account should be taken of the size and nature of the online forum and the intrusiveness of the study. Consideration should be made of the likely effect of the request to conduct research and the research itself on research participants and their communities. The process of requesting consent to research may in itself alter group dynamics (Sixsmith & Murray, 2001).
Informed consent
Research conducted in virtual environments that have been conceptualized as private settings requires the informed consent of research participants. Obtaining informed consent in virtual environments is more problematic than in off-line research, as participants are frequently geographically dispersed. In addition, research participants may be reluctant to divulge details of off-line identities required for the signing of consent forms. Further, it is difficult to verify factors that may affect an individual’s ability to provide informed consent, such as age, mental competency, and comprehension of risk (Skitka & Sargis, 2006).
A range of options has been suggested for obtaining informed consent in online research (Bruckman, 1997; Flicker, Haans, & Skinner, 2004; Jacobson, 1999; Kralik, Warren, Price, Koch, & Pignone, 2005; Roberts, Smith, & Pollock, 2004; Smith & Leigh, 1997), and these have been summarized in Table 2. Selection of a method for obtaining informed consent will necessarily be dependent upon the type of virtual environment, the level of anonymity required by research participants, and their access to high-level computing facilities. Regardless of the method used, the information about the research should be presented in a format that the research participants can keep and refer back to at any time before, during, or after their research participation.
Table 1. Dimensions of public and private space in virtual environments
| Accessibility: | Accessible to all | * | Restricted membership |
| Users’ perceptions: | Public | * | Private |
| Community statement: | Research permitted | * | Research prohibited |
| Topic sensitivity: | Low | * | High |
| Permanency of records: | Public archives | * | Private logs only |
Table 2. Methods of obtaining informed consent in online research
| Format of information | How consent obtained | |
| Signed consent: | Hard copy or electronic | Post, fax, or e-mail |
| Implied consent: | Electronic | Gateway WWW page |
| E-mail consent | ||
| Logging of consent | ||
| Use of password protected site |
Table 3. Factors that decrease the anonymity afforded by pseudonyms
• Use of name, derivation of name, or nickname
• Use of same pseudonym across virtual environments with differing requirements for identification
• Self-disclosure
• Active seeking of identifying information by others
Table 4. Levels of anonymity (site, pseudonym, and quotations)
• Identify site, use online pseudonym, and directly quote
• Identify site, use pseudonym of online pseudonym, and directly quote
• Identify site, use pseudonym of online pseudonym, and paraphrase
• Do not identify site, use online pseudonym, and directly quote
• Do not identify site, use pseudonym of online pseudonym, and directly quote
• Do not identify site, use pseudonym of online pseudonym, and paraphrase
Skitka and Sargis (2006) reviewed Internet-based psychological research conducted over a two-year period. In almost two thirds (62.5%) of studies reviewed, the researchers clearly obtained informed consent from research participants; 12.5% did not obtain informed consent, and it was unclear in the remaining 25% whether informed consent was obtained or not. Of concern, deception was used in 18% of studies, with less than half indicating that debriefing had been provided to research participants.
Consideration should also be given to seeking the cooperation of community gatekeepers2 and advising the community of the research being undertaken. Advising communities of a research project requires the public identification of the researcher. In some circumstances, the decision to research within a particular virtual environment may be made after the researcher has been either an active participant or “lurker” within that environment. We recommend that researchers make their researcher status overt as soon as the research process begins. This may include identifying as a researcher in pseudonyms (Roberts et al., 2004), descriptions (Allen, 1996), or objects (Reid, 1996); linking between research and social identities (Roberts et al., 2004); and posting information about the research.
Advising communities of a research project may take ongoing effort in public virtual environments without membership boundaries. Identifying oneself as a researcher once within an online group does not mean that absent or future members of the group are also informed of the researcher’s role (Sixsmith & Murray, 2001). There may be a need to re-identify researcher status and restate and clarify the role of the researcher on an ongoing basis.
PROTECTING ANONYMITY AND CONFIDENTIALITY
Individuals typically adopt a pseudonym (or pseudonyms) for use in virtual environments, providing a level of anonymity. Some studies include pseudonyms in reports and publications on the basis that they provide information relevant to the study (Langer & Beckman, 2005). While it has been argued that research involving pseudonymous characters is exempt from regulations governing human subjects as “true” (off-line) identities are not known (Jacobson, 1999), there are often links between the individual and their pseudonyms that decrease the level of anonymity a pseudonym provides (Allen, 1996; Bruckman, 2002; Jacobson, 1996). These are presented in Table 3. The combination of these factors means that researchers cannot assume that pseudonyms provide adequate protection for off-line identities. The degree of anonymity conferred in virtual environments does not reduce the ethical requirements for researchers to protect the anonymity of research participants and virtual interaction settings (Waskul & Douglass, 1996).
Protecting the anonymity3 of the individual extends to protecting the anonymity of their pseudonym(s), as representations of the individual online. Pseudonyms themselves gain reputations over time (Bruckman, 2002). Researchers can provide varying levels of protection to research participants’ anonymity (see Table 4). The practice of replacing existing pseudonyms with other pseudonyms in research materials confers little additional protection to the existing pseudonym when text searches can be used to identify source documents (Allen, 1996) or where the individual has a distinctive, recognisable writing style (Markham, 2005).
Further, other community members may seek to identify disguised identities and may share this information with others (Bruckman, 2002).
In addition to protecting the anonymity of research participants in research reports, the data collected need to be kept secure in order to protect confidentiality. Maintaining the security of data collected in computer-mediated research poses unique difficulties. Confidentiality may be breached at the site of data collection, during transmission of data, or in the storage of data. Sites at which data is collected may not be secure and may be subject to surveillance by gatekeepers, “hackers” (Rhodes, Bowie, & Hergenrather, 2003), other computer users, or others physically present in the location at the time (Kralik et al., 2005). Confidentiality of data may be breached during data transmission where another party intercepts data (Nosek, Banaji, & Greenwald, 2002). This may include the Internet service provider of the research participant or researcher. Employers may also monitor employees’ e-mail (Sipior & Ward, 1995; Weisband & Reinig, 1995). Confidentiality of data may be breached during storage by hackers, employers, or as a result of “open records” legislation (Pittenger, 2003).
Online researchers need to provide the most secure forms of data collection, transmission, and storage possible, aiming to minimize the risks of unauthorized persons gaining access to research data at any stage of the research process. The procedures used to ensure this will differ according to the type of virtual media used. Using a “perspective of reasonableness,” Kralik et al. (2005) argue that the chances of a breach of security are small and need to be weighed against possible gains from the research.
ownership of words
The ownership of electronic messages has been contested. It is still unclear whether the individual who authored a message, the community to which it was sent, or anyone who has access to the message is the owner of the electronic message. Electronic postings may be considered original works protected by copyright, although this has not yet been legally tested (Sixsmith & Murray, 2001). If informed consent is not obtained to use electronic messages, copyright provisions suggest that they are subject to “fair dealing” for research purposes, and should be attributed to the author (Australian Copyright Council, 2001). Researchers who neither obtain informed consent, nor reference the material they are quoting, risk violating both ethical and copyright standards. With the consent of the research participant, quotes in research may be attributed to the individual, their online pseudonym, or used anonymously. Respect for research participants is demonstrated through asking, and abiding by, their preferences for anonymity, pseudonymity, or identification. However, this is not without its problems. Individual preferences for identification may not be consistent with the norms or wishes of the community. There can be tensions between respecting copyright entitlements of individuals and protecting the privacy of other participants within a virtual environment. Roberts et al. (2004) highlighted the potential for negative impacts on the privacy of other virtual environment members when one research participant’s work is fully attributed, including information on the virtual environment.
Returning research findings to the community
A requirement in some ethical codes is to provide research participants with information about the outcome of the research. In addition to being a requirement, this can demonstrate respect for the individuals who participated in the research. A summary of research findings can be provided to research participants in hardcopy or electronic format. Where research participants are reluctant to provide contact information that may link their online and off-line identities, the summary can be placed on a Web site or sent through the messaging system of the virtual community (Roberts et al., 2004).
future trends
Rapidly changing technologies will result in the development of an increasing range of virtual environments and tools that may be used for research purposes. The precise characteristics of these new virtual environments and tools may vary greatly from those available today. Concern has already been raised over the potential for participant harm when non-human agents used in the research process evoke negative responses or unwanted arousal (Palomares & Flanagin, 2005). Privacy issues have been raised in relation to Web-cams, which have the potential to capture data from both consenting research participants and other non-consenting individuals in their proximity (Palomares & Flanagin, 2005).
Decisions made regarding research methodologies and designs in virtual environments need to be embedded in the basic principles of ethical research and may require multi-layered considerations. Before conducting research within each new type of environment or with a new type of research tool, researchers will need to address the intrusiveness of the proposed research, the perceived privacy of the research setting, the vulnerability of the community, the potential for harm to individuals and/or the community, and how confidentiality will be maintained and intellectual property rights respected in their research proposals (Eysenbach & Till, 2001). This requires a consideration of the likely impacts of the research on both research participants and the communities in which the research is conducted. It should be guided by researchers’ knowledge and adherence to the “Netiquette” and social norms of the virtual environments concerned. Guided by the principles outlined in their ethical codes, researchers will need to develop ethically defensible strategies that balance the needs of research participants and their online communities, offering protection to both, while providing the validity assurances required by the readers/users of the research.
conclusion
Our approach to the conduct of ethical research in virtual environments is based on a human research perspective, explicitly recognizing that communication online is conducted by individuals who interact via their online identities. Our focus is therefore on individuals rather than texts. We privilege the rights of individuals to make informed consent about whether or not traces of their interaction online (e.g., logs or postings) can be used for research purposes. We believe this approach is consistent with general ethical guidelines for human research in the social sciences. Alternative perspectives to the conduct of ethical research in virtual environments place a stronger emphasis on the cultural production of texts and performance (Bassett & O’Riordan, 2002; White, 2002) reflecting calls for an ethical pluralism in Internet research that recognizes a range of ethical perspectives as legitimate (Ess, 2002).
Regardless of the perspective adopted, all research should comply with the principles of ethical research as outlined in the relevant professional association’s code of ethics or by Institutional Review Boards. In the absence of specific guidelines for online research and where review committees are unfamiliar with online research issues (Keller & Lee, 2003), we recommend that researchers be guided by the principles outlined in their code, adapting the guidelines for use in virtual environments as necessary. In addition to protecting the rights of research participations, researchers have a social responsibility to ensure online research is conducted in an ethically defensible manner to maintain the viability of virtual environments as research media (Colvin & Lanigan, 2005).
KEY TERMs
Computer-Mediated Communication: Communication between two or more individuals that occurs via computer networks. Computer-mediated communication may be text, audio, graphics, or video based and occur synchronously (in “real time”) or asynchronously (delayed).
Informed Consent: An individual’s freely given consent to participate in research based on information provided by the researcher(s) about the research, possible risks associated with the research, and the voluntary nature of participation.
Informed consent must be obtained without coercion or undue influence.
Netiquette: The etiquette, or social rules, associated with communicating online. Netiquette may vary across virtual environments.
Private Space: Off-line, private space refers to geographical areas that are not for general or public use (e.g., your home). Online, the term private space is commonly used to refer to virtual environments, or parts of virtual environments that have restrictions on who may access them.
Pseudonym: The fictitious name adopted for use within a virtual environment. An individual may consistently use the same pseudonym or adopt several pseudonyms for use within and between virtual environments.
Public Space: Off-line, public space refers to geographical areas that are accessible to the general public (e.g., streets). Online, the term public space is commonly used to refer to virtual environments that do not have restrictions on access.
Virtual Identity: Representation of the individual in a virtual environment. The form of representation varies across virtual environments and may range from a pseudonym only (Internet relay chat), a pseudonym combined with a character description (multi-user dimensions), through to graphical representations, or avatars, in graphics-based environments.
An individual may have multiple virtual identities.
