What is special about the ethical issues in online research?
esearch online no doubt raises some new ethical challenges.1 One question that naturally arises is this: what can we learn from the work that has been done on the ethics of research over the last years? In this paper I will argue that, at one level, the ethical problems of Internet research are structurally similar to problems known from more traditional fields of research within the social sciences and the humanities. Still, I will argue, there are aspects of these ethical issues that are special to Internet research. I will try to characterise these special aspects towards the end of the paper.
Let me start by sketching an example that I have borrowed from the very helpful discussion in Scharf.2 I believe this is a fairly representative example of the kinds of ethical issues that confronts the online researcher. One example obviously cannot illustrate everything, but Scharfs case highlights a number of important issues regarding observation, recording and reporting in online research.
Scharf describes her research project on communication in a discussion list about breast cancer. After some time as a passive member of the list, Scharf decided that she wanted to undertake research on aspects of the interaction in this forum. While she at first was only passively "lurking," her role becomes more and more that of a participant observer. When she posted messages on the list, Scharf explains, she was careful to introduce herself as both a person interested in the topic of the list, as well as a researcher studying the communication on the list. "In retrospect, I believe I had a sense early on that it was prudent to let fellow list members that I had two reasons for participating in this forum. This inclination was especially prompted by both the highly personal nature of much of the content discussed and the knowledge that a majority of people who subscribe to this on-line group are survivors struggling to cope with their disease and carry on with their lives, a network of suffering and courageous people deserving my respect and candor."3
When she a little later began to write a paper based on her observations, Scharf wanted to quote from postings made by members on the list. She decided that for ethical reasons - she had to obtain the consent from the authors of those postings. She found their email addresses, wrote personal notes to each, and waited for their response. Eventually, all her informants gave her permission to quote from their postings. In this process she also received additional information from her informants. "In sum, although the informed-consent procedure I chose to use required extra effort and time, it ensured that explicit permission was given; confidentiality had been satisfactorily been maintained; quoting out of context had not undermined the original authors intent; and my interpretation of others comments had not amounted to erroneous appropriation."4
Even though this was a list that was open to everyone interested in the problem of breast cancer, Scharf did not treat the postings as unproblematically in the "public domain," and therefore freely available for research. She seems to feel that she is "lurking" (Scharfs term) in the period when she is just observing the group, that she should reveal to her research interests when she is posting, and that she is obliged to ask for the consent to use other peoples words. Finally, she makes information about her research paper available to the members of the group.
Scharf mentions one more challenge that a project like this has to face, namely the criticism that research activity like hers does not contribute constructively to the purpose of the list, but is the mere harvesting of the words of others. This kind of criticism has also been directed against researchers of communication in MOOs.5 The point of this criticism seems to be that the research does not itself contribute to the building of the virtual community that the other participants are engaged in, but only uses the efforts of others to serve their own personal purposes. Scharf argues that this criticism can be met if the researcher takes into consideration whether the purpose of the research is compatible with the purpose of the forum; whether the members of the group under study is appropriately informed about the researchers presence, purpose and identity; the need to contact the individuals whose words she wants to quote; the obligation to give some form of feedback to the group; and, importantly, takes into consideration the obligation to "maintain and demonstrate a respectful sensitivity toward the psychological boundaries, purposes, vulnerabilities, and privacy of the individual members of a self-defined virtual community, even though its discourse is publicly accessible."6
I think Scharfs case study illuminates a set of ethical issues that are typical of online research:
In one way, these are just new versions of ethical issues that are well known in traditional research ethics, i.e., the research subjects right to information consent and respect for privacy. At the same time it is clear, however, that the context in which these problems arise gives them a new dimension, so that our intuitions about what is the right way to handle the problems is not always clear. It is important to see the problems from the perspective of traditional research ethics and to ask in what sense they are new challenges. Let me start with the traditional side.
Traditional research ethics: problems, values and the balancing of interest
My point of departure here will be the Norwegian effort to develop ethical guidelines for research in the social sciences and the humanities.7 Let me briefly recapitulate the history of how these guidelines were developed. The guidelines are given by NESH, a central body for research ethics in Norway.8 The first edition of the guidelines was issued in 1994, and a second, revised version was published in 1999. The guidelines were developed through an extensive dialogue with the Norwegian research community. The first draft of the guidelines was formulated after a series of open meetings at various research institutions, and the draft was revised a number of times as the result of comments from the research community and deliberations in the committee. After some years of experience with using the guidelines in practice, it became clear that the guidelines needed to be updated and clarified in several points. NESH therefore initiated a new round of hearings among research institutions in Norway, and NESH got a relatively large number of suggestions for improvements. On the basis of this input, together with suggestions from within the committee, a draft for a revised set of guidelines was formulated. This was distributed among those who had made suggestions in the first round, and a number of further suggestions came in. After some further deliberations, the committee published the revised set of guidelines in February 1999.
This history is of little general interest in itself. But my point is that, because of this long process of dialogue with the research community, the guidelines goes a long way I feel - towards summing up the ethical standard among Norwegian researchers in the social sciences and the humanities. It is not just a rough set of principles formulated by a committee, in isolation from the field it is supposed to guide. This is what makes the guidelines interesting as an account of what the issues of "traditional" research ethics are within the social sciences and the humanities.
The structure of the NESH-guidelines
The need for research ethics arises when there are conflicts between the central interests of research i.e., the interest in new and better knowledge and other of interests that are external to the research process. Examples are the conflict between research and respect for the integrity and autonomy of research subjects (e.g. issues pertaining to information, consent, privacy), the conflict between the integrity of research and economic interests (in contract research, for example), and the conflict between the individual researchers interest in getting publications and intellectual honesty (i.e., the problem of plagiarism and fraud in different forms).
The guidelines have six chapters. Five of these are concerned with the different types of conflict that can arise in the research process, and try to indicate some of the morally important concerns that should be taken into consideration in different types of situations. The ambition is not of course - to give algorithms for solving ethical problems that might arise in the research process, but, rather, to help identify the issues that are involved in different situations. It is recognized that each situation is unique, and that it is the people who are involved in a particular situation who have to figure out how the problem should be handled. The guidelines are intended as just that a guide for those who are already motivated to think ethically.9
The guidelines identify two main categories of conflict of interests that can arise in the research process. One type is where the integrity of the research is at stake, when the freedom of the research process is under pressure from economic interest, for example. In such cases, the guidelines are concerned with how the integrity of the research process can be protected. The other main type of issue arises when there is a conflict between the interests of the research and the integrity of the research subject. In such cases, the main question is how the integrity of the research subjects can be protected.
The first kind of conflict is characteristic of the problems treated in parts 3-6 of the guidelines. The chapter called Groups and Social Institutions deals, on the one hand, with the need for protection of the interests of non-individual entities like institutions and social groups. It is emphasized, for example, that "[r]esearchers have a special responsibility for safeguarding the interests of under-privileged groups throughout the research process." (§ 22). But this section of the guidelines is also concerned with a different type of problem, namely the protection of the integrity of the research process itself. For example, it is said that "[r]esearchers must not allow themselves to become inordinately dependent on informants. Research into social problems can reveal reprehensible or unlawful circumstances such as plans to commit violent acts or failure to care for children -- exposing the researcher to conflicting loyalties, especially in view of the demand for confidentiality." (§ 23) The duty to bring the results of the research back to the subjects of research is also emphasized: "[e]very effort should be made to report the results of research into other countries and cultures back to the countries, cultures and persons who made the research possible." (§ 41)
The next chapter treats ethical issues pertaining to behaviour within the research community. For example, problems of plagiarism and theft of ideas are addressed: "Researchers using or quoting the ideas, publications or research materials of others must declare their sources." (§ 27) The fifth chapter deals with contract research, i.e., ethical problems that might arise when the research is under pressure from economic interests. It is stated, for example, that "[r]esearch results must not be influenced by client interests. The researcher has the right and duty to point out the limited validity of results, for instance in relation to planned political decisions." (§ 37) The final chapter, on Imparting Research through the Media, deals with the protection of the integrity of the research when encountering the mass media. The main message is that the researches have to take into account that the media often have an agenda that is different from that of the research project. It is emphasized, therefore, that "[a] researcher is obliged when imparting research results to seek the greatest possible clarity of language in order to prevent misunderstanding and misinterpretation." (§ 42)
It is clear that issues such as those we have considered here, pertaining to the question of how to protect the integrity of the research, can arise also in the context of online research. I will not discuss this class of problems further, however. Instead, I will focus on the first major type of research ethical problems identified by the guidelines, i.e., the cases where the integrity and autonomy of the research subjects are at stake.
Concerning the protection of persons
On the questions discussed above, the guidelines speak of various concerns that the researcher ought to take into consideration in the research process. But when it comes to problems where the integrity of persons is at stake, however, the guidelines talk about obligations and rights. "The guidelines distinguish between two types of norm: those which are of the nature of ethical demands and rights that must be observed, and those that express important considerations which decision-makers should take into account in their decisions." (§ 4) Norms of the first kind are to be found exclusively in the second part of the guidelines, called Concerning the Protection of Persons.
The fundamental norm in this regard is the obligation to respect dignity of persons (§5). On the "theory" of the guidelines, this norm is the basis for deriving other strong obligations, in particular: the obligation to protect to freedom and self-determination (§6), the obligation to prevent injury and harm (§7), the obligation to obtain informed consent (§8), the obligation to inform research subjects (§9), the right of children to protection (§11), the obligation to respect individuals privacy and close relations (§12), and obligations pertaining to confidentiality (§§13, 14). In all of these areas, there are deontological constraints on the behaviour of researchers, according to these guidelines. For example, with respect to the obligation to inform research subjects, the guidelines state that:
Persons participating in research must be given all the information they need for a reasonable understanding of the research field, of the consequences of participating in the research project, and of the object of the research. They must also be told who is paying for the research. (§9)
This is the default norm. There are situations, however, where it this obligation to inform is less strict:
Observation in public spaces, in streets and squares, can normally be carried out without informing those concerned. However, the registration of behaviour using technical aids (camera, video, tape recorders etc.) implies that the observed material can be stored, and thus possibly form the basis of a personal register. For the purposes of such registration people must, as a rule, be informed that recordings are being made. (§9)
This illustrates an important ambition on the part of the guidelines, namely to point out that there often are exceptions to the prima facie obligations and that often more fine-grained judgements are called for.
Another example of the same thinking is the obligation to obtain consent. Here the guidelines first state that:
Research projects which presuppose active participation must as a general rule only be initiated with the freely obtained and informed consent of the participants. The same rule applies to research that entails some risk of strain for the participants. Informants are free at any time to discontinue their participation, with no negative consequences for themselves. (§ 8)
Again, the default rule appears to be quite strict. However, situations can arise where the obligation to obtain consent is not so easy to fulfil. In many cases this will mean that the research cannot be undertaken. But not always. The guidelines mentions two cases:
Research on groups who lack the ability or the opportunity to give free and informed consent. In some cases, their interests can be safeguarded by obtaining the consent of those responsible for them.
Research in which those on whom the research is being done must be kept unaware of the problems posed and the details in the hypothesis. If the informants can not be fully informed in advance, they must be informed afterwards. Informants are entitled to help with problems that may arise in consequence of their participation in the project. They should subsequently be asked for their permission for the continued use of the collected material in the research. (§ 8)
In the first case, the problem is that some people have a reduced competence to give an informed consent. In some cases but not all - research can still be done. The other case is when the subjects cannot be told about the purpose of the research, because this would undermine the point of the project. This is the case with some studies in psycholinguistics, for example, where the point is to collect specimens of natural language. The problem is that if subjects are told in advance that their way of talking is recorded for the purpose of research, they tend to become self-conscious of their speech and will frequently try to modify the way they talk. It is better, therefore, not to tell them the real purpose of the research until after the recording has been done. This can be all acceptable, according to the guidelines, because they can be fully informed afterwards and can then consent to the use of the recording.
But the guidelines mentions also another kind of situation where the obligation to obtain informed consent is not so easy to fulfil, i.e., situations where it can be acceptable to undertake research without consent.
The freedom and self-determination of participants can in some cases be respected even when consent is not obtained in advance. Exceptions to the demand for free and informed consent can be considered in cases where the research entails little or no risk, and is necessary to obtain insight which demonstrably enhances the welfare of individuals. A typical case might involve passive participation in research based on archive materials, such as health registers. (§ 8)
The reasoning behind these exceptions to the obligation to obtain informed consent is that in the cases mentioned, there is little risk of harm or reduced autonomy. In one sense, then, the spirit of the guidelines is rule-consequentialist.10 First, a set of fundamental values is identified: respect for persons, freedom and self-determination and the interest in avoiding harm. In many cases the functions of the norms pertaining to consent and information are to express respect, to protect autonomy and to avoid harm. But in some cases, they do not have such a function. The guidelines suggest, then, that in these cases the norms of consent and information can be lifted.
This brings up an important question: to what extent is there a trade-off between the value of the research and other interests. On the one hand, we have seen that the guidelines are rather strict about the obligation to respect the integrity of persons. On the other hand, a large part of the first chapter of the guidelines is devoted to elaborating the importance of research. Hence, the guidelines certainly emphasize that it is an ethical obligation on the part of both the individual researcher and the research community to promote good research. According to the guidelines, it is precisely in the cases where there are no important values on the part of the subject at stake, where there is room for a trade-off between the value of the research and the interest.
It is important to note, however, that there is no tendency in the guidelines to say that if only the research is important enough, the respect for persons can be sacrificed. As we have seen above, the guidelines strongly emphasize the research subjects rights to respect for integrity and autonomy. Furthermore, to open up for an unrestricted trade-off between the interests of individuals and the value of research, would be clearly unacceptable. To see this, consider the case of the famous Norwegian researcher Armauer Hansen, the man who discovered the leper-germ. As part of his very important research, he implanted a needle infected with leper into the eye of an uninformed and unwilling patient. As a consequence of this, he was in 1880 brought to court and convicted, and the right to treat patients was taken away from him. Our strong intuition is, I think, not only that what he did was wrong, but also that the fact that his research had such good consequences his discoveries probably saved many lives - does not in any way compensate for the wrong he did to his patient. The history of research ethics is full of examples like this. But not in any of these cases are we willing, I think, to make a trade-off between the integrity of individuals and the value of the research in question.
It is clear, then, that the guidelines do not promote anything like such a full-blown trade-off. But still, there are these attempts, mentioned above, to open up for a balancing of interests and to bring in the value of the research into the ethical evaluation. Let me try to characterize what room the guidelines leaves for the balancing of interests in the ethical evaluation of research, by way of the following decision-rule:
Is there only minimal risk of harm? Yes / No / Exit
Are the integrity and the autonomy of research subjects adequately secured? Yes / No / Exit
Is the method adequate? Yes / No / Exit
Is the knowledge produced relevant enough? Yes / No / Exit
This decision rule is meant to express several points. The first point is that each question has to be answered affirmatively before you move on to the next. Furthermore, the questions have to be answered with a clear yes or no. These features are meant to capture the deontological constraints on the treatment of research subjects. Consider, first, the requirement that there should be only a minimal risk of harm. The idea is that if the research is possibly very harmful, you should not even ask for peoples consent to participate. Research in the social sciences is usually not that dangerous, but even such research can be harmful. The guidelines discuss the risks in this way:
People may be injured by undergoing serious mental and social strain or by incurring financial losses in consequence of the research process, or because of inadequate safekeeping of research materials. Research in the humanities or social sciences entails less dramatic risks for those taking part than where medical research is concerned, because of the lower risk of physical injury. On the other hand, the potential damage is more difficult to define and measure, and long-term effects are difficult to assess.
Examples of serious strain include stress beyond everyday levels, anxiety, reduced self-respect, reactions to invasions of one's private life, conflicts or guilt arising from being induced to act in violation of social norms, the feeling of being placed in embarrassing situations, physical discomfort, negative reactions to being told afterwards of the true purpose of research with concealed objects, impaired social relationships, or financial loss on the part of participants. (§7)
But if the risk of harm is acceptably low, one has to consider the next level of questions:
Informed consent freely given makes it possible to carry out research that involves a degree of risk of strain. (§8)
As we have seen above, the question whether the subjects integrity and autonomy are adequately protected is a complex one. And, again, heavy deontological constraints are put on the permissible treatment of research subjects. If also this second question is answered affirmatively, only then are the third and fourth questions brought into consideration. The view of the guidelines is, then, that the value of the research has a role to play in the ethical evaluation of research, but that this role is heavily constrained by other obligations.
The NESH guidelines and online research
The NESH guidelines are relevant to online research at different levels. First of all, it is clear that the obligations to respect individual integrity and autonomy, which the guidelines identify as fundamental, are equally relevant when considering how to treat subjects in online research. The same holds true for the obligations to obtain informed consent, to inform subjects, respect their privacy and close relations, etc. Furthermore, the constraints on the balancing of interests, illustrated by the decision-rule above, are also applicable to research online. Hence, with respect to the fundamental research ethical considerations, I see no reason to think that that online research raises issues that are different from other kinds of research within the social sciences and the humanities.
To be more concrete, let us look briefly at the example from Scharf, mentioned in the introduction, in the light of the NESH guidelines. As mentioned above, we should not expect the guidelines to give an algorithm for solving our ethical problems, but, rather, to point out to ethically relevant aspects of types of situations. The distinction between observation and recording was mentioned above (§9, on the obligation to inform subjects), which was an important issue also in Scharfs case. Another question that was important in Scharfs case was whether it was necessary to obtain informed consent to cite from some of the postings. One reason for doing this would be that people should be given the possibility to withdraw from the research, as mentioned by the guidelines (§8, on the obligation to obtain consent). Third, Scharf was very concerned about reporting back to her informants. Again, the obligation to do report back is treated in the guidelines §10:
Researchers are under a special obligation to report the results of their research back to those who participated in it, in a proper and comprehensible form.
Research must not entail the exploitation of informants by the researcher. Informants and respondents give something of themselves to the researcher and have a reasonable claim to something in return. Informants should, whenever possible, have an opportunity to correct misunderstandings. Information on the results should in any case be made available to whoever has contributed.
The case also raises issues pertaining to confidentiality, which is treated in the guidelines §13.
The point here is not to contribute to the discussion of Scharfs case, but only to emphasize the point that much of this thinking on the ethics of "traditional" research ethics in the social sciences and the humanities, as manifested in the NESH guidelines, for example, is directly relevant to the ethical considerations of online research. Additional evidence for this is the fact that there is a substantial amount of overlap between the set of ethical rules Scharf herself formulates on the basis of her case study, and those that we find in the NESH guidelines.
What is special about the ethical problems of research online?
But even if the ethical problems of research online in the ways indicated are similar to other problems of research ethics, they obviously have features that are unique. As indicated by the examples discussed in the introduction, this is related to the special features of online communication. Researching such forms of communication raises special problems about how to protect the flow of personal information.
To further characterise these special features of the ethical issues of online research, let me start by way of sketching an account of informational privacy.11 My first observation is that various social arenas, e.g., a public street, a private home, or a hospital register, can all be seen as channels for the flow of personal information. But such channels differ in the ways the participants informational privacy is protected. They can be characterised by the restrictions on the flow of personal information that applies to them. There are, I think, two fundamental types of restrictions; restrictions on the audience that has access to the channel, and restrictions on how sensitive information that is allowed to flow there. Let me illustrate the idea by way of the following table:
The flow of personal information in a public street is characterised by the fact that there are no restrictions on the audience that access to it. The street or any other public place is open to everyone. That is what makes it public. Our private homes, on the other hand, are characterised by the fact that there are very strict limitations on the audience. That is an important part of what makes our homes private spheres. But our private spheres are also characterised by the fact that there are few restrictions on the flow of sensitive information in this channel. Indeed, it is because there are heavy restrictions on the audience, that this channel is suited for the flow of sensitive and personal information. In a public street, in contrast, there are social norms against the dissemination of sensitive information, which are strongly related to the fact that there are no restrictions on the audience. There is, then, an inverse relationship between sensitivity and accessibility: more sensitive information that is allowed to flow in a channel, the less open to access usually is. And the more accessible a channel is, the less suited it is for the flow of sensitive information.
This relationship between accessibility and sensitivity reflects the standard way to protect personal information, i.e., to put restrictions on the audience that have access to the information. This is illustrated by the third case in the table above, the patient register in the hospital. This is, obviously, a channel where sensitive information flows. And such information is protected by restricting the audience that has access to the channel. But here there are, in addition, restrictions on the type of sensitive information that is flowing in the channel. Only information that is relevant to the treatment of the patients is supposed to flow here. Both the restrictions on the audience and the restrictions on the type of information are dictated by the purpose of the channel, i.e., medical treatment.
Now, we move in and out of different channels every day. But in order to do it, we have to know what kinds of restrictions apply to the channel, and we have to trust that these restrictions are not violated. We give up personal information in order to receive medical treatment, for example, trusting that the information is not disseminated to a wider audience. It is here that the protection of informational privacy comes in. Legal norms for the protection of privacy try to secure that personal information is not disseminated in ways that individuals who give up the information could not have reason to expect. The European Directive, for example, states that information shall only be collected for a specific purpose, that it shall not be reused in ways that are incompatible with the original purpose, and that information preferably shall be collected on the basis of an informed consent by the research subject.12 I mention only a few of the principles of this regulation, in order to illustrate the point that norms for the protection of privacy can be seen as a protection of the individuals ability to form reasonable expectations about how information about him or her will be treated in different channels. A violation of an individuals privacy occurs typically when personal information is disseminated or used in a way that is in conflict with the restrictions of the flow of personal information that the individual had reason to expect. In such a situation the basis on which the subject entered the channel is undermined.
Many of the norms of research ethics can be seen as motivated in the same way. The reason why research without the research subjects consent is often problematic, is that such research in many but not all cases will be incompatible with the assumptions about the restrictions on the channel, on the basis of which the subject entered it. The research will necessarily constitute the dissemination of personal information from a channel to a wider audience. This is not unproblematic, if such dissemination undermines the individuals control over the flow of information about themselves. In such cases, it would be unethical not to obtain the subjects consent to the intrusion into the channel. Hence, the function of norms of research ethics is to help secure that peoples expectations about the processing of information about themselves are not undermined.
The complexity of informational privacy in online communication
In the discussion above, I suggested a very simple trade-off between the sensitivity of information and the accessibility of the channel in which this information is flowing. On the picture above, more sensitivity means more restricted access, and vice versa. I will argue, however, that with online communication, this relationship is not so simple.
It is a characteristic feature of many forms of online communication, I think, that quite sensitive information is allowed to flow in these channels, despite their being accessible to a much wider audience than normally has access to private channels. There are plenty of examples of this with communication in MOOs, chatrooms, discussion- and newsgroups, and bulletin boards. In these fora, the sensitive and personal information is not protected by restricting the audience that has access to the channel, but by several other means. First there is the anonymity offered by the use of nicknames. Here, holding back the identifying information, i.e., the information that links the information to me as a real world subject, reduces the sensitivity of the information. Second, the sensitive information is protected by the ephemeral nature of the information. In many fora, in particular in chatrooms and in MOOs, the interaction may last for a very short time, and it may not necessarily be part of a larger pattern of interaction. Again, this is an important form of protection of sensitive information, a form that compensates for the relative accessibility of the channel.
But there is also a third way that sensitive information is protected in such fora, which is related to the fact that they are the basis for various forms of community. Remember the example from Scharf on the discussion group for victims of breast cancer. For this to function as a support group, there has to be a climate of mutual respect and trust among the participants on the list. This also involves norms of confidentiality and respect for privacy. In a case like this, Scharf argues that the researcher has to "maintain and demonstrate a respectful sensitivity toward the psychological boundaries, purposes, vulnerabilities, and privacy of the individual members of a self-defined virtual community, even though its discourse is publicly accessible."13 Indeed, the climate of trust in a self-defined community like this can be threatened if information is disseminated without the consent of those concerned. Not only can this be harmful to the people whose words are quoted, but it can also potentially be a threat to the social dynamics that made the group an interesting object of study in the first place. People who take part in a community like this ought to be able to form reasonable expectations to the effect that information is not leaked into a different context without their consent.
This is related to another significant feature of Internet communication, namely the fragmentation of audiences. It is noted by many commentators how the Internet can be used and is used - to design our own informational environment, so that the information we get is filtered in such a way that we only get the information in which we are interested.14 With more individual control over what information one can access, and less need for general intermediaries, there is more fragmentation of audiences. Online communication is part of this picture. In many electronic fora, there is a restricted audience, the majority of which has a common interest in the forum. Even if there are no formal or physical restrictions on the audience, it will in fact be reasonable for participants to expect a much more limited and homogenous audience. Hence, it is not reasonable to treat communication in such a forum as public in the same way we would treat communication in a major newspaper or, say, a major television channel.
The point here is to suggest that there are special kinds of restrictions on online channels of communication, restrictions less typically found with offline channels. In the offline world, there is often an inverse relation between the sensitivity of the information and accessibility to the channel, i.e., restrictions on the audience that has access to the channel is the traditional way to protect sensitive information. In the online world, there are often not any formal or physical restrictions on access to channels of communication. Instead, the sensitivity of the information is protected by the ephemeral nature of the communication, pseudonymity, or people rely on norms of confidentiality in the community in which they take part. These types of restrictions on the flow of information will be parts of the conditions that form the basis for the individuals communicative activity in the online channel.
I argued above that a violation of privacy occurs when the individuals reasonable expectations are frustrated concerning which restrictions apply to the processing of personal information. In the same way, we should ask what reasonable expectations people have formed concerning the flow of personal information, when they are communicating in online contexts. This way of framing the question sheds light on the important differences between observation of online communication, the recording of the communication, and citing people when reporting studies of such communication. To passively observe in an open forum must be unproblematic, as long as there are no restrictions on the audience that can have access to the forum. The recording of the communication in the forum without consent is problematic, however, if the nature of the communication is ephemeral. That is because you then add a new feature to the processing of the information in the channel. Hence you violate the reasonable expectations of the people under study. And the quoting of research when reporting from the research is even more problematic, because that in fact opens up a new channel for the flow of information into a different context altogether. That will be problematic to do without the permission of the people quoted, because they had good reason to believe that their statements were made in the context of a certain community where certain norms and values applied. This is more of a problem the more sensitive the information is.
One obvious consequence to this is that the distinction between the private and the public becomes much more complex in electronic fora than it usually is. For this reason, it is not helpful when the NESH guidelines, for example, states that "Observation in public spaces, in streets and squares, can normally be carried out without informing those concerned." (§7) With electronic fora, one cannot rely on a simple public / private distinction when considering the ethics of research. And, in fact, the NESH is aware of the problems with the private/public distinction in electronic fora:
The distinction between the private and the public domain may be difficult to draw with regard to information concerning behaviour which is imparted and stored electronically, for instance on the Internet. When using material drawn from such interaction, researchers must give necessary consideration to the fact that people's perceptions of what is private and what is public communication in such media may vary.
While this is on the right track, more work clearly needs to be done to chart the moral complexities of online research.
Dag Elgesem. The structure of rights in Directive 95/46/EC on the protection of individuals with regard to the processing of personal data and their free movement of such data. Ethics and Information Technology, 1 (3): 283-293, 1999.
Chris Mann and Fiona Stewart. Online Communication and Qualitative Research: A Handbook for Researching Online. SAGE, London, 2000.
T.M. Scanlon. Rights, Goals and Fairness. In Samuel Scheffler, editor, Consequentialism and its critics. Oxford University Press, Oxford, 1989.
Barbara Scharf. Beyond Netiquette: The Ethics of Doing Naturalistic Discourse Research on the Internet. In Steve Jones, editor, Doing Internet Research. Sage, London, 1999, pp. 243-256.
Andrew Shapiro. The Control Revolution. Public Affairs, New York, 1998.
Cass Sunstein. Republic.com. Princeton University Press, Princeton, 2001.
Michele White. Regulating research: The problem of theorizing community on LambdaMOO. Ethics and Information Technology, 4 (1): 55-70, 2002.
1 For an excellent overview, see Chapter 3 "An ethical framework", in Chris Mann and Fiona Stewart. Online Communication and Qualitative Research: A Handbook for Researching Online. SAGE, London, 2000. BACK
8 In Norway there are three national committees for research ethics, one for medicine, one for science and technology and one for the social sciences and humanities. They have mostly advisory functions, and do not review individual cases on a regular basis. The author was director of NESH from 1994-1999. BACK
9 In Norway there are no general, ethical review of projects within the social sciences and the humanities. There is, however, a review of research projects by the data commissioner with regard to questions of data protection. BACK
11 I have developed this account in more detail: Dag Elgesem. The structure of rights in Directive 95/46/EC on the protection of individuals with regard to the processing of personal data and their free movement of such data. Ethics and Information Technology, vol. 1 (3): 283-293, 1999. BACK
12 Elgesem, 1999, Article 6. (Available at http://www.privacy.org/pi/intl_orgs/ec/final_EU_Data_Protection.html) BACK
Ethical Issues of Online Communication Research
Representations or People?
Ethics of Internet Research: Contesting the Human Subjects Research Model