Internet users and Internet research

"Internet research" may mean both research on (concerning) the Internet and research with the aid of the Internet (the Internet used as a data source or instrument in some other way, even if the object of the research is not the Internet). Research on the Internet may concern activities such as the publishing or sharing of information, social interactions or conflicts, genre or product development, agitation, etc. As many of the services and fora that are central to Internet users can also be accessed through mobile devices that are partly based on other networks, it will normally be appropriate to adopt a broad approach to what one regards as research on the Internet. The fact that the Internet is still a relatively new phenomenon and undergoing continuous, rapid development also presents special challenges.

Researchers have an obligation to respect the dignity, security and autonomy of human beings. The same rules apply to research on the Internet as to research on other human phenomena: it must avoid harming anyone and observe the requirements of freely given informed consent and of confidentiality where relevant. At the same time, it is not always clear how, or to what extent, ethical experience from other areas can be applied to Internet research.

One of the challenges is how researchers should relate to the information that is made available. This applies in particular to social media, the primary function of which is active self-presentation. The information compiled about an individual across fora and over time can be massive. Relations between different presentations and profiles are often complex. In addition comes the fact that the person the information concerns is by no means always the one publishing it. All these challenges partially overlap ethical questions raised by so-called "Big Data" research – research based on enormous quantities of information that when processed enable correspondences to be found that it would have been impossible or virtually impossible to reveal through more traditional, hypothesis-driven methodology.

Because of the short history and rapid development of the Internet, researchers may be confronted with ethical questions and issues where it is less possible to draw on a wide range of experience and clarified situations than in many more established areas of research. One challenge, for example, is the question of where to draw the line between what is regarded as private and what is to be regarded as public information. In principle, everything one finds on the Internet is publicly available as long as it is not protected by passwords or other measures. Nevertheless, this does not simply mean that it is ethically justifiable to treat all information on the Internet as public. An important indicator for the researcher of how he or she should consider a given Internet forum, is what can be regarded as the users' reasonable expectations. Dimensions that it may be useful to reflect over in order to judge this are the kind of public and participants that use the site, and the kind of information it provides. Similarly, compilations of datasets that may seem innocent enough individually may generate profiles and knowledge that can be regarded as both personal and private.

Internet research is not exempt from the requirement of freely given informed consent if reasonable doubt can be established as to whether the information is to be regarded as public. This principle applies even if finding satisfactory means of obtaining such consent may present the researcher with a challenge. By proceeding heedlessly, researchers risk harming not only individuals and the repute of the research itself, but also social structures and communities. Undesired exposure by a forum may, for example, undermine the trust upon which participation is based, and may thereby destroy the forum itself.

Some of the research ethics challenges with respect to user expectations can be analysed along two intersecting axes:

  1. Identity-building/anonymnity Although this has been a far more typical feature of people's Internet habits in the past than it is now, there are still many kinds of activity where users do not reveal their full identity, for example by operating with pseudonyms. However, this is no guarantee that a piece of research may not lead to identification, since research implies an organised focus on the subject that in most cases it would not otherwise have been exposed to. Conversely, we find typical participation in social media, which are very largely used actively for self-representation. But research that directly or indirectly includes studies of such self-representation may also encounter ethical challenges concerning both the interpretation and the recontextualisation of the original expressions.
  2. Permanence/transitoriness It may appear to some users as though their own activities and those of others on the Internet vanish virtually without trace. In some contexts, users may not have reflected over the long time perspectives that may be associated with publishing or sharing of information. A research project that entails extracting these expressions from their original context, and perhaps storing them separately or publishing them through other channels, may be in conflict with the users' expectations.

Both these axes become relevant in the context of traceability and data aggregation. Our Internet-based actions are to a large extent traced and stored, and research that compiles data from different sources and over long periods of time may provide insight into people's identity that in many ways is broader and more detailed than that possessed by the individuals themselves.

NESH's Ethical Guidelines for Internet Research are currently being circulated for comment, and according to plan will be published by the end of the year 2015.

This article has been translated from Norwegian by Beverley Wahl, Akasie språktjenester AS.