Ethical issues are complex in any setting, and are particularly complex in online interview research. Ethical issues are central to any review by the committee or for academic or scientific merit, as well as to review by institutional review boards.
The chapter (4) on this topic is the biggest in the book-- and themes related to ethical practice are woven throughout other chapters as well.
I'd like to highlight a few inter-related issues in this discussion:
Fair Information Processing, Confidentiality and Informed Consent.
Fair Information Processing refers to protection of data, spelling out that data should be collected for the purpose noted in the agreement, participants should have access to the data and the right to correct inaccuracies. Confidentiality must be honored; personal data are not to be communicated externally without explicit consent.
These issues lead to the third: informed consent. Perhaps the most important ethical principle in research involves actions by researchers to ensure that participation of subjects is voluntarily and will cause no harm.
Researchers seeking informed consent need to make clear to their subjects how material about them and/or from them will be used; the specific uses of material and how their identities will be protected are part of information subjects need to understand before signing an agreement.
Any interviewee must sign a consent agreement, regardless of how casual the interaction may be. There is no ambiguity where that is concerned. However, for other associated research activities many shades of gray are present. So, if you want to complement, prepare for or follow-up your interview with some observations, the question of consent is less clear.
Data collection in “public” face-to-face settings is typically exempt from informed consent when individuals are not identified. Observing people on the street or at a community event could be clearly described as data collection in public. The determination of public versus private space online is not clear cut or universally defined.
After reviewing the literature, I developed a continuum I hope will be useful for people who are trying to determine when trying to determine whether the interaction is public or private.
In other words, according to this model, if you are in a forum (like this one!) and chat with someone with the intention of recording/saving and using the response as data, then it is an "interview" and consent is needed. If you are observing general trends in posting and contribution to a discussion in a setting that is open to the public (such as this one) and do not quote any individual posts, then it could be argued that it is equivalent to observation of a public discussion in a local community center.