Using the Internet for Research on Human Subjects
Written by James P. Dillard, Communication Arts and Sciences

The internet is a virtual treasure trove of data. Because of the enormous amount of behavioral information that it makes available, it has become a prime location for individuals seeking to study the dynamics of human interaction. Beyond the sheer quantity of information, the internet has several other desirable properties. For instance, researchers can collect data from widely dispersed populations. This can often be accomplished at relatively low cost and in less time than similar research in the physical world. Often the data – human interactions – come in a form ready for analysis with no transcription needed. As a result, the number of research studies conducted via the internet has exploded in recent years. And, as new functions are developed and old ones are improved, it is likely that even more research will be done via the internet. Anytime that research procedures change, some reflection on how to apply general principles for the protection of human subjects to new circumstances is called for. Consideration of how the internet is used can help to resolve questions concerning the use of human subjects.

The simplest case is that of the online survey. By their very nature, surveys are intrusive. Prior to data collection, individuals must be contacted and their participation requested. So, in most respects, an online survey is no different than a paper and pencil version. Researchers need to devise an informed consent procedure, have it reviewed by the IRB, then maintain records just as they would with a physical questionnaire. Informed consent is obtained by asking respondents to read the form and indicate via a button click their acceptance of the study protocol. This despite the fact that it is impossible to verify that the person doing the clicking is the owner of the account. Researchers rely on individuals to maintain the security of their own accounts and identities. This appears to be a realistic expectation.

Some years ago, individuals began to post to the internet diaries of their personal experiences. These diaries became known as blogs, a term that is an abbreviated version of weblog. Regarding the expectation of privacy, blogs present an easy answer: Zero. There is no expectation of privacy whatsoever. Blogs exist for the purpose of having others read them. Posting a blog is the rough equivalent of reading your diary on the evening news except that many more people will have a chance to read the blog on the internet than would ever be exposed to it on the news. Thus, data obtained from a blog is wholly public behavior that can be observed, analyzed, and attributed without the need to seek consent from the blogger.

Chat rooms are sites where individuals gather to discuss issues of common interest. They differ from discussion boards or online forums insofar as chat room conversations take place in real time whereas discussion board interactions are asynchronous. Issues of timing aside, topics in any of these venues are tremendously diverse, ranging from food and gardening to cancer survivorship and divorce. All of these venues usually have rules that they require users to follow in order to maintain civility. When boards/rooms are monitored, the person in charge has the right to delete messages. But, many boards/rooms are intentionally unmonitored. The basic premise is that anyone at anytime can post anything and, by design, no one will interfere. The internet is an inherently public medium. Data generated in circumstances such as chat rooms and discussion boards are properly viewed as analogous to shouting in the town square. Internet users know this. And, they go to these sites precisely because they are public. In the final analysis, there is no expectation of privacy and no basis for preventing researchers from having unfettered access to those data and the insights that they can provide into human behavior.

ORP Response/Cautionary Statement (AKA "Second Thoughts, Doubts, Demons We Live With, On the Other Hand..., Who Knew?, What If...?, Never Thought of It in That Way")
Written by Tracie Kahler, ORP

We agree that the Internet, particularly chat rooms and blogs, have opened up all kinds of new doors for researchers. Speaking from the IRB review side of things, though, we must admit that we have nightmares about this type of research. From a regulatory perspective, all is not as black and white as one would like regarding chat rooms and blogs. Welcome to the world of "shades of gray"! In actuality, this topic is not only greatly debated in the human participant research protections world, but it is also greatly debated in the offices of 201 Kern Graduate Building (i.e., The Office for Research Protections).

A good place to start our discussion might be the definition of a human participant. According to 45 CFR §46.102[f], a human participant is defined as "a living individual about whom an investigator (whether professional or student) conducting research obtains: 1) data through intervention or interaction with the individual OR 2) identifiable private information". The key term here is "private". Federal regulations base the definition of "private information" on a participant's "reasonable expectation of privacy".

There are those on the side of the fence that say all chat rooms and all blogs are not private because anyone with an Internet connection can access this information. Therefore, communications within these contexts do not meet the definition of "human participant research" and no human participant research protection oversight is needed. Then there are those on the opposite side of the fence that say in many situations (e.g. chat rooms), participants expect privacy and do not expect their activity to be studied by researchers. Therefore, this could be considered a reasonable expectation of privacy and should have human participant research protection oversight.

A typical day at the ORP might involve any one of us jumping from one side of the fence to the other side of the fence, and then back again. This is not due to indecisiveness, but rather it is due to the myriad of questions that arise from our discussions. Par for the course, these questions beget more questions that beget even more questions. Some of the considerations we take into account might be:

1. Is this indeed a "public" forum or is there some type of registration required in order to participate in this chat room and/or blog?
2. Does the chat room and/or blog have policies regarding privacy, confidentiality, copyright, etc.?
3. If we do consider data from the chat room and/or blog to be private (for whatever reason) and the investigator chooses not to reveal himself/herself to the chat room and/or blog participants (e.g., because the investigator wishes to not bias the results/discussions), do we determine this to be a "deception" study? What if the investigator "participates" in the chat room and does not reveal herself/himself to be a "researcher"? What do we call this?
4. If we do consider data from the chat room and/or blog to be private and the investigator does not have a problem with announcing his/her presence, in what manner is this accomplished and how do we proceed with review and approval?
5. What if we determine that a chat room and/or blog is private and yet there are no personal identifiers other than screen names/pseudonyms? Are these screen names/pseudonyms truly "not identifiable"? Is it not possible in many cases to track down one's actual identity with the screen name/pseudonym?
6. If we decide that a particular chat room and/or blog is not private, can we proceed with a "non-human participant research determination"?

These questions are just the tip of the proverbial iceberg. One can see that the determination of "privacy" is much less black and white than it would appear at first glance. Thus, as simple as it may seem, much more goes on behind the scenes as the ORP and the IRBs review and approve research proposals using Internet chats or blogs.

References:

1. The Institutional Review Board by Bankert and Amdur (2005)
2. Personal experiences by ORP and/or IRB members