Facebook apparently has been manipulating the feeds that some users get to see in order to measure how it the content affects the tone of their own posts.
You can read about the research here; I’m not publicizing it, because the Facebook’s research is an abuse of users and their trust. I don’t mind them reading my posts, for they own the service, and the service is in their name. I assume they will use my data and content to make money, but I didn’t agree to allow them to manipulate me, or what I write, feel, or think. I’m also not especially optimistic about the uses the results of such research might be applied to.
The researchers claim that the research is ethical because a computer program scanned for words that were considered either “positive” or “negative,” but the Facebook content wasn’t actually read. Facebook terms of service state that user data may be used “for internal operations, including troubleshooting, data analysis, testing, research and service improvement.”
Since Facebook users agree to the terms of service, the researchers argue that this constitutes “informed consent” for their experiment.
To be “informed consent,” users—like me—must actually understand all of the reasonable implications of “research.” I assumed that “research” meant that the data itself would be analyzed. I did not give conformed consent for researchers to use me as subject for psychological experiments. Nor, I suspect, did anyone else. This may be legal. The service agreement may be enough to avoid any legal challenge. It is not, however, ethical research.
There are many codes of ethics for human research, which this undeniably was. They do not vary significantly regarding what “informed consent” means in this context. Here is the relevant section from one such code:
“Informed consent” should normally be obtained. Information should be provided as to the nature and purpose of the research project. Reasonable steps should be taken to ensure that the research participant is aware of the nature of the research and their involvement in it and that they have freely agreed to participate in the research. Researchers should address in their application for ethics approval both how informed consent can be achieved and demonstrated. Generally good practice would suggest that these objectives should be achieved through the provision of an information sheet and the signing of a consent form copies of which must be provided to the research participant. However it should be noted that in some circumstances, for example where written consent may pose a risk to the research subject, oral consent may be sufficient. In addition it should not be presumed that the fact that a form has been signed means that the individual has given consent to inclusion in the research.Participants have a right to withdraw consent to participation in research. They may choose to withdraw their participation at any time.
Does this sound consistent with “secretly manipulate the determination of what posts from friends Facebook users see to measure its effect on their own posts without their knowledge?” No, because what Facebook did had no informed consent by the subject of the research consistent with research ethics.
In seeking informed consent, fine print and implied consent to a study based on nothing related to the study methodology does not suffice. Any researcher who says otherwise is lying, or shockingly ignorant of his or her own profession.
UPDATE ( 7/3 ) : As I said…
Facts: Think Progress