Cornell’s Involvement in Controversial Facebook Emotions Study Raises Ethical Dilemmas
The Cornell Chronicle released an article this summer regarding a study about “emotional contagion” via Facebook. The article released on June 10 described a research project conducted by Prof. Jeffrey Hancock and doctoral student Jamie Guillory (now attending University of California, San Francisco) of Cornell University and Facebook researchers that investigated the communicable effects of emotional manipulation on social media sites, particularly Facebook.
In the experiment, 689,003 of Facebook’s 1.3 billion users were presented with news feeds generated by a modified content selection algorithm. The modified algorithm reduced the amount of positive or negative content of a user’s news feed. The user’s following status updates were then analyzed to determine whether the change in news feed emotion correlated with a subsequent change in expressed user emotion.
But the study published June 2 entitled “Experimental evidence of massive-scale emotional contagion through social networks” by the Proceedings of the National Academy of Science-Social Science sparked quite a bit of controversy. For one, there was initial confusion about Cornell University’s Institutional Review Board (IRB) approval of the study. Secondly, many Facebook users were surprised to find out that they had (or had not) agreed to participate in such studies when they signed up for Facebook. The seeming lack of “informed consent” gives rise to potential issues of mental and emotional health along with a concern for user privacy. Of course, there was also the aspect of potential government involvement in the study as well.
Initially, Princeton University Prof. Susan Fiske, who edited the study, expressed concern on whether the study was approved by Cornell’s IRB. In an article by The Atlantic, she noted that the Cornell researchers explained that the study was approved since “Facebook apparently manipulates people’s News Feeds all the time.” However, it was later clarified that Cornell IRB had approved the use of a “pre-existing dataset,” which implied that the data and the methods used to obtain it were already approved by another organization, presumably Facebook. The more significant implication of this statement is that it reveals the data had already been collected prior to Cornell IRB’s review.
A media statement released by the Cornell Media Relations Office confirmed that the IRB was consulted after the study had already taken place. However, the IRB defended the researchers by adding that no review of the study was required in the first place because the Cornell researchers involved had access only to the research results.
The trouble with Cornell IRB’s approval is that it unquestioningly accepted Facebook’s approval process. While the Cornell researchers did not have access to confidential user data in any way, the ethical validity of the study was left up to Facebook. As an independent review board, IRB would have been wiser to analyze the study independently and conduct the appropriate actions necessary if ethical issues were found.
The American Psychological Association (APA) requires informed consent, or consent using “language that is reasonably understandable to that person or persons,” of research participants. While the current Facebook terms of service marginally seems to obtain informed consent of its users, the terms of service at the time of the study—before the policy update in May 2012—does not according to The Atlantic.
The APA further requires the notification of all study participants “no later than at the conclusion of data collection” for the “deceptive research” that the study conducted. Under the same guidelines, it also requires the study to “permit participants to withdraw their data.” No actions that suggest compliance with these guidelines have been reported by the study.
Some other issues raised about the study have been addressed to varying degrees. User privacy seems to raise much less concern because the study reports that no confidential user data was seen by researchers; user data was analyzed by software that reported only the results of the findings.
However, a great ethical concern lies in the fact that this study altered the mental health of people, unbeknownst to them. While most find no difficulty in coping with slightly more negative articles, the possibility remains that those that are mentally unstable could have been more drastically affected by the study. The extent of the study’s health effects are uncertain and are likely to remain so due to the difficulty in measuring such complex effects.
An initial press release by Cornell University claimed the research was funded in part by outside funding sources, including the U.S. Government. It was later corrected in an update from Cornell’s Media Relations Office that “While Prof Hancock, like many researchers, has conducted work funded by the federal government during his career, at no time did Professor Hancock or his postdoctoral associate Jamie Guillory request or receive outside funding to support their work on this PNAS paper,” according to an article from The Guardian.
Cornell University is one of several universities nationwide that is involved with the Department of Defense’s Minerva Research Initiative, which funds universities in research that improves “DoD’s basic understanding of the social, cultural, behavioral, and political forces that shape regions of the world of strategic importance to the U.S.” according to the Minerva Research Initiative’s website. Prof. Hancock has conducted research funded by the federal government, including the Minerva Research Initiative, in the past as with many other professors at Cornell and from other universities. However, the University insists that no federal funding was requested or granted for this study.
Prof. Hancock was unavailable for commenting.
Mark LaPointe is a junior in the College of Engineering. He can be reached at mnl38@cornell.edu