SAN FRANCISCO — The scientific journal that published a study by Facebook and two U.S. universities examining people’s online mood swings regrets how the social experiment was handled.
In a note of contrition, the Proceedings of the National Academy of Sciences concluded that the decision to manipulate the content appearing on the Facebook pages of about 700,000 people without their consent may have violated some principles of academic research.
The journal also pointed out that, as a for-profit company governed by its own terms of service, Facebook had no obligation to adhere to those scientific principles.
“It is nevertheless a matter of concern that the collection of the data by Facebook may have involved practices that were not fully consistent with the principles of obtaining informed consent and allowing participants to opt out,” wrote Inder Verma, the Washington, D.C.-based journal’s editor in chief.
The unusual “editorial expression of concern” surfaced Thursday, a day after Facebook Chief Operating Officer Sheryl Sandberg apologized, acknowledging that the world’s largest social network should have done a better job communicating about the experiment.
Facebook allowed researchers to manipulate the content that appeared in the main section, or “news feed,” of a small fraction of the social network’s nearly 1.3 billion users.
The data scientists conducted the study during one week in January 2012. They were trying to collect evidence to prove their thesis that people’s moods could spread like an “emotional contagion” depending on what they were reading.
Although their findings were published a month ago, the experiment didn’t trigger outrage until the past few days, after blogs and essays in The New York Times and The Atlantic raised red flags about the ethics of treating people like laboratory rats without their permission.
Privacy regulators in the U.K. and France opened inquiries into whether Facebook violated any laws.
Facebook’s data-use policy says the California company can deploy user information for “internal operations, including troubleshooting, data analysis, testing, research and service improvement.”