Three Letters About the Facebook Study


My colleague Leslie Meltzer Henry and I have sent letters asking three institutions—the Proceedings of the National Academy of Sciences, the federal Office for Human Research Protections, and the Federal Trade Commission—to investigate the Facebook emotional manipulation study. We wrote three letters, rather than one, because responsibility for the study was diffused across PNAS, Cornell, and Facebook, and it is important that each of them be held accountable for its role in the research. The letters overlap, but each has a different focus.

  • Our letter to PNAS deals with the journal’s commitment to publish articles on human subjects research only when participants gave informed consent and an IRB reviewed the substance of the research. We explain why emotional manipulation study met neither of those conditions, and why the only appropriate response by PNAS is to retract the article.
  • Our letter to OHRP deals with the Cornell IRB’s flawed reasoning in treating the emotional manipulation study as research conducted independently by Facebook. We unpack the conflicting statements given to justify the study, and show that none of them stands up to close scrutiny.
  • Our letter to the FTC deals with the heightened concerns that arise when consumers are subject to active manipulation and not just passive surveillance. We explain why conducting psychological experiments on consumers without informed consent or oversight can be a deceptive and unfair trade practice.

Our letters deal with cleaning up the mistakes of the past. But they also look to the future. The Facebook emotional manipulation study offers an opportunity to put corporate human subjects research on a firmer ethical footing, one in which individuals given meaningful informed consent and in which there is meaningful oversight. We invite PNAS, OHRP, and the FTC to take leading roles in establishing appropriate ethical rules for research in an age of big data and constant experiments.

UPDATE, July 17, 2014, 1:30 PM: I am reliably informed that Cornell has “unchecked the box”; its most recent Federalwide Work Agreement now commits to apply the Common Rule only to federally funded research, not to all research undertaken at Cornell. (I made the mistake of relying on the version of its FWA that the Cornell IRB posted on its own website; I regret the error.) This affects the issue of the OHRP’s jurisdiction, but not the soundness of the Cornell IRB’s reasoning, which rested on the activities of Cornell affiliates rather than on the source of funding.

UPDATE, July 24, 2014, 2:00 PM: The letter to the FTC overstates the effects of the Bakshy et al. link-removal study when it describes the study as making some links “effectively unshareable.” Links were removed from News Feeds on a per-user basis, so removed links were still seen by other users.