International outrage at Facebook’s study on thousands of its users without their consent has raised questions about the ethics of research done by private companies.
Facebook’s Emotional Manipulation Study has led to concerns about how such research is carried out since it was published in the Proceedings of the US National Academy of Sciences.
Facebook and one of the study authors, Facebook’s Adam Kramer, have this week responded to the concerns. But now the PNAS journal has also published an “Expression of Concern”.
PNAS editor-in-chief Inder M Verma acknowledges that “questions have been raised about the principles of informed consent and opportunity to opt out” and that this is concerning.
Done by the book
But Verma argues that the paper was consistent with Facebook’s policies at the time, and with the regulatory framework within the United States, known as the Common Rule.
This rule requires oversight by a research ethics committee and adherence with common practices regarding informed consent only if a study receives federal funding or is associated with an institution receiving federal funding.
In regards to informed consent in the Facebook study, which the Expression of Concern acknowledges is problematic, there are now larger problems.
While the paper presents a façade of informed consent by agreement to Facebook’s Data Use Policy, we have now found out – thanks to Forbes – that the clauses regarding research were only added to this policy four months after the experiment took place.
So rather than mere consent, as I termed it in my earlier piece, we would be better to describe this study as having no consent at all.
Ethics approval or not
But is the journal correct to think that the experiment – carried out by researchers from both Facebook and Cornell University – didn’t need research ethics committee approval?
The editor-in-chief’s argument is that because this experiment was conducted by Facebook for internal purposes, the university’s Institutional Review Board determined that the project “didn’t fall under Cornell’s Human Research Protection Program”.
[… ] as a private company Facebook was under no obligation to conform to the provisions of the Common Rule when it collected the data used by the authors.
This is reinforced by the statement from Cornell, which says that its researchers – Professor Jeffrey Hancock and then doctoral student Jamie Guillory – did not participate in data collection and did not have access to user data.
Their work was limited to initial discussions, analysing the research results and working with colleagues from Facebook to prepare the peer-reviewed paper […]
Because the research was conducted independently by Facebook and Professor Hancock had access only to results – and not to any individual, identifiable data at any time – Cornell University’s Institutional Review Board concluded that he was not directly engaged in human research […]
This is a matter of interpretation – when is someone involved in human research?
In this case the academics were involved in the study design and the analysis afterwards. The only thing they didn’t do was collect the data themselves.
Does that exclude their institution from making sure that they behave ethically? Or does it really just constitute a run around any ethics regulation?
International concern from regulators
A further regulatory question is whether this research needed to be approved in other jurisdictions such as the UK, Ireland and Australia.
This depends on where the 689,003 Facebook users whose news feeds were manipulated without their knowledge actually live. Regulators in some countries are now pursuing this.
The more interesting question is should the standards of ethical conduct of research and regulation not apply simply because Facebook is a private company?
This seems a difficult position to defend, particularly in the face of studies like this one, which may have exposed a significant number of people to risks (however minimal) without consent.
It is worth keeping in mind given that 6.7% of Americans suffer from clinical depression. So approximately 46,000 of Facebook’s research base may have been suffering depression at the time of the study, done over a single week in 2012.
Even a relatively minor alteration in their emotional state prompted by their Facebook news feed manipulation could have had a significant impact on them.
The history of research ethics has shown us, unfortunately, that if we want research participants to be treated with respect – and research to be conducted in line with ethical principles – then we ought to regulate research with review by an independent research ethics committee.
This latest case suggests that the US ought not make a significant regulatory distinction between private and publicly funded research.