Facebook Says It Will Help Academics Use Its Data. Here’s How That’s Supposed to Work.

There has never been a time when so much data existed about human behavior. What many of us buy, sell, like, dislike, read, and tell our friends is recorded on the internet thanks to sites like Facebook. To social scientists, the company is sitting on a gold mine.

Some of that information is public, but much is not, and the company’s reach is so vast most people don’t know how far it extends. Several research projects that use Facebook data have ended as high-profile privacy-breach scandals in part because subjects didn’t know they were being studied. In the most recent and possibly the largest data breach at the company, an academic harvested information about millions of Facebook users and shared it with Cambridge Analytica, a firm that advised the Trump campaign.

One might think that in the wake of that scandal, Facebook would lock academics out. That’s what Gary King, a political scientist at Harvard University who has pitched Facebook about opening up its data for research, expected. He met with Facebook officials right before the Cambridge Analytica news broke and, to his surprise, he got a call a few days later. They wanted him to study the company’s impact on elections.

King knew that Facebook wouldn’t actually give him unfettered access to all its information. So in the last few weeks, he and Nathaniel Persily, a law professor at Stanford University, designed what they are calling a new model for industry-academic partnerships that will allow them to explore a question: What impact has social media had on democracy?

Their model attempts to give academics access to the vast troves of data the social-media company has collected, but without the preconditions companies tend to insist on to protect their proprietary secrets and the privacy of their users. The model was designed with an eye toward Facebook and the question of its impact on democracy, but it is also meant to be applied to other companies and research questions.

“You need somebody who is trusted by the company not to reveal its secrets and private data,” Persily said. “You need the scholarly enterprise be independent enough so it’s not seen as pressured by the company.”

This model is meant to accomplish both those goals. Other scholars offered differing opinions on whether they believe it will.

The idea is to let a group of scholars into Facebook to see all the data that’s available and determine what research questions can be answered. (King said most of the data will be anonymized, though there could be situations where a scholar would need to see identifying details. Everything would be protected, he said.)

The group, called a commission, would be trained on the company’s policies and would theoretically know why some information can be shared and some can’t, presumably because of legal proceedings or certain privacy laws. The scholars on the commission would not be able to do their own research or share the company’s secrets, but they would solicit proposals from others and serve as a filter for the data, meting it out to the researchers as deemed appropriate.

“Now almost all the data in the world is created in private companies that we don’t have access to,” King said. “The question is, Is there some way to unlock this data to do good?”

No Money From Facebook

Facebook will have input on who serves on the commission and will be able to scuttle research projects that “violate the company’s legal obligations, interfere with ongoing or imminent litigation, violate privacy, or compromise proprietary information,” King and Persily wrote in a paper they published on Monday. The commission will ensure that the research that is published is valuable. It will have the responsibility to report to the public if Facebook violates the agreement.

The Social Science Research Council will oversee a peer-review process. At this point, no money is coming from Facebook, though that’s not a precondition for other partnerships with other companies. The project being funded by a group of foundations: the Laura and John Arnold Foundation, the Democracy Fund, the William and Flora Hewlett Foundation, the John S. and James L. Knight Foundation, the Charles Koch Foundation, the Omidyar Network, and the Alfred P. Sloan Foundation.

A lot of questions remain unanswered, such as who will serve on the commission and how big it will it be. Since others in academe won’t know every detail about the data, a lot rests on the credibility and perceived independence of the commission. The commissioners can’t be seen as part of an effort to atone for Facebook’s past missteps. King said. That’s not the intent.

“It is undoubtedly going to be the case that some of these researchers are going to write something that Facebook is going to have heartburn over,” he said. “My own view is that’s good. It should happen sometimes.”

Samuel Gosling, a psychology professor at the University of Texas at Austin who has written about social scientists’ use of Facebook for research, said he likes the model. The dataset available at Facebook is so vast and so rich that he thinks it’s imperative to find a way to make it useable.

“In these sorts of debates what tends to get emphasized are the costs and dangers of the research,” Gosling said. “And what tends to get forgotten are the costs and dangers of not doing the research that might be helpful and valuable to society.”

William H. Dutton, a media and information-policy professor at Michigan State University, saw some issues with the model proposed by King and Persily. For one, he had reservations about using a dataset that he did not know everything about from start to finish. He was also skeptical of the idea that the commission would come up with the broad research questions in this model.

“Centralizing the development of a research agenda is never a good idea,” Dutton said.

The research question proposed — how does social media impact democracy? — is a very interesting and important one, he said, but he added that the question would be better answered by looking at all social media, not just Facebook.

“A lot of academics don’t want to be Facebook researchers, they want to be social-media researchers,” Dutton said.

There is a lot of interest in this project, however, and Dutton predicted that the commission would get many of research proposals.

“These are really distinguished people, and they may have all of this thought through,” he said. “I wish them well.”

Nell Gluckman writes about faculty issues and other topics in higher education. You can follow her on Twitter @nellgluckman, or email her at nell.gluckman@chronicle.com.