Social-network project shows promise and peril of doing social science online
In 2006, Harvard sociologists struck a mother lode of social-science data, offering a new way to answer big questions about how race and cultural tastes affect relationships.
The source: some 1,700 Facebook profiles, downloaded from an entire class of students at an "anonymous" university, that could reveal how friendships and interests evolve over time.
It was the kind of collection that hundreds of scholars would find interesting. And in 2008, the Harvard team began to realize that potential by publicly releasing part of its archive.
But today the data-sharing venture has collapsed. The Facebook archive is more like plutonium than gold—its contents yanked offline, its future release uncertain, its creators scolded by some scholars for downloading the profiles without students' knowledge and for failing to protect their privacy. Those students have been identified as Harvard College's Class of 2009.
The story of that collapse shines a light on emerging ethical challenges faced by scholars researching social networks and other online environments.
The Harvard sociologists argue that the data pulled from students' Facebook profiles could lead to great scientific benefits, and that substantial efforts have been made to protect the students. Jason Kaufman, the project's principal investigator and a research fellow at Harvard's Berkman Center for Internet & Society, points out that data were redacted to minimize the risk of identification. No student seems to have suffered any harm. Mr. Kaufman accuses his critics of acting like "academic paparazzi."
Adding to the complications, researchers like Mr. Kaufman are being asked to safeguard privacy in an era when grant-making agencies increasingly request that data be shared—as the National Science Foundation did as a condition for backing Harvard's Facebook study.
The Facebook project began to unravel in 2008, when a privacy scholar at the University of Wisconsin at Milwaukee, Michael Zimmer, showed that the "anonymous" data of Mr. Kaufman and his colleagues could be cracked to identify the source as Harvard undergraduates.
The entire story can be read here.