Is Using Facebook as Data Source Ethical?

A team of researchers from Harvard and UCLA are monitoring the activity of an entire East Coast college’s Facebook activity, reports the IHT today. Scholars are examining the ways in which people connect with one another. But, partly to keep the data clean, the Facebook users do not know they are being watched. Is this kind of study ethical?

One problem is that Facebook is not necessarily representative of the population at large. The IHT article notes that: "Eszter Hargittai, a professor at Northwestern, found in a study that Hispanic students were significantly less likely to use Facebook, and much more likely to use MySpace. White, Asian and Asian-American students, the study found, were much more likely to use Facebook and significantly less likely to use MySpace."

In my mind, this sampling issue came up the other day during Garrett Graff's talk here at the Carnegie Council on his new book “The First Campaign” (listen to his talk here). Graff was talking about how the 2008 U.S. presidential elections will be the first campaign in which technology as both the medium and the message will determine the outcome. But doesn’t that also beg the obvious question:

Won’t those who are more technology-savvy therefore have relatively more influence in the political arena?

A sample bias may also occur in that people who are engaged in politics online might be peculiar in a certain way. Do these people have more time on their hands than do others? It reminds me of a polling data problem: Polls draw from surveys conducted over the phone… What kind of person answers the phone call from a stranger? Related, the people answering the phone surveys are probably the mirror of the techo-savvy since the surveys can only legally be done on land-line phones—something many of the savvy netizens gave up altogether for mobile services.

Another problem with the Facebook study is that the subjects have not given the researchers permission to study them. Feelings are mixed on this issue reports the IHT:

Most researchers acknowledge these limits, yet they are still eager to plumb the site's vast amount of data. The site's users have mixed feelings about being put under the microscope. Katherine Kimmel, 22, a graduate student at the University of Cincinnati, said she found it "fascinating that professors are using something that started solely as a fun social networking tool for entertainment," and she suggested yet another study: how people fill out Facebook's "relationship status" box. "You're not really dating until you put it on Facebook," she said.

But Derrick Clifton, 19, a student at Pomona College in California, said, "I don't feel like academic research has a place on a Web site like Facebook." He added that if it was going to happen, professors should ask students' permission.

Clearly the data would be skewed if the subjects knew they were being studied. If the identify of the particular subjects is withheld in the research findings, it would seem OK. But that is very difficult to do, as researchers discovered last year after AOL made available the search queries of 650,000 users. From the New York Times:

Although the 650,000 AOL users were not personally identified in the data, the logs contained enough information to discern an individual’s identity in some cases.

AOL quickly withdrew the data from its research Web site, but not before it had been downloaded, reposted and made searchable at a number of Web sites. And on Monday, the company dismissed Abdur Chowdhury, the researcher who posted the data, along with another employee. Maureen Govern, AOL’s chief technology officer, resigned.