A Must-Read For BT Wonks

The academic studies of online privacy and the technologies behind data collection are coming on strong of late. I have been covering them here in recent months because as a former egghead (I say that lovingly) in the media studies field, I have seen this movie before. As legislators and regulators turn their eye towards a commercial endeavor, scholarly research and opinion that ordinarily gets shared among a relative handful of experts suddenly becomes widely cited, misquoted, turned into gospel.

Regulators need some semblance of impartial authority when diving into complex technical matters, and, frankly, academics need funding, which usually comes from government sources. The upshot is that people in the digital media industries may as well spend some time now hitting the scholarly books, because they will meet them in the legislative and FTC hearing rooms many of them will be attending in the coming years.

This week, researchers at the University of California, Berkeley issued the most cogent and comprehensive survey both of Web user tracking and privacy policies I have seen. Readers should run, not walk to to see the details and download the full report.



The UC Berkeley team licensed data from Ghostery, a Web bug detection plug-in for Firefox that scans HTML code for trackers. The Ghostery survey of 50 of the top Web destinations showed just how many Web bugs are planted by top destinations like Blogspot (100 bugs), Typepad (75), and Google (44). They also identified that among the top 50 sites, Google Analytics tracking was present 81% of the time, followed by Doubleclick (60%), Microsoft Atlas (60%) and Omniture (57%). A detailed breakdown of policies, Web bugs, and data collection types on all of the sites surveyed are available in the Site Profiles section. The research even includes a survey of press coverage of online privacy to determine which issues are being over or under-reported and the levels of awareness they create for consumers.

One of the most interesting pieces of new insight the research offers involves user complaints about privacy. Through Freedom of Information Act requests of the Federal Trade Commission and other data sets from other organizations, researchers surveyed not only what and who consumers complained about, but also the process of complaining about privacy. Complaints to the FTC focused mainly on people search services like Zabasearch and Intelius, because public display of personal information was the second most common concern among those filing complaints to the FTC. But the "lack of user control" over one's own data was the most common gripe.

The report makes an interesting distinction between data brokers like ZabaSearch and behavioral targeting networks. Both entities have no personal relationship with the consumer. They both collect information about users from multiple sources and do so invisibly. The next argument is worth quoting at length: "The striking difference between the two is that ZabaSearch displays the end result publicly. When people were made aware of ZabaSearch's practices (and a proper forum for complaints), they complained. This raises the question of whether or not users would complain about behavioral profiling if they could see the end results and knew where to file a complaint... . The data from multiple previous surveys that we looked at all point to user concerns over websites collecting information about them and using it to deliver targeted ads. Public display is only part of the reason people complained about ZabaSearch. What was of ultimate concern was control. Users want control over who can collect, share and use information about them."

The UC Berkeley researchers argue that a critical problem here is the lack of clarity and transparency, along with deceptive phrasing, in the privacy disclosures at sites. While the researchers did find that most sites did not share personal information with third parties, they criticized publishers for partial disclosures. "Web site operators should reevaluate a common practice we discovered: claiming that they do not share information with third parties, but allow third-party trackers. We think these statements are inherently contradictory," the study says. They also explore how many sites are nested within larger media corporations whose many affiliations and subsidiaries are unknown to the users.

To their credit, these researchers hold just about everyone's feet to the fire in a tighter, more fact-based argument than we usually see from others like regulators, watchdogs and other academics. The FTC itself is targeted as a part of the problem because its own complaints procedures are too hidden and difficult to use. The researchers suggest we might know more about what does and does not concern Americans about online privacy if there were visible channels for voicing those concerns. The biggest problem for consumers, regulators, perhaps even the industry is that "no one knows who is in charge of protecting privacy," the report concludes.

Both researchers and privacy advocates seem to be going through a complex process. The first stage is research like this and Cathy Dwyer's at Pace University, which involve exposing the complexity of tracking and affiliations that the digital age hath wrought.

The next stage seems to be finding ways of better explaining that complexity to consumers. The Future of Privacy Forum, for instance, has made a project of creating clearer methods for communicating to users the issues surrounding tracking and privacy. The UC Berkeley researchers, too, will make education part of the next step in their process. They are planning to craft an animation that explains the technologies of data collection and how data flows across partners online. It seems to me they have made a pretty good start here. You may agree or disagree with their methods, presumptions and conclusions about online advertising and tracking practices; still, the KnowPrivacy report is the most comprehensive approach I have seen. It is this summer's must-read for behavioral wonks.

Next story loading loading..