radboud

    Dealing with Princeton's flawed privacy research

    This has been an odd week. Last Friday I got an email from someone asking about my hobby website’s CCPA compliance, ending with

    I look forward to your reply without undue delay and at most within 45 days of this email, as required by Section 1798.130 of the California Civil Code.

    The message sounded more legitimate than the usual spam I get, as it was asking about a real law in the jurisdiction where I live, and because it referred to a real website that I operate. That last line looked to my not-a-lawyer eyes like something a professional litigant might send out when they’re trying to gather information before deciding whether to sue someone. Mass frivolous lawsuits are a thing, after all, and I dreaded the idea that I might have had to defend my personal project in court.

    This Friday, a friend told me that a researcher at Princeton sent the emails as part of a study on CCPA compliance they’re conducting with Radboud University. That changed my whole outlook: the letter came from a fake person with a fake email domain, lying about their intentions, and lying that the CCPA required me to reply to it. The stress it caused me wasn’t fake, though.

    I submitted a link to my story to Hacker News, which a few people saw. Then someone else submitted another story and it took on a life of its own. It turned out that a lot of people got these emails. The researchers stated that they used the Tranco database of “popular” websites, and my tiny little site was only ranked as high as about number 350,000 in that list. I wasn’t alone. Princeton sent similar emails to other personal projects, and stories abounded that companies had hired counsel and incurred legal expenses to reply to complete fabrications. People had been frightened and were becoming angry.

    Based on advice from Hacker News readers, I contacted Princeton’s Research Integrity & Compliance department and Institutional Review Board, and Radboud’s Research Data Management and Ethics Committee with my concerns. Radboud responded quickly. Princeton hasn’t responded.

    What especially bothers me is that I think this is an important subject to study. I’m a Californian and I support the CCPA protecting my privacy. I want to know if companies are complying with their legal obligations, and I think a large research university like Princeton is the right kind of entity to conduct an effective study. I also believe that the researchers had the right intentions and wanted to do a good job. My problem with it is that I think they made a grave error in misrepresenting their legitimate research questions as coming from a fictional person, and wrote it in a way that set off a lot of “oh no, I think I’m about to be sued” alarms.

    I suspect the data collected from misled responses is corrupted beyond repair. For instance, many entities who replied are likely to have formulated a policy solely because they received the email. I think, then, that the appropriate next steps for Princeton and Radboud are to immediately send explanation and apology emails to all the recipients of the original emails, and to delete all responses they received from recipients of the misleading messages.

    This was such an unnecessary mess. It’s a shame because this could have been crafted in a way that resulted in better data and without scaring the research subjects. Do better next time, Princeton.

    Update 2021-12-2: The researches updated their website to read, in part:

    Our top priority has been issuing a one-time follow-up message that identifies our study and that recommends disregarding prior email. We are sending those messages.

    We have also received consistent feedback encouraging us to promptly discard responses to study email. We agree, and we will delete all response data on December 31, 2021.