Expert Commentary

Misplaced confidences: Privacy and the control paradox

2012 study from Carnegie Mellon University published in Social Psychological and Personality Science on user-controlled privacy settings online.


The sense of security provided by safeguards such as sports helmets, seat belts and birth control pills often impels users to do things they wouldn’t otherwise — ski faster, drive more recklessly or forgo condom use. Scientists call this tendency to respond to safety measures in ways that counteract protection “risk homeostasis,” “the Peltzman effect” or “the control paradox.” Does this effect also apply to user-controlled privacy safeguards on social networking sites?

Unease about Internet use is particularly acute among parents. A 2012 survey, “Parents, Teens, and Online Privacy,” from Harvard’s Berkman Center for Internet and Society and the Pew Internet and American Life Project, finds that 81% of parents of younger Internet users express concern about the commercial exploitation of personal data; and 69% of parents of online teens are also concerned about the way a teen’s reputation is being managed and the future implications of disclosures. The survey estimates that only 39% of parents of teens using social networking sites have helped their children adjust privacy settings.

A 2012 study published in Social Psychological and Personality Science, “Misplaced Confidences: Privacy and the Control Paradox,” investigates the extent to which a user’s sense of control influences the type and amount of personal information a user discloses online. The researchers, from Carnegie Mellon University, conducted three survey-based experiments with more than 450 participants from a North American university on the release or accessibility of personal information online. The goal was ultimately to see how, in practice, humans respond to increased privacy controls.

  • “Paradoxically, participants were more likely to allow the publication of information about them and more likely to disclose more information of a sensitive nature, as long as they were explicitly, instead of implicitly, given control over its publication.”
  • In the first experiment, half of the participants were told that their answers would automatically be published in an online directory, while half were told that information from some participants would be randomly selected for inclusion. “People respond to manipulations of control over release of personal information in a paradoxical way: Even though lower control implied lower objective risk of accessibility and usage of personal information by others, participants were less willing to disclose if they were provided less control over information release.”
  • During the second experiment, participants were told that the online directory would be accessible by just students, or by students and faculty. Participants expressed higher levels of concern when the publication of their online profile was uncertain than a directory accessible to both students and faculty.
  • “Study two supports the central ideas that privacy concerns are affected by control over release of personal information and that reassurances about control … can distract people from concerns about potentially more hazardous accessibility.”
  • In the third experiment, participants could authorize the publication of each of their survey responses; half were given the option to share and publish explicit demographic information about themselves. Findings showed that virtually all participants granted publishing permission for their answers. “As long as people perceive control over the decision to publish personal information and the audience to whom access will be granted, they will indeed decide to publish it, even if the objective risks associated with disclosure increase dramatically.”

The researchers stress that they are not asserting that individuals should be more concerned about online privacy — or that they should disclose less information online — only that the propensity to share more is influenced by structural factors such as site controls. “Control has become a code word employed both by legislators and government bodies in proposals for enhanced privacy production [but] higher levels of control may not always service the ultimate goal of enhancing privacy.”

“Facebook’s CEO Mark Zuckerberg has repeatedly stressed the role of privacy controls as instruments to have ‘more confidence as you share things on Facebook,’ while both Senator Kerry’s bill proposal and the recent Federal Trade Commission’s Privacy Report focus on giving users more [privacy] control,” the researchers note. Given the findings of the study, such controls could have unintended effects. “The paradoxical policy implication of these findings is that the feeling of security conveyed by the provision of fine-grained privacy controls may lower concerns regarding the actual accessibility and usability of information, driving those provided with such protections to reveal more sensitive information to a larger audience.”

Tags: youth, technology, cognition, privacy, Facebook

About The Author