Privacy in Context: An Exploration of Factors that Shape What Privacy Means Across Time and Events

What is private, and among whom? The answer is always fluid because it depends on social norms, relationships, identity, and other matters of social context. This dissertation explores how 1) the concept of privacy has slowly changed over the past 50 years in a major U.S. newspaper, and 2) whether an extreme external shock, such as the Covid-19 pandemic, can change the expectations of privacy people hold. Following an introduction, Chapter 2 takes a macro perspective to analyze differences in the language used to discuss privacy over time and across domains. It draws on news from The New York Times to ask what kinds of privacy are usually reported, and how those articles are framed. It finds that the frames used to report news about legal and social aspects of privacy are similar and have not changed much in the 50-year period I studied. In contrast, reporting about technology draws on very different frames that have shifted over time. This suggests that while concepts of digital privacy have evolved over the past three decades, the social concept of privacy in other domains is quite stable. This research also poses the theory that the stark differences in frames leads to more transactional, less relational concepts of technological privacy.While Chapter 2 shows discourse about privacy is largely stable, an immensely impactful social event may shift public opinion quickly. High-profile efforts to use data to combat the Covid-19 pandemic — for example, via digital contact tracing — could lead people to accept less privacy. In Chapters 3 and 4, I consider how perceptions of the pandemic affected the privacy preferences people held. Chapter 3 draws on an ego-centered experiment to test whether relationship-centered messages about the risks of the pandemic cause people to expect less privacy. I find even though these messages do not directly discuss data or privacy, they do lower respondent expectations of privacy in some circumstances. I also conduct correlational analysis of the data to understand how attitudinal and demographic traits relate to differences in privacy expectations. This chapter suggests that more directive messaging and privacy safeguards may be needed to prevent the unnecessary erosion of privacy during times of emergency. Chapter 4 addresses an unanswered question from Chapter 3: do people accept less privacy altruistically, or does thinking of others alter expectations of risk that then inform self-motivated behavior? I draw on survey data collected over the course of the first year of the Covid-19 pandemic to explore whether privacy preferences are better predicted by personal expectations of risk or social signals about risk. I find both factors are important predictors of privacy preferences, suggesting that privacy preferences may change during emergencies both due to self-motivated and other-regarding behavior. However, this analysis also finds evidence of a contrarian effect, suggesting that using social influence as a mechanism to affect privacy preferences can backfire. Overall, this dissertation contributes to theory about privacy in two ways. First, it applies computational techniques to analyze the meaning of privacy across domains and over time. This contributes substantive context to the study of privacy, which is typically constrained to analyzing limited periods and contexts. Second, it tests whether sociological theories — such as social influence and network cognition — affect privacy preferences. It strengthens our understanding of privacy by improving our explanations and by studying these mechanisms at a socially important time. Finally, it contributes to the field of public health by considering the mechanisms that inform privacy preferences compared to other health behaviors. As public health develops digital surveillance methods, understanding privacy as a health-related behavior is useful