Fight and Flight? How Psychological Reactance and Learned Helplessness Shape Privacy Decision-Making in Algorithm-Driven Digital Contexts
- Wang, Laurent
- Advisor(s): Metzger, Miriam J
Abstract
How do online users respond to privacy risks induced by personalized algorithms? While some studies show that people engage in strategies to protect their personal information vis-à-vis algorithms, other studies instead suggest that people give up on privacy protection because they consider existing mechanisms to be futile in addressing their privacy concerns. To better understand how users negotiate such a “resist-or-resign” tension in privacy protection within algorithm-driven advertising personalization contexts, this dissertation project integrates theoretical perspectives and research on psychological reactance and learned helplessness to develop and test a parallel process model. This model aims to provide a first theoretical account for how people experience reactance and helplessness simultaneously, and how such processes may be shaped by other factors, particularly perceived benefits of personalization and perceived vulnerability to privacy risks. Once these processes are established, a final goal of this project is to test the effectiveness of three theoretically-informed message interventions in evoking reactance and reducing helplessness, with the ultimate goal of motivating privacy protection. Two studies were conducted on Prolific with nationally-representative samples of U.S. adults to fulfill the goals of this project. Study 1 featured a cross-sectional online survey (N = 1,430) to test the proposed model and its moderating factors. Study 2 employed a 2 (personalization: yes vs. no) x 4 (message exposure: offense message vs. efficacy message vs. offense-efficacy combined message vs. no message) between-subjects online experiment (N = 1,315) to establish causality implied in the model and to test the effectiveness of the message interventions. Findings from both studies support the proposed parallel processes. Specifically, when people perceived personalized algorithms to be a threat to their freedom—as a function of either self-reported awareness of algorithmic ad personalization or interaction with a personalized advertising algorithm that was experimentally manipulated—they developed both reactance that encouraged privacy protection intention, as well as helplessness that discouraged their privacy protection intention. In addition, the survey data showed that while both perceived benefits and perceived vulnerability moderated the relationship between perceived threat to freedom and helplessness, these variables did not moderate the relationship between perceived threat to freedom and reactance. Finally, the experimental data supported the effectiveness of (a) the offense message in evoking reactance, (b) the efficacy message in reducing helplessness, and (c) the offense-efficacy combined message in achieving both goals. The results of the studies are discussed in terms of both their theoretical contributions to the communication and online privacy literatures, and in terms of their practical implications for designing interventions that can help users of information and communication technologies who are targeted by personalization algorithms to better protect their privacy.