star twitter facebook envelope linkedin instagram youtube alert-red alert home left-quote chevron hamburger minus plus search triangle x

New Model Allows Firms to Predict Consumers’ Privacy Preferences Through Users’ Public Social Media Posts

Proposed Model Is More Effective, Less Intrusive and Burdensome than Using Private Data

In today’s increasingly digitally connected world, collecting consumer data has grown rapidly, as has the need to respect and protect consumers’ privacy preferences. In a new study, researchers propose a model to predict users’ personalized privacy preferences by examining their social media posts. Their work suggests that psychosocial traits from public data have greater predictive power than private information.

The study, by researchers at Carnegie Mellon University and the University of Maryland, is published in Information Systems Research.

“Individuals today generate and share vast amounts of information about themselves in the public domain, which can provide a valuable multifaceted view of their behaviors, attitudes, and preferences,” says Beibei Li, professor of IT and management at Carnegie Mellon’s Heinz College, who coauthored the study.

Existing approaches to learning individuals’ privacy preferences lean heavily on seeking explicit user input or accessing private information, which can be seen as intrusive and burdensome. Public posts on social media are considered public data due to their broad accessibility and the implied consent granted by users when choosing to share such information.

In constructing their model, which is rooted in psychological and privacy theories, researchers used deep learning and natural language processing algorithms to identify from users’ tweets (on Twitter, now called X) five categories of theory-driven psychosocial traits that can influence individuals’ privacy preferences: lifestyle activities and habits, personality, risk preference, privacy-related economic preferences, and emotional states.

Then they conducted experiments to explore the predictive power of public data compared to private information. Even without knowing an individual’s private information (e.g., occupation, demographic characteristics), the latent traits revealed by users’ social media posts provided valuable insights about their privacy preferences. In addition:

  • Psychosocial traits gleaned from tweets were more successful in determining users’ privacy preferences than was private information (e.g., demographic data).

  • In public data, lifestyle, risk preference, and emotional states were most important for predicting individual privacy preferences, while in private information, age, employment status, and gender were most important.

  • Psychosocial traits were more important predictors of privacy preferences for young people (ages 18-34), high earners (more than $50,000 annually), and full-time employees.

  • The proposed model can also help platforms and policymakers forecast the consequences of privacy policies ahead of time by simulating user bases and the effects of different levels of privacy risk; without requiring user input, managers can use the model to modulate risk levels and avoid potential discriminatory effects.
“By understanding consumers’ personalized privacy preferences, companies and policymakers can enhance consumers’ trust and improve users’ experience through better privacy policies,” explains Wen Wang, assistant professor of decision, operations, and information technologies at the University of Maryland’s Smith Business School, who coauthored the study. “This understanding also ensures compliance with evolving privacy laws and fosters a digital environment that respects users’ privacy.”

Among the study’s limitations, the authors note that they studied only users who had accounts on Twitter and who actively posted their tweets publicly.


Summarized from an article in Information Systsems Research, Learning Personalized Privacy Preference From Public Data, by Wang, W (University of Maryland), and Li, B (Carnegie Mellon University). Copyright 2024. All rights reserved. 

About Heinz College of Information Systems and Public Policy
The Heinz College of Information Systems and Public Policy is home to two internationally recognized graduate-level institutions at Carnegie Mellon University: the School of Information Systems and Management and the School of Public Policy and Management. This unique colocation combined with its expertise in analytics set Heinz College apart in the areas of cybersecurity, health care, the future of work, smart cities, and arts & entertainment. In 2016, INFORMS named Heinz College the #1 academic program for Analytics Education. For more information, please visit