On the Prevalence of Personally Identifiable Information (PII) in the Social Sciences

Replication packages are central to transparency and reproducibility in empirical research, yet they may inadvertently contain personally identifiable information (PII) enabling the identification of research participants whose identities were never meant to be public. We systematically analyze 327 replication packages associated with articles using author-collected data from 11 leading journals across economics, political science, and psychology to quantify the prevalence of PII. Using a semi-automated detection tool and manual validation, we document the types of PII present, assess risk factors associated with their occurrence, and propose guidelines for safer data sharing. We find that 21.1% of packages contain information that could lead to identification of the research participants, with 14.4% of packages containing direct identifiers such as names, email addresses, or phone numbers. The rates range from 16.7% in psychology to 17.8% in politics and 25.9% in economics. We find that PII rates are much higher in field experiments (29.2%) and online experiments (26.6%) compared to laboratory experiments (8.3%). Packages containing data of participants in low- or lower-middle-income countries are also substantially more likely to contain PII. Our findings highlight critical gaps in current research data-sharing practices and inform journal policies to better protect participant privacy while maintaining open science standards.