Revisiting the uniqueness of simple demographics in the US population

By Philippe Golle. Workshop on Privacy in the Electronic Society, 2006

359 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Sociology and Political Science, Transportation

View Paper: https://www.semanticscholar.org/paper/3e768de55050260559e61576ce5a3182ebe63d75

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Sociology and Political Science

In Transportation

Citations in the corpus

Relationships and data sanitization: a study in scarlet (2010, result): “As a simple example, studies have established that the ZIP code, birthday, and gender and nothing more uniquely identify approximately 63% [38] or 87% [84] of the people in the United States.”
Privacy Violations Using Microtargeted Ads: A Case Study (2010, background): “As pointed out by [43], 87% of all Americans (or 63% in follow-up work by [14]) can be uniquely identified using zip code, birth date, and gender.”
Networked privacy: How teenagers negotiate context in social media (2014, background): “Computer scientists have shown that the very concept of “personally identifiable information” is, at best, murky (Sweeney, 2000), as people can be identified even from supposedly anonymized datasets from which PII has been removed (Ohm, 2010).”
Estimating the success of re-identifications in incomplete datasets using generative models (2019, methodology): “We train our model on the 5% Public Use Microdata Sample (PUMS) ﬁ les using ZIP code, date of birth, and gender and validate it using the last national estimate 50 .”
Anonymization of location data does not work: a large-scale measurement study (2011, background): “The fraction of identifiable individuals ranges from 87% based on the 1990 census data [25] to 63% based on the 2000 census data [7].”
Protecting privacy when mining and sharing user data (2012, background): “As pointed out by [174], 87% of all Americans (or 63% in follow-up work by [67]) can be uniquely identified using zip code, birth date, and gender.”
Data Safe Havens in health research and healthcare (2015, background): “For example, in relation to Criterion 8—which focuses on the appropriate management of data that are potentially identifying - even data items that are individually non-disclosive can become highly identifying if combined together (Golle, 2006).”
Protecting privacy using k-anonymity. (2008, neutral): “32,58,84,85”
Analyzing Leakage of Personally Identifiable Information in Language Models (2023, neutral): “The combination of quasi-identiﬁers ‘gender’, ‘birth date’, and ‘postal code’ re-identify between 63 and 87% of the U.S. population [21].”