When Excessive Perturbation Goes Wrong and Why IPUMS-International Relies Instead on Sampling, Suppression, Swapping, and Other Minimally Harmful Methods to Protect Privacy of Census Microdata

Abstract

IPUMS-International disseminates population census microdata at no cost for 69 countries. Currently, a series of 212 samples totaling almost a half billion person records are available to researchers. Registration is required for researchers to gain access to the microdata. Statistics from Google Analytics show that IPUMS-International's lengthy, probing registration form is an effective deterrent for unqualified applicants. To protect data privacy, we rely principally on sampling, suppression of geographic detail, swapping of records across geographic boundaries, and other minimally harmful methods such as top and bottom coding. We do not use excessively perturbative methods. A recent case of perturbation gone wrong- the household samples of the 2000 census of the USA (PUMS), the 2003-2006 American Community Survey, and the 2004-2009 Current Population Survey-, an empirical study of the impact of perturbation on the usability of UK census microdata-the Individual SARs of the 1991 census of the UK-, and a mathematical demonstration in a timely compendium of statistical confidentiality practices confirm the wisdom of IPUMS microdata management protocols and statistical disclosure controls.

Keywords: IPUMS-International; data dissemination; data privacy; microdata samples; population census; statistical disclosure controls.

Fig. 1

Sex ratio in Census 2000 Microdata: original and “corrected” ages.

Fig. 2

Mean difference between husband’s age minus wife’s, Census 2000 and 2001 American Community Survey.