Errors 2000 Census PUMS file
On February 2, 2010, a NY Times blog entry "Can You Trust Census Data?" examines a working paper that highlights a flaw in the confidentiality protection methods used by the Census Bureau with the 2000 public use micro sample file, mainly affecting the 65+ population.
The paper to which the author refers is a National Bureau of Economic Research working paper. Note that one of the authors is J. Trent Alexander at the Minnesota Population Center.