Users of the Public Use Microdata Sample (PUMS) in DataFerrett found that the tool was incorrectly applying person weights (PWGT) instead of housing weights (WGT) to housing variables when making custom tables for an unknown period of time. Users who attempted to switch weights saw the same total regardless of the weight used. Users who downloaded the data from DataFerrett or other sources, such as the PUMS FTP website, had no issues. This issue was first reported in late June 2015, and was corrected on September 3, 2015; we are not certain how long it was a problem.

The issue affected only 1-year PUMS (2008 through 2013) and only those who used the “Make a Table” option in DataFerrett.