Interpreting Work: Classifying Occupations in the Public Use Microdata Samples

Describes the process involved in creating the occupational codes for the 1850, 1880, and 1920 Public Use Microdata Samples (PUMS). PUMS are random samples of households taken from the US decennial censuses and coded into computer format for use by historians and social scientists. Much of the work for the 1850, 1880, and 1920 PUMS has been done by the Social History Research Laboratory at the University of Minnesota. The process for encoding occupational data was especially difficult because each census recorded such information differently. The laboratory solved this problem by using a common classification scheme and creating a data dictionary to translate the manuscript census information into codes. Additionally, they used occupational groupings of the 1950 census to create a supplementary coding scheme that provided more detailed information. In many instances the PUMS clearly demonstrate that the Census Office's shoddy procedures often led to dramatic undercounts of many occupational categories, particularly female domestic servants.