Historic data from thousands of tapes were recovered before the Census Bureau’s Unisys mainframe computer was decommissioned in 2010. Most of these data still need to be developed for research use.

What was Recovered?*

• A wide variety of Economic (Business) data were recovered from nearly all sectors of the economy, including, Manufacturing, Retail, Services, Wholesale, Mining, Transportation, Construction, Agriculture, and International Trade.

• Several types of data were recovered:
  • Data from censuses and surveys not currently held by CES (learn about our currently available economic data)
  • Additional data for censuses and surveys currently held by CES including additional years, data items, and/or data flags and other processing variables
• See our summary of recovered historical economic microdata  (18 KB) for more information

• Additional years of Annual Survey of Manufactures data going back as far as the early 1950s were recovered
• An early look at the recovered Annual Survey of Manufactures data is in: 
  Becker, Randy A. and Cheryl Grim. 2011. “Newly Recovered Microdata on U.S. Manufacturing Plants from the 1950s and 1960s: Some Early Glimpses.”  (3.6 MB) CES Discussion Paper CES-WP-11-29.

What was Recovered?*

• A wide variety of Demographic (people and household) data were recovered, including:
  • Many Current Population Survey supplements (all months)
  • Income Surveys Development Program (forerunner of Survey of Income and Program Participation)

• Several types of data were recovered:
  • Data from censuses and surveys not currently held by CES (learn about our currently available demographic data)
  • Additional data for censuses and surveys currently held by CES including additional years, data items, and/or data flags and other processing variables
• See our summary of recovered historical demographic microdata (42 KB) for more information

Converting the Data

• For technical reasons, the data were typically downloaded from the Unisys in two forms, assuming the data were all in:
  
  • Excess-3 character set
  • FIELDATA character set
• If the character set of the underlying data is known, recovered files can be converted to usable formats (e.g., ASCII) through known mappings

Challenges

• Data in a record might employ multiple character sets
• Record layout explaining variables, their lengths, and character set employed may no longer exist or may be incorrect
  • A file showing the first 100 records using 5 different character set assumptions created when the data were downloaded can help reveal variable length and character set
  • Existing microdata and/or published data can help in determining variable content
• Data files may have other problems such as missing variables or observations, corrupted data, not matching the description in the paper records, etc.

Step 1: Contact the Recovered Data Team

• Email the Recovered Data Team the following information:
  
  • Recovered data of interest
  • Brief description of planned research
  • Whether or not the recovered data is critical to the planned project
• Recovered data team will respond with basic information about the recovered data of interest

Next Steps

• If research does not hinge critically on recovered data, submit an RDC project proposal
• If research does hinge critically on recovered data, request 6–month feasibility access

Interested in these Recovered Data?

Email the Recovered Data Team to learn more.

*A subsequent effort recovered data from additional tapes. Contact the Recovered Data Team for interest in specific surveys and years not in the lists above or among the restricted-use microdata already available to researchers.