Preserved in Full, for Future Generations
Spring 2001, Vol. 33, No. 1
By John W. Carlin
Archivist of the United States
Of all the records that the National Archives and Records Administration holds, few are mined for information more thoroughly and more frequently than those of the U.S. Bureau of the Census.
We have preserved on microfilm all the census records dating back to 1790. Historians and scholars consult them for their accounts of our national experience. Sociologists and demographers examine them to map social, cultural, and economic change. Genealogists continually pour over them to trace family histories.
Indeed, next year, when we make public for the first time the records of the 1930 decennial census, there will be long lines of these folks waiting outside our doors for a first look at these seventy-two-year-old documents. (Access to census records is restricted by law for seventy-two years for privacy reasons.)
Among "the people's records" we keep, the census documents are some of the most valued.
That is why we have taken such care to ensure that all the records of the 2000 decennial census are properly preserved until they are made public in 2072. There is a lot in them: about 200 terabytes, or 200 million million bytes, of information in digital image files. It is the equivalent of information printed on paper from fifty thousand trees.
The first thing that researchers in 2072 will see when they look at the 2000 census records is the new electronic Individual Census Record File (ICRF), which we have appraised as a permanent record. This electronic file is an extract of the information individuals listed on their census questionnaire, and it will provide researchers with access to all the information entered on the 2000 census forms. It also will serve as an index to the forms, which we are preserving on microfilm.
With this ICRF, researchers in 2072 will, for the first time, be able to search and retrieve all individual responses from the 2000 census by the individual's name and address. No longer will they have to spend valuable time searching through paper indexes or rolls of microfilm. In addition to ICRF, we are preserving images of the actual filled-out census forms on microfilm.
Since 1960, the Census Bureau has created microfilm copies of the paper census questionnaires and then processed, or "read," the microfilm to tabulate the census. NARA has appraised these microfilm records as permanently valuable and has preserved them for future researchers.
Last year, after the 2000 census, the Census Bureau created digital images of the paper questionnaires and forms that Americans filled out; it did not rely on microfilm images for the enumeration. After the information in the digital images was inspected and verified, the Census Bureau, with our approval, destroyed the paper forms. The bureau proposed to preserve these digital images of the forms for only ten years, and to transfer the final verified information contained in them to NARA in the ICRF.
Based on the public comment we received on the Census Bureau proposal, NARA and the bureau agreed that the records should be kept permanently. However, preserving the digital images in electronic form presented numerous technical challenges because we do not yet have the capability for long-term digital image preservation.
After many discussions and studies, NARA and the Census Bureau decided that microfilm is still the most appropriate method to assure the long-term preservation of digital images. We have preserved the records of every census since 1790 on microfilm, creating a two-hundred-year-old record of decennial censuses at NARA.
As a result, the Census Bureau is going to transfer the digital images of the 2000 census to microfilm via Computer Output Microfilm (COM). For our part, we will make sure that the COM images of the forms are accessible by researchers in 2072 and beyond using the index information that will be contained in the ICRF.
We have not closed the door on having the records accessible by computer; in case that is the chosen method seventy-one years from now. So-called "blowback" technology could be used in the future to convert the records back to digital format if that's what is desired by researchers in 2072 and beyond.
We believe we have made wise decisions to preserve the 2000 census. We listened to our constituencies, and I believe we acted in their best interests and in the best interests of the nation. I think future researchers will agree.