Skip to main content


A sample text widget

Etiam pulvinar consectetur dolor sed malesuada. Ut convallis euismod dolor nec pretium. Nunc ut tristique massa.

Nam sodales mi vitae dolor ullamcorper et vulputate enim accumsan. Morbi orci magna, tincidunt vitae molestie nec, molestie at mi. Nulla nulla lorem, suscipit in posuere in, interdum non magna.

Decennial zero obs data files now accessible on compute nodes

Print Friendly, PDF & Email

The Decennial zero obs data files announced earlier are now available on the VirtualRDC compute nodes.

Zero obs data sets

The "zero obs" files are SAS® datasets that contain no records. Their structure, however, corresponds to the internal historical decennial census microdata files available in the Census Bureau's national network of Research Data Centers (RDC).  Researchers can use the zero obs files to test programs before entering an RDC, although these programs would yield no output because the zero obs files contain no data.  The structure of the VirtualRDC replicates the exact directory and filename structure of the Census RDC, allowing for program sequences to be developed and easily ported to the restricted environment of the RDCs.

Also, researchers can produce PROC CONTENTS of the zero obs files to obtain a complete list of all variables in the internal historical decennial census microdata files available to researchers.  Doing so will provide a list of all variables within each dataset, including the variable name, type, length, label and the position of the variable within the records.  For more detailed documentation on the internal versions of the historical decennial census microdata files, please refer to the documentation on the ICSPR website at


Historical decennial census microdata files are available for the 1970, 1980, 1990 and 2000 censuses, but zero obs files are currently available only for the 1970, 1980 and 1990 data.  The datasets available for each decennial census year are the 100 percent (or "short form") microdata files and the sample (or "long form") microdata files.  Each of these datasets is divided into separate files for each record type: geographic, household and person records.

File names

The naming convention for the (zero obs and actual) files is

t13[state][census year][dataset][record type].sas7bdat


  • state is the FIPS/Postal abbreviation for the state (and 'us' is used for national files;
  • census year is given in four digits
    • 1970
    • 1980
    • 1990
  • dataset is represented by a single character
    • h - 100 percent files
    • s - sample files
  • record type is represented by a single character
    • g - geographic record
    • h - household record
    • p - person record

For example, the file for the person records for New York from the 1980 sample file is t13ny1980sp.sas7bdat

Directory structure

        [census year]cen/
                [census year][RECORD TYPE]/

where [census year] and [record type] are defined as above (record type is in CAPITAL LETTERS). For example, the above NY 1980 sample file for person records can be found at


Sample code

The following code is a sample setup for a program that wishes to analyze the sample files for Florida, New York, Oregon, and Tennessee from the 1980 person files:

libname cen1980s '/decennial/1980cen/1980S/microdata';
data cen1980;
set cen1980s.t13fl1980sp

The code will run without error on the VirtualRDC, yielding a well-defined (but empty) file ' WORK.CEN1980' which could be used in further 'analysis'.