What is the ACS?
The ACS is a project of the U.S. Census Bureau that has replaced the decennial census as the key source of information about American population and housing characteristics. The IPUMS database contains samples from the 2000-2016 ACS.
In every census from 1960-2000, about 1-in-6 American households have received a "long form" census questionnaire (the source for the IPUMS samples). The long form consisted of more than 50 questions, while the short form contained only 7. Some respondents found the long form to be a burden, and the Census Bureau was concerned that the unpopularity of the long form contributed to declining response rates. Also, many users of Census Bureau products need annual data, rather than data for 10-year intervals.
The Census Bureau developed the ACS to address these problems. The ACS provides an annual snapshot of the American population similar to that provided by the decennial census long form. The Census Bureau has produced test data from the ACS since 1996. The 2000 ACS provided the survey's first nationally-representative data; the Puerto Rican Community Survey (PRCS) became available starting in 2005 and provides equivalent information for Puerto Rico.
How large are the samples?
The 2000 ACS is an approximately 1-in-750 public use sample consisting of 372,000 person records. Public use samples from the 2001-onward ACS are even larger. The 2001-2004 samples each represent approximately 0.4% of the population, including more than 1,000,000 person records per sample. The 2005-onward ACS datasets are full 1% samples containing more than 2,800,000 person records. The 2006 ACS is the first to contain group quarters. Researchers analyzing multiple ACS samples over time should remove group quarters cases, since they are available only in the 2006 data onward.
Multi-year data are also available that describe characteristics for three-year periods starting with the 2005-2007 period. Data for the 2006-2008 period was released in early 2010. In early 2011, the Census Bureau began releasing microdata for five-year periods, starting with the 2005-2009 period. For more information, see our multi-year ACS page.
What variables are available?
The ACS questionnaires are nearly identical to the 2000 census long form. The ACS samples also contain several questions on involvement in government programs and a fertility question not asked in the census. Starting in 2008, the ACS and PRCS included health insurance questions for the first time; starting in 2009, there is an item on the field of one's bachelor's degree (if any).
Geographic variables differ, however. The 2000-2004 ACS samples contain no geographic information below the state level; they do not identify PUMAs, cities, or metropolitan areas. Starting in 2005, the ACS samples identify Public Use Microdata Areas (PUMAs) and metropolitan status (METRO). The IPUMS version of the 2005-2007 ACS provides the following additional geographic identifiers: CITY, METAREA, PUMASUPR, MIGTYPE1, MIGMET1, MIGCITY1, MIGPUMS1, PWTYPE, PWMETRO, PWCITY, and PWPUMAS. These variables were constructed at the University of Minnesota and are not available via the Census Bureau.
Aside from differences in the availability of variables, there have been changes to the layout and wording of the ACS/PRCS questionnaire. The most dramatic of these occurred between 2002 and 2003; between 2004 and 2005; and between 2007 and 2008. We have noted major substantive changes in the documentation, and users can also follow the links to questions and instructions for each variable.
Is the ACS different from the C2SS?
No. IPUMS originally referred to the 2000 public use sample of ACS data as the Census 2000 Supplementary Survey (C2SS). Since we now have ACS data from 2001 onwards, we refer to the C2SS as the 2000 ACS file.
How do I get the data?
Users can access the ACS and PRCS data via the IPUMS data extract system.