American Community Survey (ACS)


Data on Health and Well-being of American Indians, Alaska Natives, and Other Native Americans, Data Catalog

American Community Survey (ACS)

[ Main Page of Report | Contents of Report ]

Sponsor: U.S. Department of Commerce/U.S. Census Bureau
Description: The American Community Survey (ACS) was designed to provide current estimates of community change, and intended to replace the decennial Census long form by collecting and producing updated population and housing information every year instead of every 10 years. About three million households are to be surveyed each year. The ACS collects information from U.S. households similar to what was collected on the Census 2000 long form, such as income and employment, commute time to work, home value and expenses, type of housing, household composition, health status, and veteran status. The ACS began testing in 1996 and expanded to a national demonstration design from 2000 through 2004. Full implementation into all counties began in 2005.

Each year, a subsample of the ACS is selected to construct the Public Use Microdata (PUMS). This data set is available online for all researchers. Use of the full sample of the ACS is restricted and access is difficult to obtain. However, the full sample data are used to create the published population estimates from the ACS. Both the PUMS data and the restricted-access full ACS sample are discussed in this profile.

Relevant Policy Issues: Demographic and Economic Indicators, Income Disparities, Unemployment Rates, Economic Assistance Program Participation Rates, Measures of Well-being for Families/Households, Housing Quality, Type of Housing, Housing Ownership, and Rental Unit Quality and Cost.
Data Type(s): Survey
Unit of Analysis: Individual and Household.
Identification of AI/AN/NA: Race is self-reported. One household member fills out the ACS questionnaire and reports race for himself/herself and all other members of the household. Instructions in the ACS survey for reporting race are: What is this persons race? Mark one or more races to indicate what this person considered himself/herself to be. Response categories include:
  • White
  • Black or African American
  • American Indian or Alaska Native Print name of enrolled or principal tribe (AI/AN)
  • Asian Indian
  • Chinese
  • Filipino
  • Japanese
  • Korean
  • Vietnamese
  • Other Asian Print race
  • Native Hawaiian (NH)
  • Guamanian or Chamorro
  • Samoan
  • Other Pacific Islander Print race below (OPI)
  • Some other race Print race below

Additionally, the questionnaire contains several questions that either require or permit respondents to write-in their responses. The write-in fields cover the following topics: race, Hispanic origin, place of birth, ancestry, migration, language, place of work, industry and occupation. These write-in responses are then coded. Using these coded responses, further race classification of the data is possible, such as separating American Indian from Alaska Native.

AI/AN/NA Population in Data Set: As noted earlier, there is both the restricted-access full sample of the ACS and the Public Use Microdata set (PUMS). The restricted-access sample provides race categories in greater detail than the PUMS data, while the race categories for the PUMS data are combined into fewer categories.

At the time of this publication, the data from the 2005 full implementation of the ACS are not yet available. The PUMS data will contain about 40 percent of the full ACS sample and represent 1 percent of the population. Based on information from the pilot testing, the PUMS data should contain an adequate number of AI/AN/NA individuals to support analysis targeting this population.

AI/AN/NA Subpopulations: The ACS questionnaire asks for AI/AN individuals to give the name of their enrolled or principal tribe. Total population estimates are provided for the following tribal groupings: Cherokee, Chippewa, Navajo and Sioux. Tribal affiliation is not available in the PUMS data set, but is present in the full restricted ACS data.

The Native Hawaiian/Other Pacific Islander race category is broken out into the following subcategories and total population estimates are reported separately: Native Hawaiian, Guamanian or Chamorro, Samoan, Other Pacific Islander. Again, this level of detail is only available in the restricted use ACS data. The PUMS data only reports the combined category Native Hawaiian/Other Pacific Islander.

Geographic Scope: The geographic scope of the ACS is national. The ACS identifies the following individual geographic areas: nation, state, county, county subdivision, place-county, place, metropolitan statistical area (MSA)/consolidated metro statistical area (CMSA)/primary metropolitan statistical area (PMSA), and Congressional district. The list of geographic identifiers will expand with the release of the 2005 data.

Data from the 2004 ACS are available for over 800 geographic areas, including 244 counties, 203 Congressional districts, most metropolitan areas of 250,000 population or more, all 50 states, and the District of Columbia. From the mid-1990s to 2004, the ACS survey was administered to selected sites. The ACS data collection effort was fully implemented in 2005. The Census Bureau plans to begin releasing the 2005 survey estimates in the summer of 2006, but only for areas with populations of 65,000 or more. For areas with populations of 20,000 or more, data release is planned for the summer of 2008, with estimates for all areasdown to census tract/block group levelby the summer of 2010.

Date or Frequency: Data collection is conducted on a monthly basis. Results are compiled and published annually. At the time of this reporting, the most current available data are for 2004. Release dates are noted in the Geographic Scope section above.
Data Collection Methodology: Surveys are mailed every month to a random sample of addresses in each county. The self-enumeration procedure uses several mailing pieces, including a prenotification letter, the American Community Survey questionnaire, a reminder card, and a replacement questionnaire if the original questionnaire is not returned in a timely manner. If a household does not respond in 6 weeks, Census Bureau staff will attempt to contact the respondent by telephone to complete the survey. If that, too, fails, about one in every three addresses remaining will be visited by Census Bureau staff for an in-person interview.
Participation: Mandatory
Response Rate: The ACS does not report an unweighted response rate. Weighting is used because not all housing units have the same probability of selection. The weighted response rate for the 2004 ACS is reported as 93.1 percent, due to a special budget issue. From 2001 to 2003, response rates ranged from 96.7 percent to 97.7 percent.
Sampling Methodology: The 2004 ACS used a two-stage stratified annual sample of approximately 830,000 housing units. Step one of the sampling design was to divide the U.S. into primary sampling units (PSUs); step two was grouping the PSUs into strata based on independent information. After that, one pair of PSUs from each stratum was selected. The 2004 population estimates were derived from 568,966 final interviews. Beginning in 2005, the first stage of sample was no longer used, as all counties will be included in the sample.
Oversample of AI/AN/NA Population: A larger proportion of addresses are sampled for small governmental units including American Indian reservations. The monthly sample size is designed to approximate the sampling ratio of Census 2000 Long Form, including the oversampling of small governmental units.
Analysis: The ACS is a weighted data set. The PUMS methodology report gives a detailed explanation on how to apply appropriate weights when conducting analysis and how to calculate standard errors for the survey variable(s) of interest. Each variable in the ACS data set has an associated design effect. The design effects for the race categories in the ACS are listed as:
  • White alone = 2.5
  • Black or African American alone = 3.1
  • AI/AN alone, Asian alone, NH/PI alone, some other race alone = 3.0
Authorization: The American Community Survey is part of the 10-year Census. As such, its legal authority derives from the same statutes that authorize the Census: Title 13 of the U.S. Code, Sections 141 and 193.
Strengths: The data set contains a large number of AI/AN/NA respondents. The data are collected on key policy issues, including housing and economic well-being. There are multiple years of data available.

The AI/AN/NA population is oversampled in the ACS, thereby improving the reliability of estimates for this population by reducing the variance. Coverage rates for the 2004 ACS for AI/AN are reported as 100 percent and as 90 percent for Native Hawaiian and other Pacific Islanders.

Limitations: The tabulations prepared from the PUMS are based on a subset of the 2004 American Community Survey (ACS) sample. Estimates from the ACS PUMS file are expected to be different from the previously released ACS estimates because they are subject to additional sampling error and further data processing operations. The full data set is difficult to obtain (see below).
Access Requirements and Use Restrictions: The ACS PUMS data are available for download at no cost at the Census Bureaus American FactFinder website:

It may be possible for researchers to access the full ACS survey data set, but obtaining permission for this is difficult. Interested researchers can send an application to the Center of Economic Studies. If approved, researchers will need to work with the data at a Census Bureau data site, or at research centers located in various cities.

Contact Information: For general information about the scope and content of the American Community Survey, call 1 (888) 456-7215 or email

Additional information can be found online at:

Where to?

Top of Page

Main Page of Report | Contents of Report

Home Pages:
Human Services Policy (HSP)
Assistant Secretary for Planning and Evaluation (ASPE)
U.S. Department of Health and Human Services (HHS)