© CQ Press 2011
Data products from the American Community Survey (ACS) include tables and maps that contain estimates of population and housing characteristics and computer files that contain individual person and housing data protected to ensure the confidentiality of respondents. Rather than collect information as part of the ten-year census cycle, the U.S. Census Bureau now collects detailed socioeconomic and housing information every year using the ACS. Sample size limits, however, require that multiple years of data be combined to provide reliable estimates for geographic areas with fewer than 65,000 people. Thus, the ACS provides three different sets of data products: one-year estimates (for areas with 65,000 or more persons), three-year estimates (for areas with 20,000 or more persons), and five-year estimates (for all areas). ACS data products are produced for geographic areas within the
DEFINING THE DATA PRODUCTS
ACS data products were initially designed to be comparable to Census 2000 long-form sample products. However, based on extensive input from data users, the ACS data products have been redesigned and expanded, and additional tables have been added to reflect new content in the ACS that was not included in the Census 2000 long form. There are some differences between the data products provided for the one-year, three-year, and five-year ACS data. Table 1 provides a comparison of the ACS data products with the Census 2000 long-form sample products. These data are available through American FactFinder (AFF) on the Census Bureau’s Web site, accessible either directly or as downloadable files.
The ACS was fully implemented in 2005, and one-year ACS data products are available for each year thereafter. Three-year ACS data products started in 2008, with the release of the 2005–2007 period estimates and subsequent releases for 2006–2008 and 2007–2009. The first five-year ACS data products for 2005–2009 were released at the end of 2010. Annual releases of one-year, three-year, and five-year ACS data products are planned for 2011 and each year thereafter.
ACS data products can be divided into two broad categories: tabulated data products, and microdata files. The tabulated data products contain precalculated estimates of frequently requested totals, percentages, means, medians, and ratios. The microdata files, known as Public Use Microdata Sample (PUMS) files, contain a sample of the individual person and housing data and can be used to produce estimates not already published in the tabulated data products.
One important difference between ACS data products and Census 2000 data products is the inclusion of 90-percent margins of error (MOE) for all estimates in ACS tabulated data products. The MOE is a measure of the precision or reliability of an estimate. The larger the margin of error for an estimate, the less reliable or precise is that estimate. These margins of error can also be used to calculate 90-percent confidence intervals, which indicate that data users can be 90-percent confident that the true population value will usually fall within the range defined by the confidence interval. The margins of error in ACS tabulated data products provide an important tool to help data users understand the reliability of ACS estimates and to draw appropriate conclusions from the data. Measures of reliability are not provided in ACS PUMS files but can be calculated using either a standard formula or the replicate weights that are included in the files.
TABULATED DATA PRODUCTS
The range of ACS tabulated data products is very similar to that for Census 2000. Tabulated data products range from simple profiles for novice users with limited information needs to very detailed tables for advanced users requiring very specific types of information. Note that some of the tabulated data products for one-year ACS data are not provided for three- and five-year ACS data.
Data Profiles provide four separate fact sheets on the broad demographic, social, economic, and housing characteristics for different geographic areas. Data Profiles display derived measures such as means and medians, estimated totals, and percent distributions. These are similar to the Census 2000 Data Profiles DP-1, DP-2, DP-3, and DP-4 provided in the Census 2000 Quick Tables data product. See a sample from a Missouri city. And from a Missouri County.
Narrative Profiles provide clear, concise, textual descriptions and graphical displays of selected measures included in the data profiles. They are a new data product for the ACS, which was not available for Census 2000 data.
Selected Population Profiles
Selected Population Profiles (SPPs) provide a predefined set of selected characteristics from the four Data Profiles for a specific race, Hispanic origin, ancestry, or country of birth population group. Many of the race, Hispanic origin and ancestry group choices are similar to those included in the Census 2000 Summary Files 2 and 4. SPPs are produced for states, congressional districts, and all other geographic areas with a total population of 500,000 or more. However, for an area to receive a one-year SPP, it must also have a population of at least 65,000 in the race, ethnic, ancestry, or country of birth group of interest, and to receive a three-year SPP, it must have a population of at least 20,000 for the group of interest. Selected Population Profiles were not included in the Census 2000 data products and are not currently published for five-year ACS data. See a sample from Missouri.
Comparison Profiles compare Data Profile estimates and distributions across time and denote statistically significant changes between the previous year’s and the current year’s estimates. The Comparison Profiles are currently only available for one-year estimates beginning with the 2006 ACS. There are four types of Comparison Profiles—demographic, social, economic, and housing—and they display the same characteristics as those included in the corresponding Data Profiles product. Comparison Profiles were not included in the Census 2000 data products. See a sample from Missouri.
The Detailed Tables provide the most in-depth ACS data available on all topics and geographic areas and serve as the source data for many other data products. Many detailed tables also have a corresponding “collapsed” version that contains less detail. Geographies that do not have enough respondents to support publication of the full detailed table may only receive the collapsed table. There are more than 1,400 tables in the ACS Detailed Tables product, including tables iterated or repeated for nine race and Hispanic origin groups, as well as allocation tables that show how many of the responses are imputed (that is, provided by Census Bureau edit routines because the data are missing).
Subject Tables display detailed ACS data on a particular topic, such as employment, education, or income, generally drawn from multiple detailed tables. There are currently over 60 ACS Subject Tables, which are similar to those in the Census 2000 Quick Tables product. See a sample from Missouri.
Ranking Tables provide state rankings of key ACS variables. Ranking Tables can be displayed in three ways—as charts, tables, and tabular displays that allow for the testing of statistical significance. The data in the Ranking Tables come directly from the detailed tables. Ranking Tables are only produced for the one-year ACS data and were not included in the Census 2000 data products. There are currently over 90 ACS Ranking Tables.
Geographic Comparison Tables
Geographic Comparison Tables (GCTs) contain the same ACS variables that are included in the Ranking Tables, but compare other types of geographic areas in addition to states. They are produced for one-year, three-year, and five-year ACS data. GCTs are produced for states and substate geographies, including counties, places, congressional districts, and metropolitan areas. See a sample from Missouri.
Thematic maps provide graphical displays of the same ACS estimates available in the Geographic Comparison Tables and Ranking Tables. Different shades of color are used to display variations in the data across geographic areas. Data users can also highlight areas with statistically different values from a selected state, county, metropolitan area, or congressional district of interest. See a sample from Missouri.
Data users may want to download summary files because this product provides a level of flexibility for the user that is not available from American FactFinder (AFF), the Census Bureau’s primary Web site for the dissemination of ACS data products. This flexibility can prove to be very useful, for example, when performing more complex aggregation of geographic areas or table cells. The standard summary files published as part of the one-year and three-year data products contain all of the Detailed Tables published on AFF and are available as a series of comma-delimited text files on the Census Bureau’s File Transfer Protocol (FTP) Web site. The full set of Detailed Tables for the five-year data as well as the Detailed Tables for block groups are only available in the five-year ACS summary file product on the FTP site. The five-year Detailed Tables published on AFF are limited largely to those tables that are comparable to the Census 2000 tables down to the census tract level.
In addition to the standard ACS summary files, the Census Bureau plans to release a five-year race and Hispanic origin summary file that provides Detailed Tables for many race and Hispanic origin groups. This file will be similar to summary file 4 from the 2000 Census. Plans also include a five-year American Indian and Alaska Native Summary File that provides Detailed Tables for many American Indian and Alaska Native groups, similar to the product provided for Census 2000 data. These two additional ACS Summary Files are scheduled for release in 2012 using the five-year ACS data for 2006–2010.
Microdata are untabulated records that contain information collected about individual people and housing units. The ACS Public Use Microdata Sample (PUMS) files are extracts or samples from the confidential microdata that avoid disclosure of information about households or individuals and allow users to generate estimates that are not available in the tabulated data products.
The only geography other than nation and state shown on a PUMS file is the Public Use Microdata Area (PUMA). PUMAs are special non-overlapping areas that partition a state. Each PUMA contains a population of at least 100,000 persons. State governments drew the current ACS PUMA boundaries prior to Census 2000.
PUMS files are available as ASCII text files with comma-separated values (CSV) and in two versions of SAS data sets (PC-SAS files and UNIX files), and are published as part of the one-year, three-year, and five-year data products. The multi-year PUMS files combine annual PUMS files to create larger samples in each PUMA, covering a longer period of time. Tools are available to assist data users with the creation of tabulations, such as the Integrated Public Use Microdata Series (IPUMS) and the Census Bureau’s Federated Electronic Research, Review, Extraction, and Tabulation Tool (DataFerrett).
Similar to the decennial census, ACS data products are published not only for legal and administrative areas, but also for statistical areas. Legal and administrative areas include states, counties, incorporated places, and congressional districts. Statistical areas are defined by the Census Bureau in cooperation with state and local agencies and include regions, divisions, census designated places (CDPs), census tracts, and block groups. Since several years of ACS samples are combined to produce the multi-year data products, the geographic boundaries for multi-year estimates are always the boundary as of January 1 of the final year of the period. For example, the geographic boundaries for the 2005–2009 ACS products are as of January 1, 2009. The boundaries for the 2006–2010 ACS products are as of January 1, 2010, the same as those for the 2010 Census data products.
GENERATION OF DATA PRODUCTS
The Census Bureau’s Disclosure Review Board (DRB) reviews all ACS data products to ensure that the confidentiality of respondents has been preserved. The Census Bureau uses several statistical methods during tabulation and the creation of the PUMS files—swapping of household records, top-coding, age perturbation or modification, and table suppression—to ensure that individually identifiable data will not be released.
Data Release Rules
In addition to disclosure avoidance procedures, data release rules based on the statistical reliability of the ACS estimates are applied to the one- and three-year data products. If more than half of the estimates in a table are not statistically different from 0 (at a 90 percent confidence level), then the table fails and is not published. In order to provide for the aggregation of small geographic areas, this data release rule does not apply to the five-year data products (although some level of suppression may still occur because of disclosure avoidance rules).
CUSTOM DATA PRODUCTS
In addition to data products regularly released to the public, other data products may be requested by government agencies, private organizations and businesses, or individuals. To accommodate such requests, the Census Bureau operates a custom tabulations program for the ACS on a fee basis. Custom tabulation requests are reviewed by the DRB to assure protection of confidentiality before release. Once they are created, customized data products are available to anyone, on request. Examples of customized products include special tabulations for Lower Manhattan in New York City to help deal with the aftermath of the September 11, 2001 attacks, and for the Gulf Coast, in the wake of Hurricanes Rita and Katrina.
Linda A. Jacobsen
Sample Data Products are taken from Springfield (Phelps County) Missouri, close to the last recorded "geographic center" of the U.S. Population.
Did This Article Tell You What You Need to Know? Let Us Know...