The SAIPE program produces the following state and county estimates:
In addition, in order to implement provisions of the No Child Left Behind Act of 2001, we produce the following estimates for school districts:
The estimates are not direct counts from enumerations or administrative records, nor direct estimates from sample surveys. Instead, for states and counties, we model income and poverty estimates by combining survey data with population estimates and administrative records. For school districts, we use the model-based county estimates and inputs from the decennial census and federal tax information to produce estimates of poverty. See the Methodology page for further details on the models and see Information about Data Inputs for details on the data sources.
Beginning with the SAIPE program's estimates for 2005, data from the American Community Survey (ACS) are used in the estimation procedure; all prior years used data from the Annual Social and Economic Supplements of the Current Population Survey. Further details are given in a 2007 SAIPE report, Use of ACS Data to Produce SAIPE Model-Based Estimates of Poverty for Counties [PDF 3.4M].
The U.S. Census Bureau, with support from other Federal agencies, originally created the SAIPE program to provide more current estimates of selected income and poverty statistics than the most recent decennial census. A brief history of the SAIPE program can be found on the Origins of the Project page. Prior to the creation of the SAIPE program the decennial census was the only source of income distribution and poverty statistics for households, families, and individuals if one needed data for "small" geographic areas, e.g., counties, cities, and other substate areas. The ten-year span between the release of decennial census data left a large gap in information concerning fluctuations in the economic situation [PDF - 191k] of the nation and local areas.With the full implementation of the ACS in January 2005, single-year direct survey estimates for counties and other areas with a population of 65,000 or more are available. Ultimately, the ACS will provide direct survey estimates for all counties and school districts, as well as for other small geographic areas (e.g., census tracts). For areas with populations less than 65,000, these direct survey estimates will be based on 3-year or 5-year accumulations of ACS data, depending on population size of the area. Since modeling produces estimates with reduced sampling error, the SAIPE program will continue to produce single-year model-based estimates for all states, counties, and school districts.