SAIPE is the Small Area Income and Poverty Estimates program of the U.S. Census Bureau. We produce model-based estimates annually of income and poverty for school districts, counties and states. We do not produce projections or estimates for other geographic units such as towns, cities, or metro areas. The following estimates are produced:
School districts:
Counties and states:
For more information, please see our About SAIPE page.
Return to top.
We use statistical models to create the estimates. The models relate state and county estimates of income and poverty from the American Community Survey (ACS) to other indicators of income and poverty. These indicators are based on summary data from federal income tax returns, SNAP benefits data (formerly known as Food Stamp Program data), the most recent decennial census, intercensal population estimates, Supplemental Security Income recipiency, and economic data from the Bureau of Economic Analysis (BEA). These estimates are then combined with direct estimates from the ACS sample to provide figures which are more precise than either set alone. This is a standard method for making statistical estimates for small areas. We refer to the final combined estimates as "model-based." More information is available at Methodology.
Return to top.
Poverty status is defined by family; either everyone in the family is in poverty or no one in the family is in poverty. The characteristics of the family used to determine the poverty threshold are: number of people, number of related children under 18, and whether or not the primary householder is over age 65. Family income is then compared to the poverty threshold; if that family's income is below that threshold, the family is in poverty. For more information, please see Poverty Definition and/or Poverty.
Return to top.
Yes. In its final report, the National Academy of Sciences Panel on Estimates of Poverty for Small Geographic Areas (2000) recommended that the Census Bureau investigate using estimates from the ACS in models for small area income and poverty estimates. Full implementation of the ACS began in January 2005, and the SAIPE program incorporated those data into the 2005 estimates and forward. All prior years of estimates were created using data from the Annual Social and Economic Supplement (ASEC) of the Current Population Survey (CPS). See Estimation Procedure Changes for the 2005 Estimates for more details.
Return to top.
A confidence interval is a range of values that, with some level of certainty, contains the true value the estimate is approximating. For example, each state and county estimate is listed with a 90-percent confidence interval. This means that for a particular estimate, if we carried out our estimation procedure and created a confidence interval 100 times, the actual value that the estimate is approximating is in that interval approximately 90 out of the 100 times. For more details about confidence intervals, please check a basic explanation of confidence intervals. For information on confidence intervals of the difference between two estimates, please check the general cautions page.
Return to top.
We see two ways to improve the statistical precision of the income and poverty estimates. First, by improving the models, and second, by improving the input data. We continue to explore and assess the utility of alternate model formulations and statistical approaches. The current estimation methods use administrative record sources uniformly available for the entire nation. If access to administrative record data at both the federal and state levels were expanded, we could explore new variables and potentially improve the accuracy of the model-based estimates. We welcome suggestions, and they can be sent to hhes.saipe@census.gov.
Return to top.
With its release of 2006 and 2007 SAIPE estimates in December 2008, the SAIPE program reduced the release lag in SAIPE program estimates by an additional year. The estimates are as current as feasible given the timescale of availability of the various survey and administrative data used. For example, the SAIPE program's 2007 estimates utilized federal tax return data for 2006, which were filed in 2007, and were available to our program in 2008. The 2007 SAIPE estimates also used data from the 2007 ACS which were collected throughout 2007, tabulated and made available to the SAIPE program in August 2008.
Return to top.
Prior to release of income year 1998 data, county data were produced every two years, for odd-numbered years. The Improving America's Schools Act of 1994 required estimates of poverty for school districts every two years. Both state and county-level estimates are produced as the building blocks for these estimates. We began producing county estimates for even-numbered years in 1998 to meet the demand for more current county-level estimates. State-level estimates have been produced annually since 1995 to meet requirements of The Personal Responsibility and Work Opportunity Reconciliation Act of 1996. School district estimates are also produced annually as required by the No Child Left Behind Act of 2001.
Return to top.
Please see Challenges for information about challenges to the SAIPE program estimates.
Return to top.
Return to top.