U.S. Census Bureau

 Small Area Income & Poverty Estimates

 Model-based Estimates for States, Counties, & School Districts

Challenge Procedures for Estimates

Challenges must be received within 90 days of the original release date.

The challenge period for the 2005 estimates ends April 7, 2008. Challenges received after this date will not be considered.

The Small Area Income and Poverty Estimates (SAIPE) program produces estimates of median household income for states and counties and estimates of poverty for states, counties, and school districts. These estimates are based on statistical models that use decennial census data, household survey data, administrative records data, and population estimates. Since these are statistical estimates, they contain error; that is, the estimates will not be exactly equal to the “true” median income or “true” poverty numbers or rates. Such “true” values are generally impossible to determine.

Errors in the estimates arise from different sources. There are “statistical errors” that include sampling, nonsampling, and model errors, and “nonstatistical errors” such as processing or boundary definition errors. Statistical errors are inherent to the estimation process and cannot be corrected. We provide estimates of the magnitude of statistical errors wherever possible and to the extent possible. Nonstatistical errors can, in principle, be corrected if (1) corrected information is available and provided to resolve the error, and (2) the error is identified within the challenge deadline and before the estimates are used to make funding allocations. Systematic errors due to deficiencies in the models or procedures used are not so readily corrected, but when discovered can lead us to reexamine and possibly revise our models and procedures for a future estimation cycle.

All estimates depend on the maps that define the geographic areas of reference. If a school district's boundary is defined incorrectly, this may produce errors in the estimates. However, a boundary correction may not change the population or poverty estimates. Corrections may involve areas that showed no population in Census 2000, and any correction that involves moving a census block between 2 districts will not be moved until the SAIPE program releases new estimates. Apart from the geographic boundary information, much of the input data and model statistics cannot be made public due to confidentiality requirements.

If you believe that a SAIPE program estimate is incorrect, you may challenge the estimate and request a review of the input data. You must provide supporting evidence for why you believe the estimate is incorrect. Requests for correction that are specific and provide evidence to support the existence of a nonstatistical error will be more persuasive than requests for correction that are general and undocumented. The complainant bears the burden of proof with respect to the necessity for correction as well as the type of correction sought. Some of the inputs to our estimates, such as individual responses from the decennial census and aggregate data from federal income tax returns, are confidential and cannot be disclosed. We will examine the evidence submitted and, if warranted, investigate the production of the estimate to determine whether a nonstatistical error has been made. If a correction is warranted and the challenge was submitted within the challenge period, the estimate will be corrected. Otherwise, corrections will be incorporated at the next available opportunity.

A challenge will result in a revised estimate if either of the following is found:

  1. The U.S. Census Bureau made an error in processing input data or running the estimation models/programs.
  2. The U.S. Census Bureau made an error in preparing or processing information to define geographic boundaries (e.g., school district boundaries).

Please note that in cases where the Census Bureau processed the information correctly but was provided incorrect input data, the Census Bureau will make reasonable efforts to provide corrected estimates, but corrections of this sort cannot be guaranteed.

Challenges to state, county, or school district estimates from the SAIPE program must be submitted to Chief, Data Integration Division:

In order for us to investigate your request, you must include the following information: 1) your name, mailing address, telephone number, fax number, email address, and organizational affiliation; 2) identification of the estimate you believe is incorrect; and 3) supporting evidence for the existence of an error. Challenges must be received within 90 days of the original release date. This time limit is necessary because the estimates are used in allocation formulas and must be finalized prior to their use in these formulas. The Census Bureau will attempt to resolve your challenge within 30 days of receipt and will notify you if it will take longer.

Additional Information Regarding Challenges

The SAIPE program poverty estimates are based on the official measure of poverty as defined by the federal government.

Supplemental data cannot be used to replace the SAIPE program estimates. We cannot replace our estimates with other data such as school lunch, food stamp, or Temporary Assistance to Needy Families (TANF) participant counts because (1) such data are not true measures of official poverty as defined by the federal government and (2) alternative data are also subject to uncertainty.

Different numbers from alternative data do not necessarily indicate errors in the SAIPE program's poverty estimates. Free and reduced-price lunch (FRPL) program counts, although informative about the economic status of a school district, will not match the official federal poverty measure, since FRPL eligibility is based on the poverty guidelines published by the U.S. Department of Health and Human Services. For the SAIPE 2005 poverty estimates, the official poverty threshold for a family of four containing two related children under age 18 was $19,157, compared with $19,350 set by the 2005 poverty guidelines. Additionally, children in a similar family would be eligible for the free-lunch program as long as family income did not exceed $25,155, and eligible for the reduced-price lunch program if family income was below $35,797. (These guidelines refer to eligibility criteria for school year 2005/2006.)

Additionally, the relationship between poverty and FRPL eligibility is not consistent across geographic areas or over time. For example, in looking at Census 2000 data for counties, we found the number of children who would be eligible for free and reduced price lunches ranged from 1.0 to 14.6 times the number of children in poverty. Comparisons over time are also particularly difficult. The Food and Nutrition Service (FNS) reported average participation of 11.6 million in the Free and Reduced-Price Lunch Program in 1990 and 15.5 million in 2000 (National School Lunch Program: Participation and Lunches Served). The Current Population Survey estimated 12.7 million related children under 18 in families below the poverty threshold in 1990 and 11.0 million in 2000 (Table 3. Poverty Status of People, by Age, Race, and Hispanic Origin: 1959 to 2006). Other measures of need such as Food Stamp and TANF participation will also differ from official poverty measures.

Our current methods for producing school district poverty estimates have undergone a thorough review by a panel of experts of the National Research Council’s National Academy of Sciences (NAS). The NAS panel concluded its work with a favorable recommendation for use of the SAIPE program estimates for allocating funds under Title I, Part A of the Elementary and Secondary Education Act of 1965, re-authorized in the No Child Left Behind (NCLB) Act of 2001. The NAS panel recognized that the estimates are unavoidably subject to errors, and that errors tend to be larger for smaller geographic areas (areas with less population). School districts that have a small number of total residents will likely have relatively larger statistical estimation errors (in percentage terms) than will districts with larger populations, though the estimates for small districts are no more likely to be too high than too low.

The school district estimates are produced under an interagency agreement with the U.S. Department of Education for use in allocating funds under Title 1, Part A of the NCLB Act. Because of the high level of uncertainty for smaller districts, the NCLB Act allows states to use alternative poverty data to redistribute the Department of Education's SAIPE-based Title I, Part A allocations to school districts with fewer than 20,000 residents. Such alternative poverty data must be approved by the Department of Education; your State Department of Education may be able to provide further information about this option.

Source: U.S. Census Bureau, Data Integration Division, Small Area Estimates Branch
For assistance, please contact the Demographic Call Center Staff at 301-763-2422 or 1-866-758-1060 (toll free) or visit ask.census.gov for further information.