Executive Summary

The U.S. Department of Education uses estimates of school-age children in poverty to allocate federal funds under Title I of the Elementary and Secondary Education Act for education programs to aid disadvantaged children. Until recently, the practice has been to base that allocation on the numbers and proportions of school-age children in poverty by county from the most recent decennial census. In 1994 Congress authorized the Bureau of the Census to provide updated estimates of poor school-age children every 2 years, first for counties and subsequently for school districts. Congress also authorized a study of the Census Bureau's program for producing these small-area poverty estimates. That study is being carded out by the Committee on National Statistics' Panel on Estimates of Poverty for Small Geographic Areas, which is charged to advise the Secretaries of Education and Commerce on the appropriateness and reliability of the Census Bureau's estimates for use in Title I allocations. This is the panel's second report.

The Census Bureau's procedure for producing updated county estimates of poor school-age children uses a statistical model that combines data from several sources, including the March Current Population Survey (CPS), income tax return records, food stamp program records, and county population estimates. The data sources used in the model are generally available only 1-2 years after the period to which they refer. For this reason, the Census Bureau decided that the most recent estimates it could produce by the end of 1996 (for the Title I allocations in spring 1997) were for poor school-age children in 1993: all the data for those estimates would not be available until the end of 1995, and model development work would require considerable time after that. The Census Bureau fol-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 1
Small-Area Estimates of School-Age Children in Poverty - Interim Report 2: Evaluation of Revised 1993 County Estimates for Title I Allocations Executive Summary The U.S. Department of Education uses estimates of school-age children in poverty to allocate federal funds under Title I of the Elementary and Secondary Education Act for education programs to aid disadvantaged children. Until recently, the practice has been to base that allocation on the numbers and proportions of school-age children in poverty by county from the most recent decennial census. In 1994 Congress authorized the Bureau of the Census to provide updated estimates of poor school-age children every 2 years, first for counties and subsequently for school districts. Congress also authorized a study of the Census Bureau's program for producing these small-area poverty estimates. That study is being carded out by the Committee on National Statistics' Panel on Estimates of Poverty for Small Geographic Areas, which is charged to advise the Secretaries of Education and Commerce on the appropriateness and reliability of the Census Bureau's estimates for use in Title I allocations. This is the panel's second report. The Census Bureau's procedure for producing updated county estimates of poor school-age children uses a statistical model that combines data from several sources, including the March Current Population Survey (CPS), income tax return records, food stamp program records, and county population estimates. The data sources used in the model are generally available only 1-2 years after the period to which they refer. For this reason, the Census Bureau decided that the most recent estimates it could produce by the end of 1996 (for the Title I allocations in spring 1997) were for poor school-age children in 1993: all the data for those estimates would not be available until the end of 1995, and model development work would require considerable time after that. The Census Bureau fol-

OCR for page 1
Small-Area Estimates of School-Age Children in Poverty - Interim Report 2: Evaluation of Revised 1993 County Estimates for Title I Allocations lowed that plan and in January 1997 provided to the panel county estimates of the number of school-age children in 1994 who were living in and related to a family in poverty in 1993, which were intended to be used for Title I allocations for the 1997-1998 and 1998-1999 school years. In its first interim report (National Research Council, 1997), the panel strongly supported a model-based approach for developing county estimates of poor school-age children. In selecting a model, however, the panel noted that it is important to conduct a thorough evaluation to assess the reasonableness of the model's assumptions, to examine the model's predictions to see that they contain no identifiable systematic errors, and to compare alternative models. Such an evaluation is a critical component of a model-based approach. The panel concluded that the Census Bureau's original model was a substantial step toward the provision of more up-to-date county estimates of poor school-age children. However, there had not been time to complete a full evaluation of the model prior to release of the original 1993 estimates to the panel in January 1997. Therefore, the panel did not recommend sole use of those estimates to allocate funds under Title I. Instead, the panel recommended to the Secretaries of Education and Commerce that an average of the 1990 census estimates (which pertain to poverty status in 1989) and the Census Bureau's original model-based estimates of poor school-age children in 1993 be used to allocate Title I funds for the 1997-1998 school year. In April 1997 the Department of Education allocated Title I funds for the 1997-1998 school year on the basis of the panel's recommended averaging procedure. The department subsequently requested that the panel and the Census Bureau carry out an in-depth evaluation of the Census Bureau's model and reasonable alternative models. It further requested that, on the basis of the evaluation findings, the Census Bureau develop a revised set of county estimates of the number of poor school-age children for 1993 and that the panel evaluate the appropriateness and reliability of the revised estimates for use in Title I allocations for the 1998-1999 school year. (There was neither time, nor a legislative requirement, for the Census Bureau to produce estimates for later than 1993.) Between June and October 1997 the Census Bureau carded out extensive evaluations of its model and alternative models. On the basis of those evaluations, it revised the county model and prepared a revised set of 1993 county estimates of poor school-age children, which were provided to the panel in October 1997. The panel commends the Census Bureau for its work in developing a model-based approach for updated county-level estimates of school-age children in poverty. It also commends the Census Bureau's efforts, carded out in a short time period, to fully evaluate the original county model and alternatives to it. The panel has undertaken a full assessment of the Census Bureau's work and the evaluation results. On the basis of its review, the panel makes the following recommendation for use of the revised 1993 estimates:

OCR for page 1
Small-Area Estimates of School-Age Children in Poverty - Interim Report 2: Evaluation of Revised 1993 County Estimates for Title I Allocations The panel recommends to the Secretaries of Education and Commerce that the Census Bureau's revised 1993 county estimates of poor school-age children be used in the Title I allocations for the 1998-1999 school year. The revised estimates should not be averaged with estimates from the 1990 census, as was done for the allocations for the 1997-1998 school year. The panel concluded that the Census Bureau's revised county model provides estimates of poor school-age children in 1993 that are demonstrably superior to estimates from the outdated 1990 census. The census estimates do not reflect the major changes in the distribution of poverty that occurred between 1989 and 1993. The panel also concluded that the Census Bureau's revised county model performs as well as or better than reasonable alternative models that are practicable for implementation at this time. Performance was measured principally by examining the assumptions of the underlying regression models and comparing the predictions from the Census Bureau's model and alternative models for 1989 with 1990 census estimates of the numbers and proportions of poor school-age children for categories of counties. The changes that were made in the revised county model (principally, modifying one of the predictor variables in the regression equation) improved its performance relative to the original model. The panel notes that the revised county model, like other models, has both strengths and weaknesses. Some level of error in model-based estimates (or in any estimates obtained from a census, from a survey, or indirectly from a model) is inevitable and is not a reason for rejecting such estimates. Yet the county model can very likely be improved with continuing research and development. It may also be possible to reduce the time lag of the estimates. In addition, the model will need to change to accommodate changes in the available data that occur in future years. Hence, the panel recommends that the Census Bureau continue research and development for further improving model-based county estimates of poor school-age children. At the same time, the Bureau is required by law to develop updated estimates for school districts: this is a challenging task, given such factors as the small size of many school districts, the different ways in which districts are defined, changes in district boundaries over time, and the scarcity of relevant data for estimation. For developing updated estimates of poor school-age children for counties and school districts, as well as other small-area estimates of income and poverty, the Census Bureau will need to provide adequate staff and other resources to support a small-area estimation program on a continuing basis.

OCR for page 1
Small-Area Estimates of School-Age Children in Poverty - Interim Report 2: Evaluation of Revised 1993 County Estimates for Title I Allocations This page in the original is blank.