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 8
Small-Area Estimates of School-Age Children in Poverty: Interim Report I: Evaluation of 1993 County Estimates for Title I Allocations 2 Poverty Estimates Based on Census and CPS Data CENSUS DATA Traditionally, the decennial census has been the source of small-area income and poverty estimates, with each census being used until data from the next census are available. The 1980 census data, covering income and poverty for 1979, were used for Title I allocations through the 1993–1994 school year; the 1990 census data with 1989 income and poverty data were not available for use until the 1994–1995 school year. As a result of changes over time, the data become increasingly outdated over the course of a decade or so and do not reflect current socioeconomic conditions and demographic distributions in the population. For small geographic areas, the changes in a decade, or less, can be substantial. Over the course of a few years a county can experience rapid population growth from new suburban development and expansion, or rapid population loss from outmigration in response to a decrease in employment opportunities. Likewise, poverty rates can increase or decrease substantially because of a rise or decline of an industry, migration, or changes in other economic and social conditions. Census data that may have accurately represented the population at the time the census was taken will not reflect subsequent socioeconomic and demographic changes. As a consequence, areas that experience either large demographic or economic shifts or both over the decade may be disproportionately overfunded or underfunded under Title I allocations that are based on census estimates. Concerns about using decennial census income data that become outdated were reinforced by changes observed between the 1980 and 1990 censuses. Nationally, the number of poor children aged 5–17 rose by 5 percent over the 10-
OCR for page 9
Small-Area Estimates of School-Age Children in Poverty: Interim Report I: Evaluation of 1993 County Estimates for Title I Allocations year period, from 7.7 million to 8.1 million. At the state level, there was considerable variability: 24 states and the District of Columbia experienced declines in the number of poor children aged 5–17 of up to 34 percent; 15 states saw increases of up to 25 percent; 8 states had increases ranging between 25 and 50 percent; and 3 states had increases between 50 and 67 percent (Moskowitz et al., 1993:71). When considering the use of 1990 census data for current allocations, there is similar concern. Income data collected in the 1990 census are referenced to 1989; they do not capture the recession that began in 1990 or the changes in the proportion and geographic distribution of people below the poverty level that resulted from the subsequent rise in unemployment. In addition to not being current, decennial census data on income are themselves estimates, and as such they are subject to sampling error because the data are collected from only a sample of households. In the 1990 census, income data were collected on the "long form" that was mailed to about 1 out of every 6 households—or about 15 million households in the United States. Sampling rates varied from 1 in 2 for very small counties and places (with an estimated 1988 population of less than 2,500) to 1 in 8 for very populous census tracts (or equivalent areas). Although sampling errors are relatively small for large geographic areas, such as states, the sampling errors for smaller geographic areas can be large relative to the estimate. Table 2-1 provides information on the amount of error due to sampling variability in the estimated numbers of poor school-age children (related children aged 5–17) by county from the 1990 census. For example, for 63 counties, the margin of error due to sampling variability is less than 5 percent of the estimated number of poor school-age children.1 The estimates for these counties are thus fairly precise. Moreover, these counties, although a small percentage (2%) of all 3,138 counties in 1990, are large ones: they contained 37 percent of the nation's poor school-age children estimated by the 1990 census. However, for 1,405 counties, the margin of error due to sampling variability is 25 percent or more of the estimated number of poor school-age children. Although these counties contained only 6.4 percent of the poor school-age children in the nation estimated by the 1990 census, the imprecision in their estimates is of concern for the Title I allocation. 1 The margin of error is expressed in Table 2-1 as the relative width of the 90-percent confidence interval; that is, the width of the interval as a percentage of the estimated number. Confidence intervals for a sample estimate are ranges that include the average result of all possible samples with a known probability; they are constructed from the estimate and its standard error (the measure of the magnitude of sampling variability of the estimate). The 90-percent confidence interval for an estimate is 1.645 standard errors below the estimate to 1.645 standard errors above the estimate: there is a 90-percent chance that the 90-percent confidence interval includes the average estimate of all possible samples.
OCR for page 10
Small-Area Estimates of School-Age Children in Poverty: Interim Report I: Evaluation of 1993 County Estimates for Title I Allocations TABLE 2-1 Distribution of Counties by Relative Widths of the 90-Percent Confidence Interval for the Estimated Number of Poor Related Children Aged 5–17 in 1989:1990 Census Relative Width of Counties Poor Children Confidence Intervala Number Percent Number Percent All Counties 3,138 100.0 7,544,737 100.0 Less than 5% 63 2.0 2,818,997 37.4 5 to 10% 236 7.5 1,846,546 24.5 10 to 15% 466 14.9 1,258,897 16.7 15 to 20% 538 17.1 761,149 10.1 20 to 25% 430 13.7 372,733 4.9 25 to 50% 1,061 33.8 449,464 6.0 50 to 75% 238 7.6 31,585 0.4 More than 75% 106 3.4 5,366 (Z)b NOTE: Three counties with no poor related children aged 5–17 in the sampled households are excluded from the table. a The relative width of the confidence interval is the percentage that the width of the 90 percent confidence interval represents of the estimated number of poor related children aged 5–17 in a county. The 90 percent confidence interval is 3.29 times the standard error of the estimate. b Less than .05 percent SOURCE: Data from Bureau of the Census. CPS DATA The CPS is designed primarily to provide monthly estimates of labor force participation, employment, and unemployment. Every March, the CPS collects additional data on income for the prior calendar year from which poverty rates can be determined. The CPS is therefore a more timely data source on poverty than the census. Indeed, the annual March Income Supplement to the CPS provides the official national measure of poverty.2 The March Income Supplement also serves as a basis for some federal fund allocations (Office of Management and Budget, 1993). The CPS sample size is not large enough to produce detailed information on the changes that occur over time in the geographical distribution of the popula- 2 The Survey of Income and Program Participation (SIPP) is another source of up-to-date income and poverty data. Two Committee on National Statistics panels have recommended that SIPP become the official source of annual national poverty estimates in place of the March CPS (see Citro and Kalton. 1993; Citro and Michael, 1995).
OCR for page 11
Small-Area Estimates of School-Age Children in Poverty: Interim Report I: Evaluation of 1993 County Estimates for Title I Allocations TABLE 2-2 Change in the Total Estimated Number of Poor Persons between 1989 and 1993 for Selected States: March CPS Data State Change in Poverty (in percent) Standard Error of Estimate (in percentage points)a United States 24.5 1.8 Florida 52.3 12.8 California 44.2 9.4 New Jersey 35.8 15.3 Ohio 27.2 12.3 Pennsylvania 26.8 12.1 New York 26.1 8.7 North Carolina 23.0 11.4 Massachusetts 19.6 13.9 Michigan 19.0 10.6 Illinois 7.3 10.1 Texas 4.1 8.0 a 3.29 times the standard error gives the 90-percent confidence interval. SOURCE: Data from Bureau of the Census. tion in poverty, but the survey can provide some useful indicators. They can illustrate how large changes can occur over short periods of time and how different areas can experience substantially different rates of change. Consider, first, the changes in the distribution in the number of poor people of all ages between 1990 and 1994 (income in 1989 and 1993). The CPS sample is sufficiently large to estimate such changes for 11 states, although the estimates are subject to large sampling errors; see Table 2-2.3 Overall, the estimated total number of poor people in the country increased by 24.5 percent, but with a wide range across states: 52 percent for Florida and 44 percent for California, but only 7 percent for Illinois and only 4 percent for Texas. Statistical sampling error affects the precision of these estimates, but it is still clear that there were changes over the period and that they differed among states. The CPS data, when grouped by selected categories of counties and averaged over 3 years to improve precision, show similar changes in the estimated number of school-age children in poverty, which increased for the nation as a whole by 19.6 percent between 1989 and 1993; see Table 2-3. The increase is evident for counties in all regions of the country, in metropolitan and nonmetropolitan areas, 3 For these 11 states, the sample was designed to meet reliability requirements for consecutive monthly changes in the unemployment rate.
OCR for page 12
Small-Area Estimates of School-Age Children in Poverty: Interim Report I: Evaluation of 1993 County Estimates for Title I Allocations TABLE 2-3 Estimated Number of Related Children Aged 5-17 in Poverty by Selected Categories of Counties: 1989 and 1993, March CPS Data County Category Children in Poverty, Income Year 1989a Children in Poverty, Income Year 1993b Change in Poverty between 1989 and 1993 (in percent) U.S. Total 8,036.000 9,613.000 19.6* Metropolitan Central 5,608,000 6.853,000 22.2* Other 362,000 471,000 30.1* Nonmetropolitan 2,066,000 2,289,000 10.8* Regionc Northeast 1,312,000 1,636.000 24.7* Northcentral 1,754.000 1,986,000 13.3* South 3,396.000 3,813,000 15.7* West 1,674,000 2,178,000 30.1* Population Size Under 9,999 202,000 243,000 20.3 10.000–49.999 1.489,000 1,538.000 3.3 50,000–99.999 759,000 927,000 22.2* 100,000–499,999 2,143,000 2,448,000 14.2* 500,000–999,999 1,229.000 1.510.000 22.9* 1 million and over 2.214,000 2,947.000 33.1* * Statistically significant difference from 0 using a 10 percent significance level. a The estimates are 3-year centered averages. For 1989 estimates, averages of March 1989, 1990, and 1991 CPS data were used (reported income in 1988, 1989, and 1990, with population controls derived from the 1980 census). b The estimates are 3-year centered averages. For 1993 estimates, averages of March 1993, 1994, and 1995 CPS data were used (reported income in 1992, 1993, and 1994, with population controls derived from the 1990 census, including an adjustment for the estimated undercount beginning with the March 1994 CPS). c The Census Bureau regions are as follows: Northeast: Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, Connecticut, New York, New Jersey, and Pennsylvania; Northcentral: Ohio, Indiana, Illinois, Michigan, Wisconsin, Missouri, Minnesota, Iowa, North Dakota, South Dakota, Nebraska, and Kansas; South: Delaware, Maryland, District of Columbia, Virginia, West Virginia, North Carolina, South Carolina, Georgia, Florida, Kentucky, Tennessee, Alabama, Mississippi, Arkansas, Louisiana, Oklahoma, and Texas; West: Montana, Idaho, Wyoming, Colorado, New Mexico, Arizona, Utah, Nevada, Washington, Oregon, California, Alaska, and Hawaii. SOURCE: Data from Bureau of the Census.
OCR for page 13
Small-Area Estimates of School-Age Children in Poverty: Interim Report I: Evaluation of 1993 County Estimates for Title I Allocations and in all population size categories, but there is substantial variation in the size of the increase. The largest increases are for counties with a population size in the category of I million or more (33.1%), "other" (noncentral) counties in metropolitan areas (30.1%), and counties in the West region (30.1%). The smallest increases are for counties with a population size in the category of 10,000–49,999 (3.3%, not statistically significantly different from zero), counties in nonmetropolitan areas (10.8%), and counties in the Northcentral region (13.3%).4 Although the CPS provides more current data than the decennial census, its much smaller sample size limits its ability to produce estimates for smaller areas. The March CPS collects data from only about 60,000 households (50,000 beginning in 1996), containing about 28,000 children aged 5–17, compared with about 15 million households for the 1990 census.5 For all but a few very large counties, the CPS sample size is too small to produce reliable estimates. In fact, there is no CPS sample in over one-half of U.S. counties; only about 1,300 counties of 3,143 counties (in 1994) are represented in the sample. And for those counties for which CPS sample data are available, the estimates of poverty and of the population aged 5–17 are, as a rule, extremely imprecise because of small sample sizes. However, as discussed in Section 3, a model-based approach that combines CPS estimates with administrative data in a statistical model can be used to yield estimates for counties that are more up to date than census estimates and have acceptable prediction errors. The Census Bureau's county-level model increases the CPS sample size for counties by combining 3 years of data.6 DIFFERENCES BETWEEN CENSUS AND CPS DATA The census and the CPS differ in other ways besides sample size. Even for a census year, the decennial census and the CPS do not produce identical results 4 The increases in the number of poor school-age children between 1989 and 1993 are the result of increases in the number of school-age children, as well as of increases in the poverty rate for this group. Consequently, for the United States as a whole, the poverty rate for school-age children increased by less than the increase in the number of poor school-age children (11.1% versus 19.6%). The increase in the poverty rate for school-age children, like the increase in their number, varied across regions of the country and types of counties. 5 In turn, the SIPP sample size, currently 37,000 households, is smaller than that of the March CPS. 6 By combining 3 years of data from the March 1993, 1994, and 1995 CPS to produce estimates for 1993, the number of counties represented in the sample increases from about 1,300 to about 1,500. A new 1990 census-based sample design was introduced beginning in the April 1994 CPS; some counties are included in both the new design and the old (1980 census-based) design, but other counties are included in only one design. The average number of sample households for counties represented in one or more of the 3 years is 113; for counties with populations under 10,000, the average number of sample households is 28, and for counties with 500,000 or more people, the average number of sample households is 701. However, several hundred (mostly small) counties with CPS sample households lack any sample households with poor school-age children (see Coder et al., 1996:Tables 1, 3).
OCR for page 14
Small-Area Estimates of School-Age Children in Poverty: Interim Report I: Evaluation of 1993 County Estimates for Title I Allocations TABLE 2-4 Census and March CPS Estimates of Related Children Aged 5–17 in Poverty in 1989, by Selected Categories of Counties County Category Children in Poverty, 1990 Census Children in Poverty. March CPSa Percentage Difference: CPS-Census as Percent of Census U.S. Total 7,545,000 8,036,000 6.5* Metropolitan Central 5,021,000 5,608,000 11.7* Other 347.000 362.000 4.4* Nonmetropolitan 2,177,000 2.066,000 -5.1* Regionb Northeast 1.180,000 1,312,000 11.2* Northcentral 1.641,000 1.754,000 6.8* South 3.174,000 3,396,0O0 3.9* West 1.550,000 1,674.000 8.0* Population Size Under 9,999 197,000 202,000 2.5 10,000–49,999 1,489,000 1,489,000 0 50,000–99,999 843,000 759,000 -9.9* 100,000–499,999 1,990,000 2,143,000 7.7* 500.000–999,999 1,124,000 1,229.000 9.3* 1 million and over 1,901,000 2,214,000 16.5* *Statistically significant difference from 0 using a 10 percent significance level. a The CPS estimates are 3-year centered averages of the March 1989, 1990, and 1991 CPS data (reported income in 1988, 1989, and 1990, with population controls derived from the 1980 census). b See Table 2-3 for the states in each region. SOURCE: Data from Bureau of the Census. with regard to children in poverty. Table 2-4 shows the differences between the 1990 census (1989 income) estimates of the number of poor school-age children and the 1989 CPS estimates for the nation as a whole and for various subcategories of counties. (The CPS estimates are averages of income data for 1988, 1989, and 1990; averaging is used to improve precision given the small CPS sample size in smaller areas; see Appendix Table B-5 for a similar comparison of poverty rates.) Overall, for the U.S. population, the CPS provides an estimate of the number of poor school-age children that is 6.5 percent higher than the decennial census.7 7 Some portion of the differences shown for the United States and various kinds of subnational areas may be due to the use of 3-year centered averages for the CPS-based estimates, which included
OCR for page 15
Small-Area Estimates of School-Age Children in Poverty: Interim Report I: Evaluation of 1993 County Estimates for Title I Allocations For most groups of counties, the CPS estimate is also higher than the census estimate, and there is a suggestion of a pattern in which the ratio of the CPS estimate to the census estimate of poor school-age children in 1989 may increase as a function of county size. The panel conducted an analysis to determine whether there were statistically significant differences among the CPS-census ratios for counties grouped by population size and other characteristics, but did not find such differences. However, this work was very preliminary and needs to be extended.8 Though not yet fully researched and understood, differences between census and CPS estimates of poverty may result from the different ways the income data are obtained. The census and CPS use the same official poverty thresholds9 to determine poverty status, income is counted in both as annual money income received in the previous calendar year, and both are intended to measure the same kinds of income. However, the CPS questionnaire asks respondents to provide income amounts for many more detailed categories than does the census questionnaire. For example, the decennial census asks respondents to provide a combined income amount for Supplemental Security Income (SSI), Aid to Families with Dependent Children (AFDC), and other public assistance or public welfare payments; the CPS asks separately for SSI, AFDC, and other public assistance or public welfare (including the source). Methodological research suggests that more detailed questions elicit more complete income reports (see Citro and Michael, 1995:402–405); however, the extent to which questionnaire differences affect the responses in the CPS and the census is not known.10 The CPS and the census also use somewhat different rules for defining the universe to which poverty applies (see Appendix B). For example, the CPS includes students living in college dormitories as family members in their parental households; the census considers the dormitory the place of residence and excludes residents of college dormitories from the poverty universe. The result is that somewhat more families with college students may be estimated as living in poverty in the CPS than in the census because a college student in a family in- a year (1990 from the March 1991 CPS) in which the poverty rate for school-age children was higher than in either 1989 or 1988. The difference between the 1990 census and the single-year March 1990 CPS in the number of poor school-age children for the United States in 1989 is 4.9 percent, compared with 6.5 percent for the 3-year average figure. 8 Table 2-4 indicates that the differences between the CPS and census estimates of poor school-age children in 1989 are statistically significant (i.e., significantly different from 0) for all county groups except those with small sample sizes. This finding is not surprising given the large national difference in the two estimates; however, it does not support a conclusion that differences between the ratios of CPS estimates to census estimates are statistically significant across county groups. A different comparison would be needed to establish such differences. 9 For example, for a family of four the 1993 (weighted average) poverty threshold level was $14,763. 10 Another difference is that the 1990 census questionnaire, but not the March CPS questionnaire, included a "total income" question. The intent of this question was to permit respondents to enter a single amount if they could not provide amounts by source.
OCR for page 16
Small-Area Estimates of School-Age Children in Poverty: Interim Report I: Evaluation of 1993 County Estimates for Title I Allocations creases its size and therefore its poverty threshold but likely does not add appreciably to its income. The way the data are collected may also result in differences. In the CPS, data are collected through personal contacts (mostly by telephone) made by trained field representatives. In contrast, the census primarily relies on respondents to complete and return a questionnaire by mail. Appendix B discusses in more detail these and other differences between the two data sources, such as coverage errors and the treatment of missing data. Such differences are important to consider in evaluating the appropriateness of moving from use of the decennial census to the CPS as the basis for developing estimates of poor school-age children for the purpose of the Title I allocations. TIMELINESS OF ESTIMATES The CPS provides more timely data than the decennial census; however, estimates of poor school-age children for counties that are derived from the CPS (or from another survey, such as SIPP) will still lag the allocation year by 1 or more years, depending on how quickly the data from the various sources used in the estimation model become available. The estimates that the Census Bureau produced with CPS data for the 1997–1998 Title I allocations are for the numbers of school-age children in counties in 1994 who were poor in 1993. In addition to data availability, the time lag is also caused by the decision to use 3 years of CPS data in the Census Bureau's model to improve the precision of the estimates (see Section 3). Consequently, the estimates will not capture any changes in the extent and distribution of poverty among school-age children that may have occurred since the year to which they apply. Published CPS data indicate that poverty among school-age children for the nation as a whole increased from 17.4 percent in 1989 to 20.1 percent in 1993 and then declined to 18.3 percent in 1995 (Bureau of the Census, 1990:Table 18; 1995a:Table 8; 1996:Table 2).11 Data for 1996 are not yet available, and no data are readily available with which to estimate the changes in the distribution of poverty among school-age children across states and counties. We were asked to evaluate the accuracy of the updated county-level estimates that the Census Bureau was able to produce with available data. We addressed the question of the accuracy of the estimates for the estimation year (1993), not the question of how well the estimates for 1993 predict poverty among school-age children in 1997. There are conceptual and operational issues involved in considering how one could make estimates of poor school-age children yet more timely. We briefly outline some possibly useful directions for research on timeliness in Section 6. 11 These estimates are for related children aged 6–17; estimates are not published for related children aged 5–17.
Representative terms from entire chapter: