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 33
The 2000 Census: Counting Under Adversity CHAPTER 2 Census Goals and Uses THE RESULT OF A DECENNIAL CENSUS is a collection of data of various types. One such type is basic population counts, which are used in reapportionment and redistricting, as input to federal funding formulas, and in many other ways. The other prominent type of data product from the census characterizes local areas or population groups from tabulations and analyses of the demographic and socioeconomic variables on the census long form. These census data are provided to users in a wide variety of data products (see Box 2.1). Such products include aggregate or summary tables for geographic areas ranging from the nation as a whole to individual blocks or groups of blocks. Another data product comprises public use microdata sample (PUMS) files, which are subsamples of individual person records from the census long-form sample, which have been carefully reprocessed to minimize the risk of reidentification of respondents. Both summary and microdata products serve federal, state, and local governments, private-sector organizations, researchers, the media, and, ultimately, the public. Collectively, these data are the only available source of detailed information for small geographic areas and population groups. They are also the only source of information about the entire U.S. population, including not only residents of conventional housing units but also residents of special places and group quarters like military bases, dormitories, nursing homes, and prisons.
OCR for page 34
The 2000 Census: Counting Under Adversity Box 2.1 Selected 2000 Census Data Products TABULATIONS OF BASIC (COMPLETE-COUNT) DATA ITEMS (age, sex, race, ethnicity, relationship, housing tenure) Census 2000 Redistricting Data Summary File (P.L. 94-171 File) Population counts, total and aged 18 and over, by 63 race categories; by Hispanic origin; and by geography (county, place, minor civil division, census tract, block group, block) Demographic Profile; Congressional District Demographic Profile Selected complete-count characteristics, by state, county, place, minor civil division, census tract; congressional district (106th, 108th Congresses) Summary File 1 (SF1, 286 tables in all) Population counts for 63 race categories and Hispanic, not Hispanic—down to the block level (file for each state) Population counts for many detailed race and Hispanic categories and American Indian and Alaska Native tribes—down to the census tract level (file for each state) Selected complete-count characteristics—down to the block level (file for each state) National-level file—tabulations for states, counties, and places; urban-rural tabulations Summary File 2 (SF2, 47 tables in all) Complete-count characteristics iterated for many detailed race and Hispanic categories and American Indian and Alaska Native tribes—down to the census tract level (file for each state) National-level file—tabulations for states, counties, and places; urban-rural tabulations TABULATIONS OF LONG-FORM (SAMPLE) DATA ITEMS Demographic Profile; Congressional District Demographic Profile Demographic, social, economic, and housing characteristics (three separate tables), by state, county, place, minor civil division, census tract; congressional district (106th, 108th Congresses) Summary File 3 (SF3, 813 tables in all) Population counts for ancestry groups—down to the census tract level (file for each state) Selected sample population and housing characteristics—down to the census tract and block group levels (file for each state) National-level file—tabulations for states, counties, places Summary File 4 (SF4) Sample population and housing characteristics iterated for many detailed race and Hispanic categories, American Indian and Alaska Native tribes, and ancestry groups—down to the census tract level (file for each state) National-level file—detailed tabulations for states, counties, places
OCR for page 35
The 2000 Census: Counting Under Adversity PUBLIC USE MICRODATA SAMPLE (PUMS) FILES (all items for households and persons selected from the census long-form sample) 1-Percent Sample Files, containing about 1 million household and 3 million person records Geographic identification: state, large areas of 400,000 or more population 5-Percent Sample Files, containing about 5 million household and 15 million person records Geographic identification: state, areas of 100,000 or more population NOTES: All files available on the Internet and CD-ROM/DVD, profiles also available on paper; all files have been processed to protect confidentiality. SOURCE: http://www.census.gov/population/www/censusdata/c2kproducts.html [12/1/03]. The sheer scope and variety of census data and products complicates the task of evaluating a census. This difficulty is compounded by two basic truths. First, census results can only be meaningfully assessed in the context in which those data are actually used. For example, a census could provide outstanding population count data but subpar characteristics data: this could happen if serious problems occurred with the census long form. The data from such a census would be perfectly adequate for some uses but would fail to satisfy others. Similarly, whether the data were to be used as counts or shares, or the level of geographic aggregation, might lead to different judgments about the quality of the data. For instance, a hypothetical census that—for some reason—did an excellent job of collecting information from people living in Western states but not elsewhere would provide good counts for the population living in the West but overestimate the West’s share of the total U.S. population. Furthermore, changes in census processes could improve the precision of counts while hurting the use of the same data to represent changes in counts over time. For example, the change to allow multiple responses to race and ethnicity questions in the 2000 census may make it possible to capture data on more focused demographic groups, but it may complicate inferences about the relative sizes of minority groups relative to past censuses. A comprehensive evaluation of a census must therefore strive to interpret census results in the context of all of their possible uses.
OCR for page 36
The 2000 Census: Counting Under Adversity The second basic truth—and major complication for evaluating a census—is that there is no single, dominant use of census data. While the use of census data to reapportion the U.S. House of Representatives is arguably the primary role of the American census due to the constitutional mandate to do so, it is hardly the exclusive or most demanding use of those data. Hence, there is no single, dominant metric against which census results can be compared in order to unequivocably determine how good they are. As context for our assessments and recommendations in subsequent chapters, this chapter summarizes the goals of the 2000 census and major uses of the data under four headings: (2-A) use for congressional reapportionment; (2-B) use for congressional and state and local redistricting; (2-C) other uses of the basic demographic data—age, sex, race, and ethnicity—collected from everyone enumerated in the census, including use of updated population estimates in federal funding formulae; and (2-D) uses of the additional data that are obtained on a wide range of topics from questions on the census long form. 2–A CONGRESSIONAL APPORTIONMENT Censuses play a fundamental role in the U.S. political system by providing counts of the population for each state once a decade for reallocation of seats in the U.S. House of Representatives. This primary role for census-taking is mandated in Article 1 of the U.S. Constitution and elaborated in Title 13 of the U.S. Code (specifically, 13 USC §141b), which stipulates that the Census Bureau must provide state-level population counts to the president by 9 months after Census Day (i.e., December 31 of the census year under the current schedule). Within 1 week of the opening of the next session of Congress, the president must report to the clerk of the House of Representatives the apportionment population counts for each state and the number of representatives to which each state is entitled according to the prescribed formula (2 USC §2a).1 In turn, the clerk of the 1 The current reapportionment formula, which uses the method of “equal proportions,” was written into law at the time of the 1940 census (Anderson, 1988:189). Apportionment is subject to the constitutional constraint that each state have at least one representative in the House.
OCR for page 37
The 2000 Census: Counting Under Adversity House must inform each state governor of the number of representatives to which each state is entitled (2 USC §2a). This schedule was legislated by Congress in 1929 so that reapportionment would occur automatically each decade, thereby precluding what happened after the 1920 census: beset by rural concerns over sharp population growth in urban areas—largely immigration-driven—Congress could not agree on a bill to reapportion seats on the basis of the census results (see Magnuson, 2000b). Concern over the possible effect of census coverage error on congressional reapportionment—and, more specifically, the effect of competing statistical adjustments to try to correct for said error—has fueled debate over census methodology for years. Although the use of sampling-based census estimates for reapportionment is now prohibited pursuant to a U.S. Supreme Court ruling (see Chapter 3), the potential effects of error and adjustment on apportionment are still viable concerns given the prominence of reapportionment as a use of census data. In studies related to the 1990 census, census data that were adjusted to reflect estimated net undercount shifted one or two House seats between states when input into the “method of equal proportions” formula used for reapportionment compared to results using unadjusted counts. The sensitivity of the apportionment formula to small shifts in population counts was cited by then Commerce Secretary Robert Mosbacher in his decision not to adjust the 1990 census. The major issue with regard to census data for reapportionment is the definition of who is included in the state counts. Historically, major controversies have involved the treatment of three groups: (2-A.1) noncitizens, (2-A.2) Americans overseas, and (2-A.3) people who are not counted in the census itself but who are estimated to be part of the U.S. population through a coverage estimation program, such as the Accuracy and Coverage Evaluation (A.C.E.) Program implemented for 2000. 2–A.1 Treatment of Noncitizens Since 1790 the census has had a goal to count all U.S. residents, including people who are not citizens of the United States (except for tourists and other temporary visitors from abroad and people
OCR for page 38
The 2000 Census: Counting Under Adversity who represent a foreign government). With the rise in the numbers of illegal immigrants in the 1970s and 1980s, lawsuits were filed to exclude them from the apportionment counts (Federation for American Immigration Reform (FAIR) v. Klutznick for the 1980 census and Ridge v. Verity for 1990; see Passel, 2000). The plaintiffs argued that including illegal immigrants unfairly benefited states with larger numbers of them at the expense of states with fewer of them. The defendants countered that the Constitution did not limit the apportionment counts to citizens or potential voters but included the entire population (originally counting slaves as three-fifths of a person). Defendants also argued that it was impractical to identify and exclude illegal immigrants in the census. Federal district courts decided both the FAIR and Ridge cases on the legal grounds of standing. Specifically, the courts ruled that the plaintiffs had not demonstrated injury or harm resulting from the inclusion of illegal immigrants in apportionment counts. The U.S. Supreme Court declined to hear appeals in both cases, upholding the lower court rulings. Although the district court opinions offer some commentary on the constitutional intent to provide representation to everyone and not just to potential voters, this language was not the basis for the decisions. Hence, the issue of including illegal immigrants in the apportionment counts has never been fully resolved. The 2000 census followed historical practice by striving to include all U.S. residents in the apportionment count. Indeed, the results indicated that the census counted many more illegal immigrants than had been estimated to reside in the country, based on survey and administrative records data (see Section 5-D). 2–A.2 Treatment of Americans Overseas Rules about counting Americans who live overseas have varied from census to census. Typically, when counted at all, Americans living abroad have been included in an “overseas” population category and not included in the population of specific states for apportionment or other purposes. Two recent exceptions occurred in 1970 and 1990, when federal military and civilian employees who lived abroad and could be assigned to a “home” state from administrative records were included
OCR for page 39
The 2000 Census: Counting Under Adversity in the state counts for reapportionment, but not in any other data releases. According to McMillen (2000:33): In both cases, the Census Bureau yielded to pressure from Congress to change its procedures to forestall the passage of undesirable legislation. In 1970 the Census Bureau sought to avoid being embroiled in the debate over the Vietnam War. In 1990 the Census Bureau did not want to link the counting of Americans overseas with excluding undocumented aliens from the census. Since 1995, groups representing Americans abroad have lobbied the Census Bureau to enumerate not only federal government employees who live overseas, but also private citizens who live abroad. The Bureau has resisted such pleas on grounds of the difficulties of finding such people and evaluating the completeness of their coverage. It would also be hard to determine how to allocate them to a home state of residence for inclusion in state apportionment counts. The 2000 census did not attempt to count private citizens who lived abroad but, instead, followed the 1990 practice by enumerating federal military and civilian employees (and their dependents), including them in the state apportionment counts (using the home residence listed for them in administrative records), and tabulating them as “Americans overseas” for other purposes. This population totaled 576,000 people in 2000, or about two-tenths of 1 percent of the total U.S. population. After narrowly losing the 435th and final seat in the U.S. House of Representatives to North Carolina, the state of Utah filed the first of two challenges against the apportionment totals on the grounds that the Census Bureau had not treated Americans living overseas equally in the count. Specifically, the case held that Mormon missionaries stationed abroad should be treated in the same manner as military and other federal employees stationed oversees (see Box 2.2). The case ultimately failed but—in the wake of the Utah suit and in response to congressional directives for planning for the 2010 census—the Census Bureau in 2004 will test the feasibility of counting private citizens living abroad, using France, Kuwait, and Mexico as test sites. The Census Bureau also plans a second Overseas Census Test in 2006.
OCR for page 40
The 2000 Census: Counting Under Adversity Box 2.2 Utah v. Evans: Legal Challenges to 2000 Census Under the reapportionment counts issued by the Census Bureau in December 2000, Utah fell short of the 435th (and final) seat in the U.S. House of Representatives by fewer than 1,000 people; the seat was awarded to North Carolina. As a consequence, Utah filed two legal challenges, one of which was decided by the U.S. Supreme Court and both of which raise issues likely to be revisited as the 2010 census approaches. Many of the documents filed by the state of Utah in the two cases remain available on the Internet at http://attorneygeneral.utah.gov/highprofileissues.htm [12/1/03]. Utah v. Evans I: Overseas Enumeration In its first legal filing, Utah challenged the reapportionment counts on the basis that Americans living overseas were treated unequally in the census count. Specifically, the case cited the failure to count missionaries of the Church of Jesus Christ of Latter-day Saints (who temporarily live abroad), even though military personnel and other federal employees stationed overseas are counted. To apply a uniform standard, the case argued that federal employees stationed overseas should be dropped from the apportionment counts. In April 2001, a three-judge panel of the U.S. District Court for the District of Utah ruled against the state. In a November 26, 2001, order, the U.S. Supreme Court affirmed the ruling without comment and declined to take the case on appeal (Nov. 26, 2001, order list, No. 01-283). Utah v. Evans II: Whole-Person Imputation As litigation in the overseas enumeration case proceeded, Utah’s legal team developed another challenge to the apportionment count, this time examining imputation of whole persons to the census when even household size is not known (imputation types 3–5, as in Box 4.2 in Chapter 4). Utah argued that whole-person imputation using hot-deck methods (see Appendix G) constituted sampling, which was prohibited for use in apportionment totals by the U.S. Supreme Court in 1999. If those imputations were dropped from apportionment tallies, Utah would win the 435th seat rather than North Carolina, but no other state’s apportionment total would be affected. Utah initially sought to add its case against imputation to the existing litigation on overseas enumeration, but the U.S. District Court for the District of Utah denied amendment to the case. However, it did so not on the merits of the case but rather on the grounds that it would fundamentally affect the character of the lawsuit. Consequently, the imputation challenge was filed in its own right.
OCR for page 41
The 2000 Census: Counting Under Adversity In November 2001, a three-judge panel of the U.S. District Court for the District of Utah sided with the Census Bureau and ruled that imputation did not constitute sampling. Utah appealed to the U.S. Supreme Court, which announced in late January 2002 that it would hear arguments on March 27, 2002, on both the court’s jurisdiction in the case (i.e., whether Utah had legal standing to bring the case) and on the merits of the case. On June 20, 2002, Justice Stephen Breyer delivered the opinion of a 5–4 court majority in favor of the Census Bureau. The court found that “imputation differs from sampling in respect to the nature of the enterprise, the methodology used, and the immediate objective sought. [These] differences are of both kind and degree.” Consistent with the “actual enumeration” clause in the Constitution, the opinion notes that “in this instance, where all efforts have been made to reach every household, where the methods used consist not of statistical sampling but of inference, where that inference involves a tiny percent of the population, where the alternative is to make a far less accurate assessment of the population, and where consequently manipulation of the method is highly unlikely, [methodological limits under the “actual enumeration” clause] are not exceeded” (Utah v. Evans, 536 U.S. 452, 2002). Justice Sandra Day O’Connor dissented, concluding that imputation is a form of sampling and is, therefore, prohibited under Title 13 of the U.S. Code (see Box 3.1). Accordingly, she declined to comment on the constitutionality of imputation. Justice Clarence Thomas, joined by Justice Anthony Kennedy, also dissented; they agreed with the majority opinion that imputation is not a form of sampling but disagreed on the question of constitutionality, finding imputation and “estimation techniques” inconsistent with their reading of the Constitution’s “actual Enumeration” clause and the historical record of the debate surrounding it. Justice Antonin Scalia dissented in full, arguing that Utah lacked standing to bring the case. Even if Utah succeeded in its challenge and the census totals were recalculated, the president could not and would not be compelled to issue a new apportionment slate to Congress. As a result, Scalia concluded, the federal courts are powerless to “redress the injury” claimed by Utah, and the state is ineligible to bring the case. Scalia would dismiss the case on that basis and did not comment on either the legality or constitutionality of imputation in the census process.
OCR for page 42
The 2000 Census: Counting Under Adversity 2–A.3 Treatment of Uncounted People The goal of the census is to count each U.S. resident once and only once. However, research on census coverage, which began with analyses of undercounts of draft-age men and young children in the 1940 census, estimated a net undercount of the population in every census from 1950 to 1990. (The net undercount is the difference between the number of missed people who should have been included in the census—omissions—and the number of people who were included in the census in error, such as duplicates and other erroneous enumerations.) Research also estimated higher net undercount rates for some population groups than others, such as higher rates for blacks compared with nonblacks and children compared with adults (see Chapter 5). Beginning with the 1970 census, the Census Bureau made special efforts to improve coverage of hard-to-count population groups, although research showed that such efforts were only partly effective (see Citro, 2000c). Beginning with the 1980 census, the Bureau worked to develop a dual-systems estimation (DSE) methodology, based on data from a postenumeration survey and a sample of census records, that could be used to statistically adjust census counts for measured net undercount. The Bureau originally planned to use DSE methods to adjust 2000 census state population totals for congressional reapportionment, but a January 1999 decision by the U.S. Supreme Court precluded the use of adjusted totals for this purpose. The Bureau then planned to adjust census counts for other purposes, but that was not done. The latest Bureau DSE estimates are that the 2000 census overcounted the population by about 1.3 million people or about one-half of 1 percent of the population. This small overcount masked not only large numbers of duplicates and other erroneous enumerations, but also large numbers of omissions. In Chapters 5 and 6 we review the long controversy over census adjustment and assess what is known about population coverage in 2000. On a related note, the second lawsuit filed by the state of Utah challenged the inclusion of certain kinds of census imputations—persons who were not directly enumerated but whose characteristics were imputed instead—in apportionment totals. In June 2002, the U.S. Supreme Court ruled in favor of the Census Bureau’s inclusion
OCR for page 43
The 2000 Census: Counting Under Adversity of imputations (see Box 2.2; we discuss imputation in greater detail in Section 4-D and in Appendixes F, G, and H). 2–B LEGISLATIVE REDISTRICTING A second major function of the U.S. census, which flows from its constitutional mandate to provide state apportionment counts, is to provide small-area population counts for redrawing congressional and state legislative district boundaries. Census small-area data by age, race, and ethnicity (Hispanic origin) are used to ensure that districts satisfy standards for equal population and the requirements of the Voting Rights Act for equitable representation of population groups. In this section, we review the history of redistricting with regard to population size for three time periods: through 1960, from 1960 to 1990, and since 1990. Although the historical discussion may seem to be a lengthy digression, the various battles over standards for redistricting are crucial to understanding data needs served by the census and have important ramifications for the level of geographic detail needed in census tallies. It is important to note that, in the following discussion of equality of district size, we speak principally about comparison of district populations within states and not equality of population across states, among all 435 districts of the U.S. House of Representatives. The latter issue was resolved by the U.S. Supreme Court following the 1990 census; the state of Montana challenged the “equal proportions” method used to allocate seats in the House after it lost its second House seat and Montana’s at-large district became the most populous in the nation. The Court ruled that Congress has discretion to allocate House seats as it sees fit so long as good-faith efforts are made to achieve mathematical equality in representation based on state population, even though inequities in district population (necessarily) arise when compared across states (United States Department of Commerce v. Montana, 503 U.S. 442, 1992).2 2 Following the 2000 census, Montana remained a single-district state; with a 2000 census apportionment population of 905,316, its single at-large district remains the most populous in the nation.
OCR for page 60
The 2000 Census: Counting Under Adversity has been a vehicle to obtain additional information beyond the basic head count by age, race, and sex. By 1910 the census questionnaire had grown to include 34 questions for household residents, including items on age, race, sex, household relationship, marital status, citizenship, education, health, home ownership, language, occupation, place of birth, service in the Civil War, and—in a last-minute addition by the Congress—language of mother and father (Magnuson, 2000a). The 1910 census also included special questions for American Indians, Alaska Natives, and blind and deaf people. Beginning in 1940 statistical sampling was used to accommodate additional questions without burdening the entire population: six new questions on socioeconomic status and housing adequacy were asked of a 5 percent sample of the population in 1940 (6.5 million people). The 1960 census, which first used the U.S. Postal Service for questionnaire delivery, introduced the concept of separate “short” and “long” forms. Since then, the short form has included basic items asked of all U.S. residents, and the long form has included the short-form items together with additional items asked of a sample (about one-sixth of the population in 2000, or 45 million people). The added questions on the 2000 census long form (see Appendix B) included 36 population items, covering such topics as marital status, educational attainment, place of birth, citizenship, language spoken at home, English proficiency, ancestry, military service, year moved into residence, various types of disability, responsibility for grandchildren in the home (new item in 2000), current and prior-year employment status, occupation and industry, transportation to work, and income by type. The 2000 census long form also included 26 housing items, covering such topics as market value of owned home, rent, cost of utilities, characteristics of house or apartment (e.g., number of bedrooms, heating fuel), year structure built, ownership finances (mortgage payment, taxes, home insurance), and number of vehicles (including autos, vans, and trucks).4 The census long-form-sample data have extensive uses in the public and private sectors. At present the long-form sample is generally the only source of nationwide small-area data (for counties, cities, towns, neighborhoods) on the topics it covers, and the size of 4 Numbers of questions or items should be taken as approximate: one question may have several parts that refer to distinct items.
OCR for page 61
The 2000 Census: Counting Under Adversity the long-form sample also makes it a unique resource for analysis of small population groups, such as elderly minorities. Below we briefly describe uses of long-form data by: (2-D.1) the federal government (see Citro, 2000d; National Research Council, 1995b:Apps. C, G, H, M); (2-D.2) state and local governments (see Gaines et al., 2000; National Research Council, 1995b:Apps. E, G, H); (2-D.3) the private sector (see Naymark and Hodges, 2000; National Research Council, 1995b:App. F; Spar, 2000); and (2-D.4) academia (see National Research Council, 1995b:App. D). 2–D.1 Federal Government Uses of Long-Form-Sample Data Many federal agency uses of census long-form data are mandated in law, either directly or indirectly in that the census is the only feasible data source to satisfy a mandate. Indeed, no item was included in the 2000 census if it was not deemed to serve an important federal purpose (see Section 3-B.2). Such purposes include implementation of sections of the Voting Rights Act, allocation of federal funds to states and localities, assessment of charges of employment discrimination, and planning, monitoring, and evaluation of federal programs. Since §203 was added to the Voting Rights Act in 1975, census long-form data have played a key role in ensuring that localities assist voters who have trouble reading English. §203 required counties, cities, and townships to provide election materials and oral assistance in another language as well as in English in areas for which the Census Bureau determined that 5 percent of the voting-age citizens were of a single-language minority and had an illiteracy rate in English (defined as failure to complete fifth grade) higher than the illiteracy rate in English of the entire nation. The 1982 amendments asked the Census Bureau to investigate the usefulness of 1980 census long-form questions on mother tongue and English-speaking ability for determining coverage under the bilingual assistance provision. On the basis of the Bureau’s research, the definition of a covered area became one in which 5 percent of the citizens of voting age spoke a single minority language, said
OCR for page 62
The 2000 Census: Counting Under Adversity they did not speak English very well, and had a higher illiteracy rate than the nation as a whole. The result of this change was to reduce the number of covered areas from about 500 following the 1970 census to about 200 following the 1980 census; about 300 covered areas were identified after the 1990 census. The 1992 amendments extended the bilingual voting assistance provision to 2007 and made some additional minor changes to the definition (see Bureau of the Census, 1976; Kominski, 1985, 1992). About 300 areas were identified as covered under the bilingual assistance provision after the 2000 census (Federal Register, 67[144, July 26, 2002]:48871–48877). Over $200 billion of federal funds are allocated each year to states and localities by means of formulas, many of which directly or indirectly make use of census long-form data (see National Research Council, 2001b, 2003b). Examples include: Medicaid (estimated $145 billion obligated in fiscal year 2002): Reimburses a percentage of each state’s expenditures for medical care services for low-income elderly and disabled people and families with dependent children by a formula that uses per capita income estimates from the U.S. Bureau of Economic Analysis (BEA). BEA develops these estimates by using data from a wide range of administrative records, the decennial census long-form sample and other censuses and surveys, and census-based population estimates (as denominators). The Medicaid formula is also used to allocate funds under other programs (e.g., Foster Care-Title IV-E, with estimated obligations of $5.1 billion in fiscal 2002). Title 1 of the Elementary and Secondary Education Act (estimated $9.5 billion obligated in fiscal 2002): Allocates funds to school districts to meet the needs of educationally disadvantaged children on the basis of a formula that includes estimates of poor school-age children. The estimates were previously derived from the most recent
OCR for page 63
The 2000 Census: Counting Under Adversity census long-form sample; currently, the estimates are obtained from statistical models developed by the Census Bureau, which include census poverty data as one input. Special Education Grants to States (estimated $7.5 billion obligated in fiscal 2000): Allocates funds to states for the education of handicapped children in part on the basis of a formula that includes the number of children in the age ranges mandated by the state’s program and the number of children in poverty in those age ranges. Census long-form-sample data are the source for the estimates. Community Development Block Grants, Entitlement Grants Program ($3 billion authorized in fiscal 2002): Allocates 30 percent of funds to states and 70 percent of funds to localities (metropolitan cities and urban counties) on the basis of the larger amount computed under two formulas. Both formulas use census complete-count and long-form-sample data—total population, poverty population, and overcrowded housing units in the first formula and total population, poverty population, and housing units built before 1940 in the second formula. Home Investment Partnership Program (estimated $1.8 billion obligated in fiscal 2002): Allocates funds to states, cities, urban counties, and consortia of local governments on the basis of a formula that uses census long-form-sample data, such as estimates of rental units built before 1950 occupied by poor families. Workforce Investment Act Adult Program and Youth Activities (estimated $1 billion obligated in fiscal 2002): These programs allocate funds to service delivery areas (one or more counties or cities of 200,000 or more population) by use of formulas that include census long-form-sample data for specified age groups (people ages 22–72 and people ages 16–21, respectively) in families with income not more than the higher of the official poverty line or 70 percent of the Department of Labor lower living standard income level.
OCR for page 64
The 2000 Census: Counting Under Adversity Many federal agencies use census long-form-sample data for program planning, monitoring, evaluation, enforcement, and development of special statistics and surveys. For example: The U.S. Equal Employment Opportunity Commission regularly uses census long-form-sample labor force data on occupation and industry for ZIP codes and other geographic areas to analyze statistical evidence in class action charges of employment discrimination by age, sex, race, or ethnicity. In addition, federal agencies, as employers, use such data to evaluate their own recruitment and promotion systems. The U.S. Department of Transportation uses census long-form-sample data on disability for traffic analysis zones to monitor compliance with the Federal Transit Act and the Americans with Disabilities Act. The Department has helped develop for every census since 1960 a transportation planning package, which provides local governments with detailed information on commuting flows in their areas (cross-tabulating place of residence by place of work for traffic analysis zones made up of groups of census blocks). The U.S. Department of Housing and Urban Development uses census long-form-sample data on rental housing condition to construct fair market rents for non-metropolitan counties and all but the largest metropolitan areas (for which it uses the American Housing Survey). Fair market rents are used to determine rent subsidies to eligible families under the Section 8 Housing Assistance Payments Program; for areas for which census data are used, fair market rates are set at a percentile of the distribution of gross rents (including utility costs) for two-bedroom apartments with electricity and complete kitchen and plumbing facilities and into which the occupant moved within the last 5 years. Since 1950, the U.S. Office of Management and Budget has used census long-form-sample data on commuting patterns to help define metropolitan statistical areas (MSAs).
OCR for page 65
The 2000 Census: Counting Under Adversity In turn, there are many federal agency applications of MSAs, such as determining eligibility for fund allocations. Following every census since 1960, the Science Resources Statistics Division of the National Science Foundation has commissioned a survey by the Census Bureau of a sample of college graduates drawn from the census long-form records, with oversampling of people who reported a science or engineering occupation. People identified as scientists and engineers in the survey are followed up every 2 years to track changes in their training and employment. 2–D.2 State and Local Government Uses of Long-Form-Sample Data State and local governments use small-area census data extensively for such purposes as fund allocation, program planning and evaluation, facility planning, disaster planning, and economic development and marketing. Examples of formula allocation by states include: allocation of over $200 million in Colorado child welfare block grants to counties with a formula that includes census estimates of the county’s number of children below 200 percent of the poverty line; allocation of $5 million in Florida Intensive Crisis Counseling funds to counties with a formula that includes census estimates of the number of poor female-headed households with dependent children (see Butcher and Dunton, 1999). Examples of state and local government program planning and evaluation include: using census long-form-sample data on English language proficiency, language spoken at home, and years of school completed to identify neighborhoods in which residents may require special assistance in using local government services; developing socioeconomic profiles of census tracts to use in applications for funds from federal housing and other grant programs. Examples of facility planning include: using census long-form-sample data on commuting patterns to help redesign
OCR for page 66
The 2000 Census: Counting Under Adversity bus routes and plan new roads; comparing distributions of census small-area characteristics with corresponding distributions of service users (e.g., people visiting health clinics) to determine where new service facilities are most needed (such studies are often done by geocoding administrative records addresses of service users to census geographic areas). Examples of disaster planning include: using census long-form-sample data on vehicle ownership and disability together with basic population data to estimate numbers of people who might need to be evacuated in the case of a disaster (e.g., from an area prone to flooding) and how many of them might need transportation assistance; using census place of work and place of residence data to estimate daytime populations for use in developing disaster planning for employment centers. Examples of economic development and marketing include: using census long-form-sample data on educational attainment, occupation, and labor force participation of the adult population to attract prospective businesses by informing them of the availability of an appropriately skilled labor force; using census long-form-sample data on ancestry and language to inform businesses of opportunities for serving particular markets. 2–D.3 Private-Sector Uses of Long-Form-Sample Data Retail establishments and restaurants, banks and other financial institutions, media and advertising firms, insurance companies, utility companies, health care providers, and many other segments of the business world use census long-form-sample data, as do nonprofit organizations. An entire industry developed after the 1970 census to repackage census data, use the data to develop lifestyle cluster systems (neighborhood types that correlate with consumer behavior and provide a convenient way of condensing and applying census data), and supply address coding and computer mapping services to make the data more useful. Typical private-sector uses of census data include site location, targeting of advertising and services, workforce development, and assessment of compliance with government requirements, such as equal employment opportunity.
OCR for page 67
The 2000 Census: Counting Under Adversity Many retail and service businesses use census population, poverty, income, and labor force data for specific locations to study the possibilities for retail outlets. Some radio stations relate callers’ ZIP codes to census tract profiles to determine areas with residents who are partial to the station’s format, using the information to develop a marketing plan to target those areas. Some churches compare members’ characteristics from a survey to census characteristics for the surrounding community to determine ways in which members are similar to and distinct from the community and use the information to develop an outreach program. All banks are required to use median household income and income distributions by census tract to ensure compliance with federal mortgage lending guidelines regarding race, as well as for meeting other regulatory requirements. Most employers are required to use local-area data on occupation by sex, age, race, and ethnicity to assess compliance with federal equal opportunity employment and anti-discrimination laws. 2–D.4 Research Uses of Long-Form-Sample Data Research studies using long-form-sample data have developed basic understanding of key social processes, such as migration flows and the social and economic effects of the aging of the population. They have also contributed important information to policy makers and the public on such topics as demographic changes, trends in educational attainment, concentrations of people in poverty, and local exposure to environmental hazards. Some research applications use tabular data for small areas. Others use public use microdata sample (PUMS) files, which are available for most censuses back to 1850—the first census to obtain data about individuals rather than summary data for whole households (see Box 2.1; Ruggles, 2000). The aging of the population has led to heightened research attention on the characteristics of older people and how they are
OCR for page 68
The 2000 Census: Counting Under Adversity changing over time. Census small-area tabulations support analyses of migration flows and concentrations of the elderly by geographic area in total and for subgroups defined by such characteristics as income, living arrangements, labor force attachment, and extent of disabilities. Microdata sample files support detailed analyses of different groups of the elderly, such as those with different disabilities and the relationship of disability to education, labor force attachment, and income. Such analyses are not often possible with other data sources, which have much smaller sample sizes than the microdata sample files. The availability of these files for previous censuses permits rich historical analyses of the aging of the U.S. population and the different experiences of the elderly over time. Census data support research on the process of education and the consequences of education on individuals’ life chances. Education researchers use small-area census data on median years of school completed, average income, and employment patterns to understand the challenges facing different school districts. They use microdata sample files to study how age and race groups differ on such factors as educational attainment, income, and housing quality. Path-breaking analyses of 1970 and 1980 census small-area tabulations revealed a large increase during the 1970s in the number of urban neighborhoods in which more than 40 percent of the families had incomes below the official poverty line. The number of people living in such areas also increased dramatically, and a large proportion of them were blacks living in a handful of the nation’s largest cities. These tabulations stimulated research with census and other data on the correlates and consequences of concentrated urban poverty. For example, detailed longitudinal data for families in the ongoing Panel Survey of Income Dynamics were augmented by census characteristics of the neighborhoods in which the sample families lived, permitting rich contextual analyses of family outcomes. Census socioeconomic profiles for small areas, when related to environmental data, permit analyses of the distribution of
OCR for page 69
The 2000 Census: Counting Under Adversity environmental risks for different population groups and are an essential basis for studies of health effects of such local hazards as toxic waste dumps. Census small-area data are an unparalleled resource for analysis of migration flows over time among regions, states, counties, and places and the consequences for changing demographic and economic conditions among different parts of the country. Microdata sample files permit detailed analysis of the characteristics of long-distance migrants versus those who migrate shorter distances or not at all.
OCR for page 70
The 2000 Census: Counting Under Adversity This page intentionally left blank.
Representative terms from entire chapter: