The Challenge of Classifying Heterogeneity

The federal government is evaluating its standards for racial and ethnic classification embodied in Directive 15 of the U.S. Office of Management and Budget (OMB). Changing sources and numbers of immigrants, complex patterns of regional and metropolitan concentrations of racial and ethnic populations, and evolving trends in intermarriage and self-identification have put pressures on the existing system of classification and have raised questions about its acceptance and utiltity.1

Definitions of race and ethnicity touch every American individually. But the categorization of race and ethnicity is also a public issue with society-wide implications. The challenge facing the federal government is to use a set of categories that is uniform and practical for administration while also acknowledging a diverse population and individual perceptions of identity.

The motives for racial and ethnic classification are widely recognized to be in potential conflict: As retiring Supreme Court Justice Harry Blackmun has stated, (New York Times, April 7, 1994:A13): “In order to get beyond racism, we

1  

On August 28, 1995, OMB issued a Federal Register notice (Vol. 60, No. 166:44674-44693) that summarizes the suggestions for change that OMB has received. The notice also sets forth proposed principles to be used in possible revision of the directive. OMB proposes to rely on results from the March 1996 Census Bureau’s National Content Test and the June 1996 Census Bureau’s Race and Ethnic Targeted Test to understand better the impact of various alternatives data collection procedures for self-identified race and ethnic data. OMB intends to publish a Federal Register notice on the research results in Spring 1997 and to publish a final decision regarding any changes to Directive 15 in mid-1997.



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 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification The Challenge of Classifying Heterogeneity The federal government is evaluating its standards for racial and ethnic classification embodied in Directive 15 of the U.S. Office of Management and Budget (OMB). Changing sources and numbers of immigrants, complex patterns of regional and metropolitan concentrations of racial and ethnic populations, and evolving trends in intermarriage and self-identification have put pressures on the existing system of classification and have raised questions about its acceptance and utiltity.1 Definitions of race and ethnicity touch every American individually. But the categorization of race and ethnicity is also a public issue with society-wide implications. The challenge facing the federal government is to use a set of categories that is uniform and practical for administration while also acknowledging a diverse population and individual perceptions of identity. The motives for racial and ethnic classification are widely recognized to be in potential conflict: As retiring Supreme Court Justice Harry Blackmun has stated, (New York Times, April 7, 1994:A13): “In order to get beyond racism, we 1   On August 28, 1995, OMB issued a Federal Register notice (Vol. 60, No. 166:44674-44693) that summarizes the suggestions for change that OMB has received. The notice also sets forth proposed principles to be used in possible revision of the directive. OMB proposes to rely on results from the March 1996 Census Bureau’s National Content Test and the June 1996 Census Bureau’s Race and Ethnic Targeted Test to understand better the impact of various alternatives data collection procedures for self-identified race and ethnic data. OMB intends to publish a Federal Register notice on the research results in Spring 1997 and to publish a final decision regarding any changes to Directive 15 in mid-1997.

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification must first take account of race. There is no other way.” This is the conflict of those who dream of a “raceless” society, while at the same time wanting to assure that all groups have equal opportunity. Equality cannot be ascertained without measurement, but measurement itself, say critics of classification, may strengthen the distinctions one was hoping to minimize. A NATION OF CHANGE The current federal standards, issued in 1977, are the latest in a history of racial and ethnic categorizations that predate the nation ’s founding. Those categorizations have always been closely associated with the politics and social agendas of their times: they sparked controversy in the writing of the Constitution and have been a source of contention ever since. The first U.S. census, mandated by the Constitution and conducted in 1790, was primarily concerned with political apportionment. It distinguished holders of the franchise —white male property owners—from the general population, and, for taxation purposes, it also differentiated American Indians who were subject to taxation from the larger American Indian population. That first census also classified residents of African ancestry as part of either the free or slave population. During the nineteenth century, changes in categorization reflected increased immigration and race consciousness. After the Civil War, in response to numerous race-based laws, the census added detailed subcategorizations of the African-ancestry population by color and blood quantum (e.g., mulatto, quadroon). Increasing subclassification of the immigrant population also began in the middle of the nineteenth century. Chinese and Japanese residents were enumerated separately, even though their numbers were minuscule: for example, tabulations of the 1870 census showed a count of 47 males and 8 females of “ Japanese race” (Bureau of the Census, 1975). As the century came to an end, immigration from new sources in southern and eastern Europe significantly increased the diversity of the European-origin population: these new immigrants were considered to be of different races than the English and northern Europeans, and the census classifications reflected this difference. Following the introduction of immigration restrictions, beginning in 1882 with the Chinese Exclusion Act and culminating in the Immigration Act of 1924 (which imposed national origin quotas on immigration), the foreign born became a smaller and smaller portion of the white population, and the intensity of identity decreased for many white ethnic groups.2 2   As the composition of the population has changed, so too have the criteria of classification. A major change in the 1980 census was to omit the question asking place of birth of respondents’ parents because, it was believed, so little of the population was still foreign born. Less than 20 years later, however, a new wave of immigration is well under way, and now there are significant pressures to reintroduce the parental nativity question in the decennial census. For a detailed discussion of the history of racial and ethnic classification in the census, see Edmonston and Schultze (1995:Ch. 7).

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification In the last 30 years, several factors have led to major concern about the current racial and ethnic classification: changes in laws relating to minorities; growth in immigration, particularly from non-European sources; the geographic concentration of racial and ethnic minorities, especially in inner-city schools; increased racial and ethnic intermarriage; and increased interest in and reliance on self-identification—rather than observer identification—for classification. As federal and state laws shifted from exclusion of nonwhites to inclusion based on minority status, the demand for clear and consistent statistical definition of minority groups increased. Antidiscrimination and equal opportunity laws—such as the Civil Rights Act of 1964, the Voting Rights Act of 1965, the Fair Housing Act of 1968, the Equal Credit Opportunity Act of 1974, and the Home Mortgage Disclosure Act of 1975—have increased the relevance of collecting and publishing racial and ethnic data. A new period of immigration began with implementation of the 1965 amendments to the Immigration and Nationality Act, through which national origin quotas were extensively revised. In recent decades, two-thirds of U.S. immigrants have come from Asia, Mexico, and Central America, changing the composition of the minority population. In the 1990 census, blacks were reported to comprise 12 percent of the population; Hispanics, 9 percent; Asian and Pacific Islanders, 3 percent; and American Indians and Alaskan Natives, 1 percent (Bureau of the Census, 1992:Table 16). As the number of persons in these categories has increased, so has the contention over the meaning and utility of these and related racial and ethnic classifications. Population projections based on assumed rates of natural increase and continued immigration show further increases in the size of minority populations. In the next 50 years, the non-Hispanic white population is projected to decrease from the 1990 figure of 75 percent to 59 percent of the total U.S. population, and the relative size of the minority populations is projected to change significantly as Asian and Hispanic groups grow rapidly, with the black population projected to remain at about its current relative size. Projections for 2040 show the following population breakdown: Hispanics, 18 percent; blacks, 12 percent; and Asians and Pacific Islanders, 10 percent; and American Indians and Alaskan Natives, 1 percent (Edmonston and Passel, 1994). 3 The current size of the minority population is not unprecedented. In 1790, the time of the first census, blacks constituted about 19 percent of the population; by 1990, the minority groups of blacks, Asian and Pacific Islanders, Hispanics, and American Indians constituted 25 percent of the population. Yet the combination of population projections for the future and the other factors noted above have led to the current debate about racial and ethnic populations and how they are classified. 3   These projections, like those developed by the Bureau of the Census, assume no racial or ethnic intermarriage.

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification DIRECTIVE 15 AND ITS USES Current federal government standards for racial and ethnic classification are embodied in Directive 15 of the U.S. Office of Management and Budget (OMB), “Race and Ethnic Standards for Federal Statistics and Administrative Reporting,” which covers all federal agencies. After several years of development, Directive 15 was issued in 1977 in recognition of the need for consistent race and ethnicity categories. 4 The directive has three components: (1) definitions for five basic race and ethnicity categories; (2) designation of minimum data collection and reporting formats; and (3) formats for data presentation (see Appendix B for the complete text of Directive 15). In sharp contrast with all other statistical directives, the federal standards on race and ethnicity are intended for administrative as well as statistical purposes. They serve as the lens through which administrators and other interested parties view the operation of federal programs. Thus, the directive has played an influential role in determining both data availability and administrative policies. Directive 15 has remained unchanged since it took effect in 1977. 5 The basic race and ethnicity categories specified in Directive 15 are American Indian or Alaskan Native, Asian or Pacific Islander, black, Hispanic, and white. Four of these categories are regarded as “race” and one as “ethnicity.” As shown in Table 1, the race and ethnicity categories can be separated into two formats or combined into one format with five categories. Table 2 shows the criteria for classifying people according to Directive 15. Many observers have noted that Directive 15 is permissive, that is, it man 4   Prior to the directive, several agencies (including the former Office of Education and the Equal Employment Opportunity Commission) had their own categorization policies. In 1974 the Federal Interagency Committee on Education created an Ad Hoc Committee on Racial and Ethnic Definitions to recommend a uniform set of categories for education data, and the report of that committee, released in 1975, became the basis for Directive 15. A year later Congress passed Public Law 94-311, “Joint Resolution Relating to the Publication of Economic and Social Statistics of Americans of Hispanic Origin or Descent. ” It required OMB and several departments—Commerce; Labor; Agriculture; and Health, Education, and Welfare—to develop a variety of programs to collect economic and social statistics on the status of Hispanics. Between the 1975 interagency committee recommendation and the adoption of administration-wide standards in 1978, changes in terminology and classification were made: “Negro ” and “Caucasian” were replaced with “black” and “white,” and persons whose origins were the Indian subcontinent were reclassified from the white category to the Asian or Pacific Islander category. 5   In 1988 OMB circulated a draft revision, adding an “other” race category and requiring classification by self-identification. The proposal was supported by many multiracial and multiethnic groups and some educational institutions, but it was strongly opposed by the relevant committees in Congress, federal agencies—including the Civil Rights Division of the Department of Justice, the Department of Health and Human Services. the Equal Employment Opportunity Commission, and the Office of Personnel Management—and by some nonprofit organizations and large corporations (Katzen, 1993).

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification TABLE 1 Directive 15 Classification Matrix Characteristic Separate Format Combined Format Race White Black Asian and Pacific Islander American Indian or Alaskan Native White, not of Hispanic origin Black, not of Hispanic origin Asian or Pacific Islander American Indian or Alaskan Native Hispanic Ethnicity Hispanic origin Not of Hispanic origin   dates a set of categories, but it allows any finer categorization that can be aggregated into the mandated categories (e.g., specific ethnic groups could be added to an agency’s administrative records). Few agencies other than the Census Bureau have availed themselves of the directive’s permissiveness, but it is likely that some of the changes favored by various interest groups could be accommodated within the general guidelines of the Directive 15. Other features of Directive 15 are not permissive. Agencies may combine data from the major categories in summary reports, but the term “nonwhite” cannot be used to describe such an aggregate. It is not permissible to aggregate data into an “all other” group, except when a report or analysis focuses on one or two identifiable racial or ethnic groups. Moreover, the directive recognizes only one “ethnic” distinction, namely that between Hispanic and non-Hispanic origin, and it requires that Hispanic origin be separate from “racial” distinctions. Individuals are not permitted multiple responses, nor is there an option for “other” or “none of the above.” The directive acknowledges the existence of multiple identities, but specifies the use of the “category which most closely reflects the individual’s recognition in his community” (U.S. Office of Management and Budget, 1978). The language of the present Directive 15 may cause some confusion and misunderstanding. It uses such terms as “majority race” and “principal minority race, ” which reflect the directive’s roots in the legal experiences of civil rights formulation and enforcement. But Directive 15 does not confer legal status. Even though the directive identifies “protected groups,” being defined as a separate category is not the same as being recognized as a protected class for civil rights compliance or being an officially designated ethnic group for other than federal statistical and administrative reporting purposes. The purposes of the directive are for the record-keeping, collection, and presentation of data. The preamble to Directive 15 is unequivocal: “These classifications should not . . . be viewed as determinants of eligibility for participation in any Federal program.” At the same time, statistical data following the standard are used for administrative purposes (e.g., monitoring compliance with civil rights legislation).

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification TABLE 2 Criteria for Classifying People According to Policy Directive 15   Definitional Criteria Racial or Ethnic Group Original People of Cultural Origina Cultural Identification, Affiliationb Racec Race American Indian, Alaska Native North America   Yes   Asian/Pacific Islander Asia, Pacific Islands Black Africa     Black White Europe, North Africa, Middle East Ethnicity Hispanic Mexico, Puerto Rico, Central or South America, other Spanish countries Mexico, Puerto Rico, Central or South America, other Spanish countries     aUsed only to define Hispanics. bUsed to define American Indians and Alaska Natives. cUsed only to define blacks. Source: Hahn (1994) There is a symbiotic relationship between categories for the tabulation of data and the processes of group consciousness and social recognition, which in turn can be reflected in specific legislation and social policy. For example, the designation “Hispanic” was not in common usage prior to its adoption by the Bureau of the Census in the early 1970s, yet its government-wide use as a referent for heterogeneous populations has bolstered the growth of Hispanic or Latino/a identity. OMB’s Office of Statistical Policy, which is responsible for Directive 15, does not and cannot decide which groups should benefit from governmental policies and programs. Researchers as well as federal administrators have repeatedly observed that Directive 15 has been influential far beyond its original intent. The directive itself states only the modest goal of fostering the creation of “compatible, nonduplicated, exchangeable racial and ethnic data by Federal agencies” for three

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification reporting purposes: statistical, administrative, and civil rights compliance. It is clear that compliance was the paramount issue in developing the directive (see Appendix B). Any new directive must fulfill the same purposes as the existing directive. OMB officials have noted that consideration of the revision of Directive 15 is not an occasion for various racial and ethnic groups to press their claims for special recognition. Rather, OMB ’s aim is to improve and update the directive and to reflect changes in and more accurately describe the nation’s population, while preserving consistent categories for the nation’s statistical records. The present directive permits—and a revised directive might be able to encourage—an evolutionary process of change. Experience can be gained by the use of subcategories consistent with the currently mandated categories but responding to newly articulated needs. A process for facilitating the provisional adoption of such subcategories across federal agencies could be written into a new directive or instituted administratively in another way. A recurring theme in the debate about racial and ethnic classification —mirrored at the committee’s workshop and in this report—is the tension between a two-format approach like that in Directive 15 and a single-format approach. As discussed above, Directive 15 provides a choice of either four categories of race and two categories of ethnicity (the ethnic categories are Hispanic or non-Hispanic) or a combined format of five categories of race and ethnicity (see Table 1). The existence of two formats has always been problematic. Whether it is unavoidable is a matter for dispute. Some data users favor a conceptualization that builds on the combined format and eliminates the two-format approach (see below). Others recoil from the presumptions built into such combined format systems. In considering possible revision of Directive 15, OMB will have to resolve the dilemmas inherent in a federal standard with two formats. Federal Experience Virtually all federal agencies have been affected by Directive 15 since it took effect, even those not needing to meet specific statutory requirements. In addition to agencies that deal with employment, housing, education, civil rights, and federal statistics (like the census), lesser known but significant users of the standards include such agencies as the Federal Reserve Board, which tracks recipients of its housing loan program, and the Small Business Administration, which uses the categories in its minority business development programs. In addition, nearly every federal agency uses the standards for compliance with respect to its own personnel requirements. As part of the preparation for the workshop on which this report is based, federal agencies were asked to respond to a set of open-ended questions concerning four issues: (1) the use of Directive 15 for federal data collection; (2) data collection beyond the directive ’s classifications; (3) the utility and adequacy of

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification the directive; and (4) possible revisions of current categories. Many agencies responded (see list in Appendix C) and were forthcoming about their own experiences. Many agency representatives at the workshop also noted the usefulness of the survey itself in fostering interagency communication about experience with the directive · The agencies provided both general evaluations and specific comments on data consistency and comparability. Most federal agencies reported favorably on the directive, though expressing a desire for specific guidelines; none suggested it would be better off without a government-wide standard. Reactions to possible modifications were mixed, with some agencies indicating a large degree of flexibility, others expressing strong reservations, and some opposing alterations. The directive has become a strongly established standard for the collection and presentation of data on race and ethnicity for statistical administrative and compliance purposes. It is widely used and closely observed. The major exception to the standard has been the Census Bureau, which was granted an exemption allowing it to include an “other” (write-in) response to the race question in the 1980 and 1990 censuses · Only a few agencies, including the National Center for Health Statistics and the Equal Employment Opportunity Commission, have added finer classifications, as permitted by the directive. Those subcategories have been consistent (after reaggregation), as required, with the directive. The directive’s distinction between statistical and policy purposes is not clear-cut. Agencies reported that they could not distinguish “statistical” from “policy” purposes; they collect racial and ethnic data in adherence to federal policies, notably those on civil rights. Invariably, regulation and monitoring require some comparison with benchmark statistics, usually census data. The general degree of satisfaction with the current standard and wariness of change is evidenced in the following excerpts from the agencies’ survey responses: . . . the categories in Directive No. 15 are quite adequate in meeting the Commission’s need for racial and ethnic data. . . . the Commission does not recommend any changes or expansion in the categories set forth in Directive No. 15 at this time (U.S. Equal Employment Opportunity Commission). The Directive No. 15 categories have been adequate for civil rights enforcement purposes . . . (Office of Civil Rights, U.S. Department of Health and Human Services). The Directive 15 combined standard meets all OPM’s needs for race and ethnic reporting. . . . OPM opposes any changes in race/ethnic categories that either are not tied to specific program evaluation purposes or are not aggregable to the existing Directive 15 combined standard categories (U.S. Office of Personnel Management). Still, some agencies noted needs for more detailed data collection, such as the following:

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification . . . a category should be added to cover multi-racial beneficiaries . . . [This] is needed to give multi-racial/multi-ethnic families and individuals fair opportunity for identification (Food and Nutrition Service, U.S. Department of Agriculture). The costs and associated difficulties of changing classifications were of concern to the agencies. In addition to the immediate costs of implementing changes, there is a desire for historical comparability in the data. The overall sense was that while the need for statistical continuity should not be a barrier to needed change, it should serve as a strong deterrent to superficial change, and especially to untested change. Workshop participants also noted that statistical categories are often adapting to change that has already occurred, rather than to change that is anticipated. One reason is that statistical administration is usually conservative and cautious. Another is that there are conceptual and technical difficulties in capturing rapid change. Finally, statistical categories are developed in accordance with past and present federal policies and data needs, not to anticipated or possible future policies and needs. The agencies said that the directive has led to a high degree of consistency in the reporting of racial and ethnic data across the federal system and to comparability of data over time. Workshop participants noted, however, that the importance of temporal comparability is mitigated by the fact that the directive has been in place for a relatively short time, less than two decades. Agencies believe that the data are as consistent as could be expected with regard to comparability between those from observer identification and those from self-identification (although there is no evidence on this point). The patterns of consistency for category totals are all the more striking in the face of individual-level data, particularly for Hispanics, which show that individual reports vary from survey to survey and that individuals often self-classify themselves differently than would an observer (Rodriguez and Guzman, 1992; Rodriguez, 1974). Comparability of data across federal agencies is a key objective of Directive 15. Most federal agencies indicated that they use other data, primarily census data, but they did not provide details on comparability and limitations of such data sharing. Different data collection methods (e.g., personal interviews, administrative records) and different classifications schemes (the combined race and ethnicity format with five categories or the separate race and ethnicity formats) can yield different counts and inconsistencies, however.6 For an assessment of the current standards, one would have to do an inventory of the alternate methods for collecting data, consider the sources of error that may affect these different 6   Inconsistencies can also result from the use of different overall numbers for the denominators in calculating rates. This occurs because data are usually obtained from different sources (e.g., from vital statistics registration versus the census), resulting in inconsistencies within the numerator and denominator data (see Edmonston and Schultze, 1995:Appendix K).

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification methods, and identify the ways in which these factors affect the comparability of the data. Data on race and ethnicity from the decennial censuses are special in several ways. First, as noted above, race and ethnicity questions have been prominent in census data since the first census of 1790. Second, the Census Bureau not only collects data on many more racial and ethnic groups than the minimum mandated by Directive 15, but it also has an exemption that permits respondents to write in anything they wish, including “other.” Census data are unique, for instance, in having information for subgroups of Asians and Hispanics (e.g., Laotians, Colombians). Third, the Census Bureau, as the nation’s leading federal statistical agency, has more flexibility than other agencies to tabulate and present data for various needs in different combinations. Although discussions about census questions and data were prominent in the workshop, the census experience is not necessarily applicable to or appropriate for agencies with more limited uses for racial and ethnic data. The minimum federal standards meet the perceived data needs of many agencies. Nonfederal Experience A number of federal agencies require local government and industry reporting of race and ethnicity on the basis of Directive 15, for information about schools, state and local government hiring, and industrial employment practices. The federal standard has also influenced data collection standards of state and local agencies. States and localities often treat Directive 15 as a default standard, in part because of the political controversy involved in creating a new standard and in part because data in this format are already collected in state and local agencies for reporting to the federal government. Workshop participants included representatives from a variety of nonfederal settings, and they reported some of their experiences. Local school districts, since 1967, and postsecondary institutions, since 1970, have reported the race and ethnicity of students, teachers, and school district employees to the National Center for Education Statistics of the U.S. Department of Education. The data are used by the agency’s Office of Civil Rights for enforcing civil rights compliance, investigating complaints, and developing policy. Racial and ethnic designations of students are made either by parents, when enrolling their children, by the students themselves, or by teachers and district officials. It is not known what proportion of students are given a racial or ethnic designation in each of these ways. Workshop participants raised questions about reporting the classification of students who do not respond to inquiries about their racial or ethnic identity. The experience reported from a few districts is that they apparently face a choice between creating a new category, such as “other” or “undesignated,” that is not in compliance with Directive 15 or reallocating the nonresponses to the five basic

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification categories. In Fairfax County, Virginia, the district reports to the federal government using an “other” designation, with many footnotes. However, the state has insisted that reporting be limited to the five categories laid out in Directive 15, so the district has proportionately reallocated the “other” students to one of the five exising categories. The race and ethnicity categories that state and local governments use for their programs and policies depend largely on the area’s demographic and socioeconomic characteristics. These categories may be consistent with, but not identical to, Directive 15. In Hawaii, for example, the Health Surveillance Program, an ongoing survey conducted by the Department of Health, allows a wide range of statistical detail with two questions: “Of what race or combination of races is: Your father? Your mother?” However, most published tabulations are limited to 10 unmixed groups (Caucasian, Japanese, Chinese, Filipino, Hawaiian, Korean, Puerto Rican, Black, Samoan, and “other unmixed or unknown ”) and two mixed categories (part Hawaiian and mixed non-Hawaiian). The state of California collects data for Filipinos separately from other Asian Americans because of the historical socioeconomic disparity between Filipinos and other Asian American groups, especially Chinese and Japanese. In both cases, the categories can be reaggregated for federal purposes to the Directive 15 classification. Some federal legislation requires more detailed categories than those in Directive 15. The Disadvantaged Minority Health Improvement Act of 1990 specifically mandates collection of data by race and ethnicity categories. It requires the secretary of the U.S. Department of Heath and Human Services to encourage states to collect vital statistics for subpopulations that are significantly represented in a given state.7 According to workshop participants, private industry uses racial and ethnic data compiled according to Directive 15 to demonstrate compliance with equal employment opportunity laws. Moreover, racial and ethnic data, primarily those gathered in the decennial censuses are used as a demographic basis for marketing decisions. Private corporations may design their own survey instruments to gather data on the race and ethnicity of their employees for federal reporting. Such forms are usually designed to comply with the reporting procedures for the Equal Employment Opportunity Commission, but they sometimes diverge from the category format of Directive 15. For example, General Motors, in its federal contract compliance data questionnaire, specifies “not of Hispanic origin” for whites, but makes no reference to Hispanic origin for blacks. One workshop participant noted that changes to existing race and ethnicity 7   The National Center for Health Statistics (1995b) encourages states to obtain more detailed data on subpopulations of Hispanics, Asian Americans, and Pacific Islanders if those groups represent a significant part of a state’s population, and they are reimbursed for a share of the extra costs of doing so.

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification categories may impose a burdensome cost on employers, involving resurveying their employees and reentering the data, because racial and ethnic data are usually gathered only at the time of hiring. In addition, affirmative action plans for federal contractors are usually based on timetables for hiring from particular racial and ethnic groups, for which consistent use of categories over time is necessary to determine if hiring goals have been met. Most social scientists consider race and ethnicity as social constructs (see below) and thus point to the limitations of racial and ethnic classification; yet many still use the Directive 15 categories in their work. This use occurs primarily because the available data come from the federal statistical system, predominantly the decennial census and large federal surveys, which must be in compliance with the existing standards. Although researchers search for more accurate categories, they do not necessarily completely reject the existing categories. Rather, they may refine, contract, or expand them; use them selectively; or use them in conjunction with other data. Academic scientists and other data users, including community-based researchers and service providers, may incorporate such other indicators as language, nativity, and ancestry. Federally funded research is not required to follow the Directive 15 categories; researchers are free to develop data collection and coding schemes that best serve their own purposes. But the directive likely influences what race and ethnicity data researchers collect, as well as what race and ethnicity data they use (Entwisle and Astone, 1994; Hauser, 1994). This influence is not universal, however; according to the Office of Minority Health in the U.S. Department of Health and Human Services, many of the grantees and contractors of the National Institutes of Health do not collect race and ethnicity data that would comply with Directive 15. SOCIAL TRENDS THAT AFFECT RACE AND ETHNICITY The current debate about racial and ethnic classification has been fueled by at least two major factors: demographic and social changes in American society and changes in the identification and conception of race and ethnicity themselves. Growth, Concentration, and Heterogeneity of Minority Groups The racial and ethnic minorities designated in Directive 15 now account for about 25 percent of the U.S. population, and by 2040 they are projected to account for about 41 percent (Edmonston and Passel, 1994). The proportions of Hispanic and Asian populations are projected to grow significantly, while those of black and American Indian and Alaskan Native populations are projected to remain roughly the same. State and local experiences further highlight the diverse changes now under way in the nation’s minority populations. The concentration of minorities in

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification some states and metropolitan areas is particularly notable; nine metropolitan areas accounted for over one-half of the nation’s growth in minority populations between 1980 and 1990 (Frey, 1993). In the four largest metropolitan areas—New York, Los Angeles, Chicago and the San Francisco Bay area—the designated minorities make up more than one-third of the populations. In 1990, Hispanics were 94 percent of the population in Laredo, Texas; blacks were 43 percent in Jackson, Mississippi; and Asians were 63 percent in Honolulu, Hawaii. In many school districts, the majority of students are now from “minority ” groups. Immigrants of minority ethnicity and race are concentrated in only a few states. Using language as an indicator of the recency of immigration, more than one-fifth of the population in six states speaks a language other than English at home: Arizona, California, Hawaii, New Mexico, New York, and Texas (Bureau of the Census, 1992:xv).8 The great heterogeneity of recent immigrants is hidden behind the breadth of categories like Asian and Pacific Islander and Hispanic. Immigration from the Caribbean and Africa adds to the diversity of the black category. The 1990 census question on race listed 10 subcategories of Asian and Pacific Islanders (API)—Asian Indian, Chinese, Filipino, Guamanian, Hawaiian, Japanese, Korean, Samoan, Vietnamese, and “ other” API. The Hispanic question listed Cuban, Mexican, Puerto Rican, and “other.” Persons were further requested to specify “other” responses. Several dozen “other” API groups and more than a dozen “other” Hispanic groups were reported in significant numbers. Write-in responses to the race question required more than 300 codes, those for American Indian tribes about 600 codes, and those for Hispanic origin another 70 codes. The 1980 and 1990 censuses included an open-ended ancestry question separate from the race and ethnicity questions; more than 600 unique codes were used in 1990.9 Racial and ethnic intermarriage introduces further heterogeneity—and complication—to questions of racial and ethnic identification. In 1990, interracial unions (for example, those involving one partner who indicated race as black and one who indicated race as white) accounted for 1.5 million married or cohabiting couples (Bureau of the Census, 1990), and interethnic couples (in which one partner indicated Hispanic origin and the other did not) also accounted for slightly more than 1.5 million couples. (There is considerable overlap in the two groups, however, since Hispanics are also classified as either white or black.) There were much smaller numbers of couples for the other race and Hispanic combinations. 8   This high figure also reflects long-time residents who do not regard English as their first language, including a sizable number of Navajos and other American Indians and Spanish-speaking people in the Southwest. 9   The 1980 question was “What is this person’s ancestry?” The 1990 question was “What is this person’s ancestry or ethnic origin?”

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification In total, about 4 percent of the nation’s 52 million couples were racially or ethnically mixed, in terms of the five core categories of Directive 15. About 4 million children are now growing up in multiracial or multiethnic households—some 2 million children in each category (Rolark et al., 1994), and the absolute and relative number of children with such mixed backgrounds is likely to increase. This group may represent as many as 8 percent of all children, depending on the overlap of multiethnic and multiracial households. The growing number of multiracial and multiethnic children has contributed support to a movement for the introduction of multiracial and multiethnic categories in census and other data. Especially in school systems, some parents and children want the right to identify themselves as multiracial, multiethnic, or of multiple backgrounds. Some parents of multiracial children have argued that their children should not be forced to choose between the races of their parents. A multirace category has been added for collecting school enrollment data in Berkeley, California; Fulton County, Georgia; Cincinnati, Ohio; and Ithaca, New York. The states of Ohio and Illinois have also mandated a multirace category for school forms. The state of Georgia passed legislation in 1994 requiring all state agency and employment forms to include the classification, “multiracial” (Graham, 1994). Since marriages across racial and ethnic lines are likely to continue to increase, single-race and single-ethnicity classification promises to face even more political challenges. Racial and Ethnic Identity Issues of classification involve not only the many distinct groups with which people may identify, but also the restriction imposed by Directive 15 and most administrative agencies that persons shall have only one race and ethnicity. Two aspects of these issues are self-identification versus observer identification and the social construction of identity. Self-Identification Versus Observer Identification Workshop participants noted that, increasingly, self-identification has become the preferred means of statistical classification; individuals are acknowledged to be the best judge of their identity. The decennial census and most surveys ask respondents to self-identify their race and ethnicity.10 Of course, 10   Individuals do not always report their own race and ethnicity on surveys. For example, in the decennial census, the household respondent, who is usually an adult, reports the race and ethnicity for other household members. In this discussion, however, the issue is whether identification is made by the respondent or someone known to the respondent rather than by a survey taker or other third-party observer.

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification there are no objective standards on the “accuracy” of these reports; respondents are what they say they are. At the same time, several federal agencies noted that they rely on administrative forms that are filled out by observers. Observer identification occurs when employers fill out paperwork for their employees, school administrators write in responses for their students, and medical personnel write in responses for patients who may not be able to respond. The U.S. Equal Employment Opportunity Commission (EEOC) noted that most racial and ethnic classification is done by employers: an employee is “included in a particular group because that employee appears to belong to that group.” The practices of self-identification and observer identification are not split along federal agency lines or between program and statistical purposes. Some agencies use a combination of the two approaches in reporting, which may affect comparability. For instance, employer data on hiring may be compared with census data on characteristics of the local market. At the workshop, some enforcement agency personnel and legal experts dissented from the preference for self-identification. In cases of racial and ethnic discrimination, observer identification is often preferred: racial and ethnic discrimination may occur against a person regardless of how that person identifies herself or himself. Identity as a Social Construct Equally as important as the issue of self-identification is the increasing recognition that race and ethnicity are subjective, personal characteristics. The dominant perspective in social science now views race and ethnicity as social constructions, which develop over time as groups share common social and political experiences (Glazer and Moynihan, 1963, 1975; Hirschman, 1983; Thompson, 1989; and Nagel, 1994). From this perspective, race and ethnicity are dynamic and vary across groups and over time. This view stands in marked contrast to the widespread popular perspective that race is biologically determined and permanent and that ethnicity is culturally determined and equally permanent. 11 Although the social science perspective on race emphasizes its social construction, the process is rooted in historical experience. Three examples illustrate the changing nature of racial and ethnic classification. The first comes from how populations of multiple racial backgrounds are named and treated in U.S. history and in different societies. In the United States today, the norm is to label anyone 11   Biological scientists also find no evidence for fixed races as they are popularly perceived. Current biological studies point to a wide variety of genetic characteristics that are shared across supposedly different races as well as very different characteristics that are not shared by groups that are supposedly of the same race (Marks, 1994).

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification with black (African American) ancestry as black, but this was not always the case. Many nineteenth century Americans recognized a population produced from the mixture of Africans and Europeans, calling them mulattos, quadroons, and octoroons (Davis, 1991). These categories were used in the census of 1890, and the mulatto category continued in 1900, 1910, and 1920. In other societies, people of multiple racial backgrounds are often seen as a new group, distinct from the parental groups. Thus, there are Eurasians (those of mixed Asian and white ancestry) in Asia, Metis (those of mixed French and indigenous ancestry) in Canada, and mestizos (those with combinations of native Indian, black, and Hispanic ancestry) in Latin America. A second example of social construction is the phenomenon sociologists call “situational ethnicity”: an individual can choose when to assert a particular ethnic or racial identity or claim. For instance, a person may report a different race or ethnicity on a census questionnaire than on an employment application. In this way, one’s ethnic or racial affiliations are discretionary and voluntary. Clearly, whether race or ethnicity are pertinent to a particular social situation is closely affected by larger social and political forces (Paden, 1967; Okamura, 1981; Weinfeld, 1981; Waters, 1990:8, 16-17). A third example of the social construction of race and ethnicity can be found in the concept of “symbolic ethnicity,” seen yearly on such occasions as St. Patrick’s Day or the Chinese New Year. If members of an immigrant group assimilate over several generations, knowledge and practice of the primary ethnic culture fades. What is left is a sense of ethnicity that shows up by wearing green and marching in a New York City parade on St. Patrick’s Day or celebrating at a Dragon Dance on Chinese New Year day in Los Angeles. Such activities allow individuals to display symbolic ethnicity even though their ethnic affiliation may have little salience in other social settings. The attachment to specific ethnic identities may change over generations and may shift for an individual over his or her lifetime (Alba, 1990; Gans, 1979; and Waters, 1990). EXPERIENCES OF OTHER COUNTRIES The United States is not alone in grappling with the challenge of measuring a racially and ethnically diverse population. Other nations with indigenous populations or new immigrant groups are also examining ways to improve their standards for racial and ethnic classification. In April 1992 the Bureau of the Census and Statistics Canada convened an international conference on the measurement of ethnicity in six countries: Australia, Canada, Malaysia, the former Soviet Union, the United Kingdom, and the United States. The statistical agencies of these nations are faced with the challenging task of how best to enumerate an increasingly heterogeneous population. Academic researchers, policy analysts, statistical agency staff, and human service providers discussed current and future

OCR for page 4
Spotlight on Heterogeneity: The Federal Standards for Racial and Ethnic Classification issues in the field of measurement of ethnicity (see Statistics Canada and Bureau of the Census, 1993). The executive summary of the conference concluded that statistical agencies cannot meet all of the diverse data needs for racial and ethnic information—such as statistical purposes, program administration, business planning, research—because of limited space on questionnaires and resources for data coding and reporting (Statistics Canada and Bureau of the Census, 1993:5). The highest priorities have been for legal requirements, such as fulfilling Directive 15 in the United States. The needs of academic and other research groups, business, ethnic communities, and nonprofit groups for baseline and small-area data are also important but have been secondary to government needs. A dominant theme of the conference was that, despite some processes of assimilation, ethnicity continues to matter. The measurement of ethnicity is determined differently in each country, depending on how homogeneous or heterogeneous a country is, the state’s ideas of inclusion or exclusion, and who decides what the officially recognized ethnic groups are. Current national measures of ethnicity, to a large extent, are determined by social issues and public policies of each country. Future measurement of ethnicity will continue to be influenced by those factors, as well as by demographic changes and the results of testing programs and research. Participants noted that ethnicity is dynamic and may change over the lifetime of an individual. Thus, inconsistencies in data over time may be the result of this dynamic nature of racial and ethnic identity rather than flaws in data collection. Ethnicity is multidimensional and includes such aspects as race, origin, ancestry, identity, language, and religion. Fluidity and multiplicity are both difficult to measure with existing methods; research is needed on new methodologies or new approaches to obtain a broader array of data and different concepts and definitions.