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 140
Modernizing the U.S. Census 7 Data on Race and Ethnicity Historically, the decennial census has included questions on race and ethnicity, although the specific questions asked and the categories for tabulating the answers have changed every 10 years. These changes have occurred because of shifts in the racial and ethnic makeup of the population, changes in social attitudes and political concerns, and the evolving needs of the federal government for data. The growing racial and ethnic diversity of the American population, changing attitudes about race and ethnicity, and the increasing use of census data have converged to make census questions on race and ethnicity the focus of attention and controversy. Those questions now play a special role in debates over census content, census methods, and public cooperation. The Civil Rights Act of 1964 and the Voting Rights Act of 1965, with extensions, amendments, and court interpretations, have expanded the need for race and ethnicity data for all levels of geography, including individual blocks. These data are required for congressional and state election redistricting, for enforcement of federal, state, and local civil rights statutes, for allocation of funds and administration of programs at every level of government, and for many related purposes (see Appendix C). The decennial census is currently the primary source of race and ethnicity data to fulfill these many legal requirements. Without a change in the legal framework, the required race and ethnicity data need to be collected in the census. Coupled with survey information and data from administrative records, census data also satisfy other legislative and programmatic requirements for data on race and ethnicity. Improvements in the measurement of errors in census data have fostered awareness in recent decades of the differential net undercounting of racial and
OCR for page 141
Modernizing the U.S. Census ethnic minorities. Since the mid-1960s, this increased knowledge about the nation's policies and programs and their importance in these undercounts have contributed to increasing efforts and spending to improve the census. Data on race and ethnicity are involved, directly and indirectly, in most of the concern and contention about costs and quality of the 2000 census. Two difficult issues have arisen with respect to the need for data on race and ethnicity in the census (see Statistics Canada and the Bureau of the Census, 1993, for an international review of census data on ethnicity).1 One is the growing public pressure for revising and expanding race and ethnicity classifications in the census as the nation becomes more diverse. The other is increasing recognition of the fluidity and accompanying ambiguity of racial and ethnic identities for many people. These issues have been brought to the fore with the reliance of the census on self-identification of race and ethnicity, which began in the 1970 census. Self-identification has been viewed as an improvement over the historically discriminatory methods of designation, such as blood quantum and observer identification. These older methods carried the presumptions that a certain amount of ancestry determined a person's racial or ethnic identity, that an observer could identify a person's race and ethnicity, that only one identification per respondent is appropriate, and that a person has the same identity in all situations in his or her lifetime. These presumptions are increasingly problematic with self-identification, especially at a time of growing heterogeneity of the American population, due not only to more diverse immigration but also to increased interracial and interethnic marriages. Moreover, self-identification in the census is not the same as in other surveys or on various types of forms. In the census, it is typically one member of the household who identifies his or her own race and the race of all other members of the household. The design of future censuses must take account of these issues. The panel believes that underlying these important issues is a much more basic concern regarding the role of race and ethnicity data. In particular, there are inconsistencies between race and ethnicity categories required for legislative and administrative purposes and the census categories of race and ethnicity, which were designed for statistical purposes. Courts have already begun to litigate the classifications because of the discrepancy between the different conceptual approaches of the law and statistics. The legal approach views individuals as potential members of protected classes, as defined by statutes and judicial decisions. The statistical approach reflects an effort to provide a comprehensive demographic profile with data that may extend beyond legal considerations. The two approaches must be reconciled before the technical questions of questionnaire design, question placement (short or long form), and labeling of groups can be solved.
OCR for page 142
Modernizing the U.S. Census HISTORICAL CONTEXT For over 200 years there have been both change and continuity in race and ethnicity data in the United States. The number of items related to race and ethnicity has increased and contracted over time as new populations have entered the United States or as selected populations have become important for policy purposes, such as a change in civil status. Table 7.1 displays the number and types of categories used in the race item of the census from 1850 to 1990. (For a discussion of the categories, see Mann, 1994.) An item on race has been included in every census since 1790, when the population was distinguished for apportionment purposes as the sum of free persons, Indians not taxed, and three-fifths of other persons. Article 1, Section 2, of the Constitution mandated the three-way race classification: within the context of race, free persons were understood to be whites; the persons counted as three-fifths were slaves and synonymous with blacks. This use of race data continued in censuses through the Civil War. The practice of the "one-drop rule"—which defined a person as black who had even one black ancestor—served to reinforce the identity of a person with any nonwhite ancestry as nonwhite. The entry of new populations has motivated items related to ethnicity and national origin, which have been included in each census since 1850.2 The questions first pertained to immigrants of European origins who were distinguished from the native-born white population. They were identified by several items related to nativity, parental birthplace, immigration, citizenship status, language status, and time of arrival in the United States. In contrast, immigrants of Asian origins were categorized as "aliens ineligible for citizenship," and they were counted in the race item as Chinese (beginning in 1870) and Japanese (beginning in 1890). By the 1890 census, the race item had expanded to eight categories: white, black, mulatto, quadroon, octoroon, Chinese, Japanese, and (American) Indian. The categories were a combination of color, tribal status, and Asian national origin. There was no single ethnicity item: in addition to previously asked items on birthplace, parents' birthplace, and citizenship status, the 1890 census included two language items: ability to speak English and mother tongue. It also included an item on number of years in the United States. Classification by race and ethnicity became more complex in the twentieth century as more items were used to identify residents who enjoyed fewer rights than the native-born white citizenry. The race question continued to be characterized by both changing and continuing categories of color, tribal status, and Asian origins. Additional Asian categories were included to record the increased immigration from Asian countries that resulted from implementation of the 1965 Immigration and Nationality Act (which abolished the national origins system of the National Origins Act of 1924, as well as the Asiatic barred zone of the Immigration Act of 1917). Also during the twentieth century, several categories
OCR for page 143
Modernizing the U.S. Census have been eliminated. For example, the Mexican category was used only once, in the race item of 1930, although it has resurfaced in the Hispanic origin item since 1970. Mixed-race categories, such as mulatto and part Hawaiian, were used for only brief periods, and none is presently used in the census. Items related to ethnicity have also increased and decreased over time, and the exact wording of items has varied. Items have continued to revolve primarily around immigration, nativity, and origins. The 1920 census (which preceded the National Origins Act of 1924, which effectively restricted immigration from Europe) included 10 items related to ethnicity. In the post-1965 period, an important change in ethnic classification was the use of self-identification items to obtain data on ethnicity. This began with the 1970 census, when the Hispanic origin item was introduced. In the 1980 census, an open-ended ancestry item for all groups was included in the long form (McKenney and Cresce, 1993). The race, Hispanic origin, and related ethnicity items on the census since 1970 are shown in Table 7.2 for the short form and Table 7.3 for the long form. These items are the result of changing policy and programmatic needs for race and ethnicity data and special research programs of the Census Bureau to test and improve the enumeration of racial minorities and persons of Hispanic origin in a mail-based census. The Spanish/Hispanic-origin item was first included in the 1970 census, and it was tested and refined in subsequent censuses (see Table 7.3). As the tables show, since 1970, there has been a proliferation of categories or subcategories to the question on race. The write-in items in the 1990 census elicited more than 300 race responses, approximately 600 American Indian tribes, 70 different Hispanic-origin groups, and more than 600 ancestry groups. In addition to reporting identification with a specific racial group, respondents could indicate an "other" race identification in two ways: (1) by checking a circle on the questionnaire to indicate an "other" race, or (2) by writing a race in a box for "other races.'' Written "other race" responses numbered more than 9.8 million in 1990, including respondents who marked only the "other race" circle without specifying a race; of these, over 97 percent also indicated Hispanic origin. There were approximately 2.5 million write-ins to the "other race" category identified during the early automated coding and editing of the race item. In all, 41 percent were reclassified to one of the major race categories (white, black, American Indian/Alaskan Native, or Asian/Pacific Islander) during this automated process and thus were not included in the final counts for "other race." CURRENT REQUIREMENTS Foremost among the legal requirements for race and ethnicity data is the Voting Rights Act of 1965. As initially passed and subsequently amended (in 1970, 1975, 1982 and 1992), the Voting Rights Act prohibits states and political
OCR for page 144
Modernizing the U.S. Census TABLE 7.1 Census Race Categories, 1850-1990 Year White Black/Negro Native Peoples Chinese Japanese Other Asian or Pacific Islander Other 1850a Black, mulatto 1860a Black, mulatto Indianb 1870 White Black, mulatto Indian Chinese 1880 White Black, mulatto Indian Chinese 1890 White Black, mulatto, quadroon, octoroon Indian Chinese Japanese 1900 White Black Indian Chinese Japanese 1910 White Black, mulatto Indian Chinese Japanese Other (+ write in) 1920 White Black, mulatto Indian Chinese Japanese Filipino, Hindu, Korean Other (+ write in) 1930c White Negro Indian Chinese Japanese Filipino, Hindu, Korean Other races, spell out in full) 1940 White Negro Indian Chinese Japanese Filipino, Hindu, Korean (Other races, spell out in full) 1950 White Negro American Indian Chinese Japanese Filipino (Other race-spell out) 1960 White Negro American Indian Chinese Japanese Filipino, Hawaiian, part Hawaiian, etc. 1970d White Negro or black Indian (American) Chinese Japanese Filipino, Hawaiian, Korean Other (print race) 1980 White Black or Negro Indian (American), Eskimo, Aleut Chinese Japanese Filipino, Korean, Vietnamese, Asian Indian, Hawaiian, Guamanian, Samoan Other (specify) 1990 White Black or Negro Indian (American), Eskimo, Aleut Chinese Japanese Filipino, Hawaiian, Korean, Vietnamese, Asian Indian, Samoan, Guamanian, other Asian or Pacific Islander Other race
OCR for page 145
Modernizing the U.S. Census a In 1850 and 1860, free persons were enumerated on the form for "free inhabitants"; slaves were enumerated on the form designated for "slave inhabitants." For the free schedule, the instructions told the enumerators: "In all cases where the person is white leave the space blank in the column marked 'Color.''' For the slave schedule, the listed categories were black (B) or mulatto (M). b Although this category was not listed on the census form, the instructions read: 5. Indians.—Indians not taxed are not to be enumerated. The families of Indians who have renounced tribal rule, and who under State or Territorial laws exercise the rights of citizens, are to be enumerated. In all such cases write "Ind." opposite their names, in column 6, under heading "Color." 9. Color.—Under heading 6, entitled "Color," in all cases where the person is white leave the space blank; in all cases where the person is black without admixture insert the letter "B;" if a mulatto, or of mixed blood, write "M;" if an Indian, write "Ind." It is very desirable to have these directions carefully observed. c In 1930, the census questionnaires included "Mexican" as a race category. d In 1970, on questionnaires used in Alaska, the categories "Aleut" and "Eskimo" were substituted for "Hawaiian" and "Korean."
OCR for page 146
Modernizing the U.S. Census TABLE 7.2 Race and Ethnicity Questions on the Short-Form Census Questionnaire, 1970, 1980, and 1990 1970 COLOR OR RACE 1980 IS THIS PERSON 1990 RACE White White White Negro or Black Black or Negro Black or Negro Indian (Amer): print tribe Japanese Indian (Amer): print tribe Japanese Chinese Eskimo Chinese Filipino Aleut Filipino Korean Asian or Pacific Islander: Hawaiian Vietnamese Chinese Korean Indian (Amer): print tribe Filipino Other: print race Asian Indian Hawaiian Hawaiian Korean Guamanian Vietnamese Samoan Japanese Eskimo Asian Indian Aleut Samoan Other: specify Guamanian Other Asian or Pacific Islander: print race Other race: print race IS THIS PERSON OF SPANISH/HISPANIC ORIGIN OR DESCENT? IS THIS PERSON OF SPANISH/HISPANIC ORIGIN OR DESCENT? No (not Spanish/Hispanic) No (not Spanish/Hispanic) Yes, Mexican, Mexican-Amer., Chicano Yes, Mexican, Mexican-Am., Chicano Yes, Puerto Rican Yes, Puerto Rican Yes, Cuban Yes, Cuban Yes, Other Spanish/Hispanic Yes, Other Spanish/Hispanic print one group NOTE: There were no Spanish/Hispanic-origin questions on the 1970 short form.
OCR for page 147
Modernizing the U.S. Census TABLE 7.3 Ethnicity and Related Questions on the Long-Form Census Questionnaire, 1970, 1980, and 1990 1970 1980 1990 15 percent sample In what state or foreign country was this person born? Is this person a naturalized citizen of the United States? When did this person come to the United States to stay? Does this person speak a language other than English at home? What is this language? How well does this person speak English? What is this person's ancestry? In what U.S. state or foreign country was this person born? Is this person a citizen of the United States? When did this person come to the United States to stay? Does this person speak a language other than English at home? What is this language? How well does this person speak English? What is this person's ancestry or ethnic origin? Where was this person born? What country was his father born in? What country was his mother born in? What language, other than English, was spoken in this person's home when he was a child? 5 percent sample Where was this person born? Is this person's origin or descent Mexican; Puerto Rican; Cuban; Central or South American, Other Spanish; No, none of these Is this person naturalized? When did he come to the United States to stay?
OCR for page 148
Modernizing the U.S. Census subdivisions from imposing or applying any "standard, practice or procedure" for voting that will result in the abridgment of the right to vote "on account of race or color" or because a citizen "is a member of a language minority." Congress passed the legislation in 1965 to implement the guarantees of rights—particularly voting rights—in the Fourteenth and Fifteenth amendments to the Constitution. The definition of "race or color" was left implicit in the law; in practice, it meant "black'' and "white" or "white" and "nonwhite." Congress added the coverage of "language minorities" in 1975 and specifically defined them as "persons who are American Indian, Asian American, Alaskan Native, or of Spanish heritage." Redistricting plans that "dilute" the votes of the "members of the protected class" come under the scrutiny of the law. In 1980 and 1990, therefore, state legislatures prepared redistricting plans that would not "dilute" the votes of minorities. There has been extensive litigation on these issues. The Census Bureau has provided the population tabulations at the block level for the redistricting plans and their challengers. However, the Census Bureau collects these data under the authority of P.L. 94-171 (passed in December 1975), which does not specifically mention race and ethnicity tabulations; it was designed to provide timely total population counts for small areas to states and localities for redistricting purposes. The addition of the race and Hispanic-origin tabulations from the short form in 1980 was an administrative decision, in response to requests by the Department of Justice and state and local governments. The categories reported are those promulgated in the federal standard for race and ethnicity data, OMB Statistical Directive 15 (1977), which included four race groups: black, white, American Indian or Alaskan Native, and Asian or Pacific Islander; and one ethnic group, Hispanic origin. In 1990, race and Hispanic-origin data were cross-tabulated in the P.L. 94-171 file, and voting-age data were also reported. Barring changes to the Voting Rights Act, the 2000 census will be required to provide race and ethnicity data for small geographic areas for the redistricting process. There are two challenges facing the Census Bureau in providing such data. The first is guaranteeing accurate and precise coverage of the minority and majority populations. Since the differential undercount of racial minorities is well known, the census without correction is itself a flawed standard and cannot provide accurate small-area data. The differential undercount has been the basis of lawsuits against the Census Bureau since 1970. The Department of Commerce has defended the Census Bureau against such lawsuits by noting that before 1990 there was no acceptable or tested method of distributing the known national undercount to small areas. Even in 1990, when an extensive evaluation was conducted to be able to adjust for the differential undercount, the accuracy of the adjustments for small areas was questioned. In planning for the 2000 census, one of the major goals of the Census Bureau is to reduce the differential undercount. The Census Bureau plans to accomplish this through a series of fundamental changes in the way the census is
OCR for page 149
Modernizing the U.S. Census conducted. One of these changes is to target potentially hard-to-count areas in advance so that special enumeration procedures will be used for the census there. Another change is the concept of an integrated one-number census, in which statistical estimation for missed persons will be directly incorporated in the census counts that are officially released. With such changes, it may be possible for Census Bureau to make inroads in correcting for the known differential undercount in the P.L. 94-171 redistricting file. The second challenge the Census Bureau faces in providing race and ethnicity data for small geographic areas is the classification of various groups, which is currently mandated by OMB Directive 15, "Race and Ethnic Standards for Federal Statistics and Administrative Reporting," issued in 1977. Prior to that directive, federal agencies used their own categorization policies, leading to a recognition by several agencies of the need for a uniform set of race and ethnicity categories. The OMB directive requires that race and ethnicity data, for both statistical and administrative purposes, be collected for a minimum set of categories: American Indian or Alaskan Native, Asian or Pacific Islander, black, Hispanic, and white. The directive explicitly allows the collection of additional detailed race and ethnicity categories, so long as they can be aggregated into the basic categories. Agencies should not, however, combine data from one major category with data from any of the other major categories. Individuals are not permitted multiple responses; however, a special exemption is made for the census to collect an "other" response. Directive 15 has remained unchanged since it took effect in 1977. In 1993, the House Subcommittee on Census, Statistics, and Postal Personnel, chaired by Congressman Thomas C. Sawyer, held four informational hearings on Directive 15. At the July 1993 hearing, the Office of Management and Budget testified that the current administration planned to reconsider revision of the directive. Directive 15 sets the framework for race and ethnicity data collected in the census, although the census has collected more detailed categories in the 1980 and 1990 censuses than required. Revisions to Directive 15 in the next few years would affect the 2000 census, possibly requiring additional major race and ethnicity categories, changing the specific groups within the major categories, and requiring the use of multiple responses. In recent years, social scientists, some advocacy groups, and other data users have questioned the conceptual foundations behind the basic categories reported in OMB Statistical Directive 15 and hence the P.L. 94-171 tabulation. Advocacy groups have raised questions about including separate or new classifications for persons of multiracial parentage, for Native Americans, and for ethnic groups such as Arabs and people from the Middle East. The Census Bureau has also been asked by data users and service providers to furnish more data on smaller racial and ethnic groups for small geographical areas for program planning and evaluation, health studies, and fund allocation. To date, few if any
OCR for page 150
Modernizing the U.S. Census challenges have questioned the use of the Statistical Directive 15 categories for the P.L. 94-171 files, although this might be a future implication of the issues raised about the directive. Thus far, the Census Bureau response has been to list subdivisions of the Statistical Directive 15 major categories or to accommodate other groups in three "other" write-in categories in the census: other race, other Asian and Pacific Islander, and other Spanish/Hispanic. In response, OMB has undertaken a major reconsideration of Directive 15 in consultation with a wide range of federal and nonfederal data users. Census Bureau evaluation of the quality of race and ethnicity data suggests continuing classification problems (see Appendix K; McKenney et al., 1993; Cresce et al., 1992). For example, the Hispanic-origin item continues to have the highest allocation rate among short-form items. About 40 percent of Hispanics identified themselves as "other race" in the 1990 census question on race, which suggests that many respondents perceive "Hispanic" as a race rather than an ethnicity category. Moreover, as the Hispanic population has grown and as Hispanics increasingly identify themselves as "other race," this category has become larger than some "other race" categories. Self-identification for race, Hispanic origin, and ancestry questions means that responses are based on self-perception and therefore are subjective, but at the same time, by definition, whatever response is recorded is an accurate response. Yet self-identification for many people is based on context; people have multiple identities that may differ by context. The political context for the current census categories—whether there are positive, neutral, or negative effects to identifying with a certain group—is only one context for identification. Race is a more salient concept for historically disadvantaged groups, particularly descendants of African American slaves. First-generation immigrants have tended to identify with their country of origin, though their children, in the second generation, have more complex patterns of identity. Even among whites, flux in identification is noted by inconsistency in reporting ancestry at different points in time. Waters (1990) suggests that many people have a range of choices and options in how they choose to identify. The increased number of people identified as American Indian and their high intermarriage rates, as well as the substantial intermarriage rates for several Asian and Pacific Islander and Hispanic populations, suggest that racial and ethnic identity is becoming more complex and may shift for a person over time as well as in social context. Racial and ethnic identity is likely to grow more complex as the pool of people with identity options increases. The high reporting of "European" as a first ancestry by American Indians and several Asian and Pacific Islanders groups may already reflect attempts to report on multiracial and multiethnic identities. The purposes of collecting race and ethnicity data provide another context for self-identification. A person may respond differently to the same wording of questions, for example, when asked in the census and when asked in a job application form. Ivan Fellegi of Statistics Canada (private communication) provides
OCR for page 151
Modernizing the U.S. Census an example of this sensitivity to racial self-identification. Self-reports of race were obtained from the Canadian Public Service Commission, the federal government's central personnel management agency, and compared with reports to the 1991 Census of Canada and to an employee survey. With results directly compared, individuals appear to make rather different reports of their racial status in the nonthreatening, anonymous context of the census: in the Canadian data, individuals were more likely to report minority racial status in the census than in the other two information systems. Questionnaire design also can provide the opportunity for variations of self-reporting. It is clear that the actual wording of race and ethnicity questions can affect responses. For example, having an explicit box to check for Mexican origin self-identification results in more persons who classify themselves as Mexican than does an open-ended question to which persons can write in the classification "Mexican."3 In the census, the race, Hispanic origin, and ancestry items may be viewed as variations of the same or similar concept, the overlap of concepts, or as different but related concepts depending on sequence, wording, and instructions. People who view them as the same or similar concepts provide consistent answers (or respond to one item and omit the others as redundant). In contrast, people who view the items as different may provide multiple responses. On the basis of 1980 and 1990 census data, researchers have found that combinations of the race, Hispanic origin, and ancestry items yield different results in a person's self-identity, including: (1) people with the same racial and ancestry identity (e.g., white, German), (2) people with a race different from ancestry (e.g., African American, Irish), (3) people with one race and multiple ancestries (e.g., white, English, Russian), and (4) people with multiple racial identities (e.g., white, Vietnamese). As we noted above, the expansion of categories and of different combinations by which people self-identify is a response to immigration and interracial and interethnic unions, as well as changes in social attitudes, political pressures, and the evolving needs of the federal government for race and ethnicity data. This proliferation increases the possibility of fluidity and inconsistency of race and ethnicity reporting over time, as people may identify differently at different times or for different purposes. FUTURE REQUIREMENTS The problems of classification are likely to get much more complex as the ethnic and racial character of the American people changes. First, the official categories are themselves a product of the legislative debates of the 1970s, before the waves of immigration of recent years and the emergence of significant interracial marriage patterns and the resulting new generation of children with identities in two or more groups. Current public debate, as manifested in congressional and OMB hearings on Statistical Directive 15, has focused on the proliferation of new and finer classifications. There are, however, statistical and
OCR for page 152
Modernizing the U.S. Census operational concerns that caution against large expansion of the number of categories. For new classifications, the completeness of coverage and accuracy of the national counts will be unknown, making it difficult to develop measures of undercount. The reliability of these counts from finer classification, especially for small geographical areas, will be questionable. Furthermore, small groups are not usually evenly distributed geographically but are quite concentrated, raising questions not just of reliability but of confidentiality and privacy (i.e., Census Bureau tabulation guidelines require that there is a minimum number of people in tabulated cells to protect the confidentiality of respondents). Operationally, the costs and burdens of collection, processing, and presentation of data about many ethnic and racial groups are significant. Although the expansion of race and ethnicity classifications is a recognition of the complexity of the racial and ethnic composition of the American population, the proliferation of categories may not be a practical solution. More broadly, a close reading of the language of the Voting Rights Act and OMB Statistical Directive 15 reveals a conceptual confusion in categorization. The 1975 amendment to the Voting Rights Act defines "persons who are American Indian, Asian American, Alaskan Native or of Spanish heritage" as "language minorities." Directive 15 defines American Indians and Asian and Pacific Islanders as "racial" groups, regardless of language facility in English. The census questionnaire is in accord with Directive 15 and considers American Indians and Asian and Pacific Islanders, along with blacks and whites, as racial groups; it further includes tribal and subgroup delineations of these smaller populations. Similarly, the question on Spanish or Hispanic origin defines an "ethnic" identity. It does not measure language ability. Federal courts have begun to consider these issues in voting rights cases. In Garza v. Los Angeles County Board of Supervisors, expert witnesses testified that not all those who identified themselves as of Spanish origin in the 1990 census should be regarded as falling under the covered rubric of ''persons of Spanish heritage," that is, a language minority (Grofman, 1992:218). Thus far, the courts have decided that self-identification places one in the group, so that "a response of Spanish origin on the census questionnaire" serves as the "defining characteristic of the covered population" (Grofman, 1992:218). The test of inclusion as a race or ethnicity classification would be an actual finding of a pattern of discrimination against a member of a protected class. However, since the current census question on race is not a question on language minority, but rather on racial or ethnic self-identity, future court challenges seem likely. Whatever the conceptual foundations should be, the census supports self-identification, implying that the census question must appear to be sensible and self-explanatory for respondents in terms of their own situation. For many Americans, multiple categories are a reality, irrespective of possible conceptual or legislative advantages for a univariate classification. It is probable that challenges to the current federal standards on race and ethnicity classification will become more serious in years to come and that
OCR for page 153
Modernizing the U.S. Census the courts will challenge the Census Bureau to validate the conceptual foundations of race and ethnicity classifications. CONCLUSIONS AND RECOMMENDATIONS Because OMB Statistical Directive 15 is currently being reviewed, it seems prudent to continue research efforts to clarify the conceptual foundation of the classifications and to alert Congress to the conceptual confusion embedded in existing law and administrative procedures. As federal agencies and other data users have documented, the programmatic and policy considerations of the existing categories are important. What is problematic is being tied to these categories rigidly, especially as they may not fit current and future populations. Conclusion 7.1 The panel concludes that there are inherent ambiguities to racial and ethnic status. We recognize that reporting of this status in the census has multiple meanings for respondents and that the context of reporting affects how people self-identify. Conclusion 7.2 The panel concludes that the census questionnaire must be developed as a well-tested means for obtaining nationally useful data on race and ethnicity to meet constitutional and legally mandated federal requirements and for other legislative and program needs and informational purposes. The growing diversity of the U.S. population, changes in the way people identify themselves, and variations in definitions of members of protected classes raise the fundamental question of whether current classifications are appropriate for the country's changing society and future legislative and other policy requirements. A research program on race and ethnicity for the 2000 census, which would now be starting very late in the planning cycle, can only begin to address some of these issues. A more extensive program of research on how persons identify themselves in terms of race and ethnicity is needed to promote meaningful classifications in a changing society. Such a program would include the extension of current, but limited, Census Bureau initiatives on cognitive research; testing of various combinations of the race, Hispanic-origin, and ancestry items; and testing for open-ended and multiple identifications. Allowing respondents to identify with more than one racial group, to report as multiracial, and to treat Hispanic origin as equivalent to a racial group may be more consistent with some respondents' self-perceptions, although it may be less consistent with other respondents' self-perceptions. Only systematic research can develop useful information on such issues. Recommendation 7.1 The panel recommends that the Census Bureau
OCR for page 154
Modernizing the U.S. Census expand its examination and testing of race and ethnicity questions to provide comprehensive information on: (1) public understanding of the concepts and acceptability of questions, (2) compatibility among the several census items and the utility of cross-tabulations, (3) the comparability of census data to race and ethnicity data collected in other federal surveys or obtained from administrative records, and (4) the quality of data for small areas and specific groups. This research needs to be given a high priority so that the results may be incorporated into the review of Statistical Directive 15 currently being conducted by the Office of Management and Budget. Recommendation 7.2 The panel recommends that the Office of Management and Budget issue a revision of Statistical Directive 15 sufficiently early to provide adequate time for planning and testing for the 2000 census and coordinated implementation of changes by all affected agencies. Needed planning activities include the review and redesign of respondent-friendly forms, with special attention to the items on race and ethnicity. In addition, because of the complexities involved, any proposed changes to these items require consultation with affected groups and data users. Such consultation must occur prior to extensive tests as part of the planning cycle. Development of promotion and outreach programs for field tests also needs to be carried out at this time, along with appropriate congressional review. Any proposed changes of race and ethnicity items will have implications for race and ethnicity data needed for statutory, legislative, and programmatic purposes that should be taken into consideration. These include drawing political boundaries, funding allocations for various geographical levels, as well as the quality of the data for small geographical areas and for such relatively small racial and ethnic groups as American Indians. NOTES 1 Both Canada and the United States collect race and ethnicity data in their censuses. About half of the countries of the world collect such data in their censuses. 2 Although the 1820 census contained a checkoff box or line pertaining to foreign nationals, the use of that information is not clear; 1850 is the census date usually cited by both historians and social scientists for the beginning of ethnic items in the census. 3 In 1990, the check-box resulted in more people responding to the Spanish/Hispanic item and was significant for the Mexican-origin category only. This
OCR for page 155
Modernizing the U.S. Census relationship between check-box and write-ins did not hold for the write-ins to the Asian and Pacific Islander category in the race item: pre-1990 testing showed that more people provided Asian and Pacific Islander write-ins than reported in the check-boxes for the Asian and Pacific Islander groups.
Representative terms from entire chapter: