Mr. Chairman and Members of the Subcommittee:
I am pleased to be here today to discuss the reporting of race and
ethnicity data. Our testimony today focuses on two issues you asked
us to address: (1) our prior work on the collection and reporting of
race and ethnicity data by the Bureau of the Census for the decennial
census, as well as by other federal agencies; and (2) state reporting
of race and ethnicity data to federal agencies for health and
educational purposes. My comments are based on our prior work in
this area, our current monitoring of plans and preparations for the
2000 Decennial Census, and limited data collection we did in selected
states in April 1997 in response to your request.
Over the years, our work has shown that the collection of these types
of data is technically complex and publicly controversial. It is
technically complex because race and ethnicity are not objectively
definable characteristics, making measurement difficult. Also, in
many instances, a person self-identifies his or her own race and
ethnicity; in other instances another party may categorize that
person's race and ethnic designation by observation, which can
produce inconsistent results. In addition, the manner in which
different organizations may ask for racial or ethnic information, as
well as how this information is compiled or aggregated, can lead to
inconsistent results. Measurement of race and ethnicity is also
controversial because some individuals have strong feelings about how
they are classified and are uncomfortable, when presented with a list
of classifications, if a particular "category" is not available for
them to select. For example, some people who are multiracial want to
be able to reflect this heritage by designating themselves as such;
however, they may not be provided this choice. Alternatively, some
people may oppose the use of a multiracial category because it could
result in a reduction in the number of individuals classified in
their racial category, and they view this as potentially reducing any
benefits this particular group may receive. Some state and federal
program or administrative officials raise concerns about a
multiracial category because it may (1) add costs from the need to
change forms and computer software, (2) not provide any analytical
benefits, or (3) result in reporting inconsistencies and impede
analyses of trends.
BACKGROUND
-------------------------------------------------- Chapter STATEMENT:1
The United States government has long collected statistics on race
and ethnicity. Such data have been used to study changes in the
social, demographic, health, and economic characteristics of various
groups in the population. Federal data collections, through
censuses, surveys, and administrative records, have provided a
historical record of the nation's population diversity.
Since the 1960s, data on race and ethnicity have been used
extensively in civil rights monitoring and enforcement covering such
areas as employment, voting rights, housing and mortgage lending,
health care services, and educational opportunities. Legislatively
based requirements in these areas created the need among federal
agencies for compatible, nonduplicative data for the specific
population groups that historically had suffered discrimination and
differential treatment on the basis of their race or ethnicity. We
have attached a listing of some of the statutes that require the
collection and reporting of racial or ethnic data.
In the mid-1970s, the Office of Management and Budget (OMB), in
conjunction with several federal agencies, undertook a collaborative
effort to standardize racial and ethnic data collected and published
by federal agencies. The result of this effort was OMB's 1977
publication of the "Race and Ethnic Standards for Federal Statistics
and Administrative Reporting" contained in Statistical Policy
Directive No. 15. These standards also implemented the requirements
of Public Law 94-311 of June 16, 1976, which called for the
collection, analysis, and publication of economic and social
statistics on persons of Spanish origin or descent. Directive 15,
which has not been changed since it was initially published, provides
standard classifications for recordkeeping, collection, and
presentation of data on race and ethnicity in federal program
administrative reporting and statistical activities. These
classifications include four races--American Indian or Alaskan
Native, Asian or Pacific Islander, Black, and White; and one
ethnicity--Hispanic origin or not of Hispanic origin.
The standard collection categories are to be used for (1) civil
rights compliance reporting by both the public and private sectors
and all levels of government; (2) new and revised general program
administrative and grant reporting by federal agencies; and (3)
statistical reporting by federal agencies. According to OMB's Chief
Statistician, even though states are not directed to follow OMB's
guidance except when reporting to the federal government, in practice
they generally do. The Directive states that its purpose is not to
limit the collection of data to the five categories--four racial and
one ethnic. However, any required reporting that uses more detail
must be organized in such a way that the additional categories can be
aggregated into the basic racial/ethnic categories. One notable
exception to the standard has been the Bureau of the Census, which
was granted an exemption allowing it to include an "other" (write-in)
response to the race question in the 1980 and 1990 Censuses.
During the past several years, the standards have come under
increasing criticism from those who believe that the minimum
categories set forth in Directive No. 15 do not reflect the
increasing diversity of our nation's population. Some have also
proposed changing the names of some categories. Because of these
criticisms, OMB is in the process of determining whether Directive
No. 15 should be modified. Among the changes being considered are
whether a multiracial category should be included; and whether "race/
ethnicity" should be asked as a single identification, or whether
"race" identification should be separate from Hispanic origin.
COLLECTION OF RACE AND
ETHNICITY DATA BY THE BUREAU OF
THE CENSUS AND OTHER FEDERAL
AGENCIES
-------------------------------------------------- Chapter STATEMENT:2
Even if federal classifications for race and ethnicity were agreed
upon, it is extremely difficult to obtain accurate and consistent
data on the number of individuals within each classification. Our
prior work on the Bureau of the Census' collection of race and
ethnicity data during the 1990 Census indicates that the question or
questions used to elicit information on an individual's race and
ethnicity will affect how an individual classifies himself or
herself. Consequently, if the results of different surveys using
different questions or a different series of questions to obtain race
and ethnic data from the same population were compared, it would be
likely that different proportions of respondents would be classified
as White, Black, Asian or Pacific Islander, or American Indian or
Alaskan Native; or of Hispanic origin. This work also demonstrated
that external pressures on the Bureau can result in the use of
questions that may not provide as accurate a count of racial and
ethnic groups as would be possible if other, alternative questions
were used. Our prior work on federal agencies' use of race and
ethnic definitions found that other factors can also affect the
quality of racial and ethnic data collected, including the fact that
all states may not use the same classifications as the federal
government, and the fact that some racial and ethnic data are
obtained by observer-identification rather than by self-reporting.
1990 DECENNIAL CENSUS
EXPERIENCE WITH COLLECTING
RACE AND ETHNIC DATA
------------------------------------------------ Chapter STATEMENT:2.1
In 1993, we reported and testified that experience from the 1990
Decennial Census provided valuable lessons for future censuses.\1 One
lesson was the need to develop a consensus on the race and ethnicity
questions as early in the decade as possible. Another was the need
for the Bureau of the Census to continue efforts to improve race and
ethnic data quality to ensure that the quality of data collected is
acceptable.
--------------------
\1 Federal Data Collection: Measuring Race and Ethnicity Is Complex
and Controversial (GAO/T-GGD-93-21, April 14, 1993) ; and Census
Reform: Early Outreach and Decisions Needed on Race and Ethnic
Questions (GAO/GGD-93-36, Jan. 28, 1993).
CONSENSUS NOT ACHIEVED
---------------------------------------------- Chapter STATEMENT:2.1.1
In the 1990 Census, the Bureau was unable to build a consensus on its
recommended version of how Asian and Pacific Islanders were to be
represented in the race question despite implementation of a special
testing and consultation program. As a result, the final format of
the race question was decided late in the decade after protracted
debate and was contrary to the Bureau's initial recommendations.
The Bureau, after testing alternative versions of the Asian and
Pacific Islander questions during the 1986 Los Angeles, Mississippi,
and National Content tests, among other tests, determined that a
short version of the race question was likely to produce data on the
Asian and Pacific Islander population that was as good as other test
versions of the race question. The short version provided a space
for Asian and Pacific Islanders to write in their specific groups,
such as Chinese or Asian Indian. Other test versions would have
called for Asian and Pacific Islanders to mark separate, prelisted,
detailed groups of this racial category, as was done in the 1980
Census. The Bureau's research suggested that the write-in option
would produce a somewhat higher proportion of the population
reporting as Asian and Pacific Islanders than would the use of a
question with prelisted groups. Despite its research, the Bureau was
unable to convince the Asian and Pacific Islander community that the
short version of the race question should be used. Responding to
congressional direction and pressures from the Asian and Pacific
Islander community, the Bureau reconsidered its original decision and
chose to include in the 1990 census a version of the race question
with prelisted Asian and Pacific Islander categories.
This experience demonstrated the need for the Bureau to begin to work
early in the decade to work with a diverse group of customers,
including organizations representing the interests of various race
and ethnic groups, to identify data needs for the 2000 census and the
best ways these needs can be met. In congressional hearings,
representatives of the Asian Pacific Islander community said that the
Bureau had not solicited their participation in the early phases of
redesigning the race question. The advisory committees representing
minority communities had not been established until 1986 for the 1990
Census. This was the same year that the major tests were held that
drove the debate on the race question and the Bureau's initial
recommendation. Several representatives said that the Bureau had
already formulated the census questions before the committees began
to meet.
In contrast to the situation in preparation for the 1990 census,
Census advisory committees for the 2000 census were chartered in
February 1994. These committees included the African-American
Population, American Indian and Alaska Native Populations, Asian and
Pacific Islander Populations, and the Hispanic Population. The
committees held a series of meetings during 1995 prior to the June
1996 National Content test, at which time the Bureau tested several
variations of questions designed to obtain race and ethnicity
information. Another meeting of these advisory committees was held
in December 1996 to discuss major findings from that survey.
According to Census officials, the results of this survey and the
race and ethnicity tests should be available in early May 1997.
DATA QUALITY ISSUES
---------------------------------------------- Chapter STATEMENT:2.1.2
Census Bureau evaluations suggested that the data from the 1990 race
and Hispanic origin questions were generally of high quality.
However, the evaluations also suggested that some problems associated
with collecting data on Hispanics that confronted the Bureau in 1980
continued. For example, in both the 1980 and 1990 Censuses, the
Bureau found that Hispanics had difficulty classifying themselves by
race, and this difficulty led to inconsistent reporting by Hispanics
when they initially completed the census questionnaire compared to
their responses when they were interviewed as part of a quality
check. Also, for both the 1980 and 1990 Censuses, the Bureau
estimated that nearly all--97.5 percent in 1990 and 95 percent in
1980--of the respondents who reported being in the "other race"
category were persons of Hispanic origin.
The problems experienced in connection with the Hispanic origin
question stemmed, at least in part, from the format and sequence of
the Hispanic origin and race questions. According to the Bureau,
both Hispanics and non-Hispanics have had difficulty dealing with
this issue. Some Hispanics equate their "Hispanicity" with race and
have had difficulty classifying themselves by the standard race
categories. In 1990, about 40 percent of Hispanics marked the "other
race" category; they either indicated they were Hispanic in the
Hispanic origin question or indicated they were Hispanic in the
write-in space provided in the race question. According to the
Bureau, some non-Hispanics, having already responded to the race
question, skipped over the Hispanic origin question when they should
have indicated that they were not of Hispanic origin.
As part of OMB's ongoing effort to determine whether Statistical
Policy Directive 15 should be revised, OMB and the Census Bureau are
working together to determine whether race/ethnicity should be asked
as a single identification or whether the race identification should
be separate from Hispanic origin or other ethnicities. Research on
the effect of keeping race/ethnicity as a single identification or
separate was conducted during the May 1995 Current Population Survey
conducted by the Census Bureau, as well as during the June 1996 Race
and Ethnic Targeted Test. Results from the May 1995 survey indicated
that, among other things, placing the Hispanic origin question before
the race question significantly reduced nonresponse to the Hispanic
origin question. The survey also found that placing the Hispanic
origin question before a race question that did not include a
multiracial option (1) reduced the percentage of persons reporting in
the "other race" category on the race question, and (2) increased
reporting by Hispanics in the White category of the race question.
The results of the June 1996 test are due in May 1997.
AGENCIES' USE OF CONSISTENT
RACE AND ETHNIC DEFINITIONS
------------------------------------------------ Chapter STATEMENT:2.2
In our 1992 report on agencies' use of consistent race and ethnic
definitions, we found that the eight agencies we reviewed, including
the Department of Health and Human Services and the Department of
Education, used the standard definitions of Directive No. 15 in the
data collection operations we examined.\2 However, we also found some
consistency problems in the agencies' reporting when they used data
not controlled by federal collection rules or data based on
observer-identification.
The consistent use of definitions by federal agencies is to be
accomplished by an OMB control procedure that is designed to help
ensure that standards are properly incorporated in data collection
efforts. OMB's Office of Information and Regulatory Affairs is
required to approve all federal data collection instruments and
methodologies before an agency begins collecting data. This control
process is designed to monitor the use of standards in the
development of the data collection methodologies. In our 1992
report, we found that another practice that helps ensure that the
policy's definitions are followed is the federal statistical
agencies' extensive use of the Census Bureau support and products
that also are governed by the rules of Directive No. 15. For
example, the National Center for Education Statistics (NCES) and the
National Center for Health Statistics (NCHS), as well as other
federal agencies, use elements of the Decennial Census or the monthly
Current Population Surveys in their reports. As a result, the data
agencies use from these sources should have been collected using
methodologies that OMB has approved.
Inconsistent reporting of racial and ethnic data by federal agencies
can arise when an agency uses state-provided data. The management of
state- or local-generated data is outside of federal jurisdiction.
State-provided data can be inconsistent, for example, if states
categorize race and ethnicity differently, if state data are
incomplete, or if states categorize multiracial/ethnic peoples
differently than specified by OMB Directive No. 15. Some school
districts collecting race and ethnic data also use a special category
to classify people of mixed race or ethnicity, thus creating
difficulties when data contained in this category are aggregated into
the five Directive 15 categories. State use of a multiracial
category could become a greater potential source of inconsistency in
the future, because the number of multiracial and ethnic families has
grown significantly over the last 20 years.
Inconsistent reporting of racial and ethnic data can also arise when
others determine an individual's race or ethnicity, generally from
observation. OMB's Directive recommends that the category most
closely reflecting the individual's recognition in his community
should be used for purposes of reporting on persons who are of mixed
racial and/or ethnic origins. Department of Education officials told
us that some states determine a student's race or ethnicity by that
of the mother, whereas others use the father's race or ethnicity.
Because the label applied by state policy may not be the same label
applied had OMB's guidance been used, inconsistencies may arise in
the racial categorization of those of mixed race.
--------------------
\2 Federal Data Collection: Agencies' Use of Consistent Race and
Ethnic Definitions (GAO/GGD-93-25, Dec. 15, 1992).
STATE COLLECTION OF RACE AND
ETHNICITY DATA FOR HEALTH AND
EDUCATIONAL PURPOSES
-------------------------------------------------- Chapter STATEMENT:3
States collect data on race and ethnicity in various programs, often
to comply with federal requirements. You asked that we provide
information on the types of race and ethnicity data states are
collecting in the areas of health and education. You specifically
asked us to include information on states that have enacted laws
requiring the use of a multiracial category. To respond to your
request, we focused on state reporting of race and ethnic data on
births and deaths in the health area and on student enrollment in
public schools in the education area. At the federal level, NCHS and
NCES have responsibility for compiling and reporting data nationally
on these topics and work cooperatively with the states to obtain the
data. Accordingly, we obtained information from NCHS and NCES on
their process for collecting state data as well as from a study
recently sponsored by NCES on racial and ethnic classifications used
by public schools.
In addition, we obtained information on the collection of race and
ethnic data from health departments in nine states--Alabama, Georgia,
Illinois, Indiana, Maryland, New Jersey, Ohio, Texas, and Washington.
Of these nine states, Georgia and Indiana have implemented state laws
requiring all state agencies to use a multiracial category under
certain circumstances. We also contacted education officials in the
District of Columbia and in 12 states--California, Georgia, Illinois,
Indiana, Maryland, Michigan, North Carolina, New York, Ohio,
Pennsylvania, Texas, and Virginia--as well as representatives from 23
local school systems in 8 of these jurisdictions. Of those states,
five have laws requiring the use of a multiracial category under
certain circumstances. We contacted local school representatives to
get an indication of the types of racial and ethnic data schools were
collecting and, if the state had implemented a law or requirement for
use of a multiracial category, whether schools were implementing the
requirement. We selected the five states that have laws requiring
schools' use of a multiracial category and North Carolina because it
has administratively mandated the use of a multiracial category in
its school system. We judgmentally selected local school systems to
achieve geographic dispersion. Due to the limited scope of our work,
we cannot project our findings to jurisdictions or school systems we
did not contact.
Health department officials in all nine states said they are
collecting race and ethnicity data from birth and death certificates,
aggregating these data, and reporting the data to NCHS, in accordance
with guidance and instructions provided by NCHS. According to these
officials, the guidance and instructions used conform to OMB's
Statistical Directive No. 15.
State education officials in the 12 states and the District of
Columbia said they are collecting race and ethnicity data from
schools and aggregating these data. Most of these officials said
that data are reported to the federal government in the five
categories contained in the directive. Officials from Georgia and
Indiana, which are two states that have implemented legislation
requiring the use of a multiracial category, said they are reporting
race and ethnicity data to the federal government in the five
categories contained in the directive. However, in order to report
these data, these officials said that individuals in the multiracial
category must be allocated to the other racial and/or ethnic
categories. In contrast, an Ohio education official said that this
state, which has a state law requiring the use of a multiracial
category, uses six categories to report these data, including a
multiracial category. Based on discussions with state and local
education officials, the forms used to collect racial and ethnic
information, the categories used to classify race and ethnicity, and
who classifies students vary from school to school.
FIVE STATES HAVE LAWS
REQUIRING THE USE OF A
MULTIRACIAL CATEGORY
------------------------------------------------ Chapter STATEMENT:3.1
Georgia, Illinois, Indiana, Michigan, and Ohio have enacted laws that
include the term multiracial with respect to collecting data on race
and ethnicity. Georgia, Indiana, and Michigan have passed
legislation requiring the addition of a multiracial category on all
state forms that have a listing of racial and ethnic classifications
from which one must select. Illinois and Ohio have similar
legislation that applies only to educational departments and schools;
however, Illinois' law specifies that the category of multiracial is
to be collected and reported only if the data are for state or local
use. Although Illinois' and Ohio's laws do not define multiracial,
the other three states' laws define multiracial as having parents of
different races. For federal reporting purposes, laws in Georgia and
Michigan provide for reallocating multiracial individuals into the
five federal categories on the basis of the rate that the general
population comprises each classification. Ohio's law, in contrast,
requires that the parent, guardian, or custodian of each student have
the opportunity to designate the appropriate federal racial category
for the student. Georgia's, Indiana's, and Ohio's laws are being
implemented. According to an education official in Illinois, that
state is delaying implementation of its law until OMB makes a
determination on revising Statistical Directive 15. Michigan's law,
according to state education officials, is to be implemented during
the 1997-1998 school year when forms are scheduled to be revised to
include a separate question to collect data on whether the individual
is also multiracial. A summary of the scope and implementation
status of laws in the five states is attached to this statement.
COLLECTION OF RACE AND
ETHNICITY DATA FOR HEALTH
PURPOSES
------------------------------------------------ Chapter STATEMENT:3.2
We focused our efforts on the collection of race and ethnicity data
for health purposes on data collected in connection with births and
deaths. NCHS is responsible for compiling national statistics on
births and deaths. To do this, it is to work cooperatively with
state health departments which administer birth registration and
death reporting systems under the laws and regulations of the states.
Birth certificates are to be used to compile annual vital statistics
on the number and rate of births by such characteristics as place of
birth and residence of mother. Population composition and growth are
to be also estimated using these data, and the data are to be used in
planning and evaluating programs in public health and other areas.
Information from death certificates is to be used for many purposes,
including assessing the general health of the population, examining
medical problems that may be more prevalent among certain population
groups, and identifying geographic areas with elevated death rates
from selected causes of death.
NCHS has developed model birth and death certificates that states can
use in recording vital data on births and deaths. Information on the
birth certificate can be provided by a number of individuals,
including the mother and/or father, a physician or other hospital
personnel, or a midwife. To obtain data on race and ethnicity, the
model birth certificate includes a question on whether the mother is
or is not of Hispanic origin, and the same question on the father as
well. NCHS guidance clearly states that this information is not part
of the race item, and that a person of Hispanic origin may be of any
race. Thus, the Race and Hispanic origin questions are asked
independently in the model certificates. A blank space follows the
Hispanic origin question so that the person completing the form can
write in the race of the mother, as well as the race of the father.
Instructions for completing that block direct that the information
should be obtained from the parent or parents, or other informant,
and that the entry reflect the response of the informant. For Asians
and Pacific Islanders, the national origin of the mother and father,
such as Chinese, Japanese, Korean, Filipino, or Hawaiian, is to be
entered. If the informant indicates that the mother and/or father is
of "mixed race," both races or ancestries are to be entered.
Birth certificates do not record the race of the child. Beginning in
1989, NCHS guidance recommends that health officials categorize
births by the race of the mother, as recorded on the birth
certificate. If the mother does not state her race on the
certificate, then the baby's race is imputed to be the race of the
father. Prior to 1989, newborns with two parents of the same race
were classified as that race; newborns with one nonwhite parent were
classified as the race of the nonwhite parent. When both parents
were nonwhite but of different races, the newborn was assigned the
father's race; except that if either parent was Hawaiian, the newborn
was classified as Hawaiian.\3
With regard to death certificates, funeral directors are responsible
for getting the death certificate completed. Information is to be
obtained from the spouse, one of the parents, children, or another
relative of the decedent, or a physician. In contrast to the birth
certificate, which requires data on whether the mother or father is
of Hispanic origin and the race of the mother or father, the death
certificate requires this information on the decedent.
According to officials in nine state health departments, these states
follow NCHS' model and guidance when obtaining this information.
Birth and/or death certificates provided us by seven of these states
confirmed that their forms provide a space in which any race or
racial makeup can be provided. For reporting purposes, NCHS has
developed a coding system and guidance for categorizing and reporting
race and ethnicity responses. This guidance includes 7 codes for
Hispanic origin, as well as up to 16 codes for race. The race
category includes White, Black, Indian, five categories for Asian or
Pacific Islander, Other Entries, and Not Reported. According to NCHS
officials, an additional 6 coding categories for other Asian or
Pacific Islanders, for a total of 11, were being used by California,
Hawaii, Illinois, New Jersey, New York State, New York City,
Minnesota, Texas, and Washington. NCHS provides the option to other
states on whether to use the more detailed categories for Asian or
Pacific Islanders. If an individual were to record his or her race
as more than one race on the certificate, NCHS' guidelines provide
that the first race listed should be used for coding purposes, or if
the individual reports percentages, the race having the highest
percentage should be coded. Thus, even though an individual can
indicate a makeup of various races, for purposes of aggregating the
data, the individual is coded as belonging to one race by health
officials.
Of the nine states whose health departments we contacted, Georgia and
Indiana have laws requiring the use of a multiracial category when
collecting race and ethnicity data. According to a Georgia
Department of Public Health official, birth and death certificates
were not modified to accommodate the law because the law requires
that forms be modified only if they contain a list of categories from
which one must select, and Georgia's birth and death certificates do
not contain such a list. This health department official said that
when an individual indicates that he or she is multiracial, the state
obtains data on which races the individual belongs to and then
follows NCHS' protocol for coding racial data. For example, if Black
is the first race mentioned, the person is coded as Black. If White
is the first race mentioned, the person is coded as White. This
official said that if some other race is mentioned first, the person
is coded as an "other" race because all of the other races are very
small in the state. According to this official, the Department has
also included a field in its database to indicate if the individual
considers himself or herself multiracial. These data are separate
and not reported to NCHS. This official said that Georgia started
collecting the multiracial data about 1 or 2 years ago, but that it
has not used the multiracial data to perform any special statistical
analyses. Furthermore, he said that the number of individuals
classified as multiracial is small, and that the data would not
likely be used for analytical purposes because a racial designation
is generally not a good indicator of health problems. Likewise,
Indiana's law, passed in 1995, requires the addition of a multiracial
category on forms that contain a listing of racial categories. An
Indiana health official said that the state has not changed its
forms, practices, or procedures for collecting racial or ethnic data
on birth or death certificates. This state uses birth and death
certificates that do not contain a listing of categories. According
to this official, the health department has not added a separate
multiracial category in the state's database so that the number of
multiracial persons could be tabulated by computer.
None of the officials in the seven state health departments we
contacted that do not require the category of multiracial indicated
the need for one or were sure of how it would be used if collected.
For example, according to an Illinois Department of Public Health
official, the multiracial classification would not be useful with
respect to conducting vital statistical or health analyses.
According to this official, if additional data were to be collected
for vital statistical health analyses, these data should include
socioeconomic data. Similarly, an Alabama Department of Public
Health official stated that, currently, the state does not need a
multiracial category because, in her view, the state has a relatively
few number of individuals who could be classified as such.
Therefore, she said that NCHS' classifications were currently
sufficient for determining different racial and ethnic groups'
lifestyles that affect health.
The Department of Health and Human Services' National Committee on
Vital and Health Statistics, Subcommittee On Health Statistics for
Minority and Other Special Populations, has reviewed state
legislation requiring the use of a multiracial category. According
to a subcommittee member, of some concern to the subcommittee is the
fact that some state laws call for reallocating the number of
multiracial people to OMB's five categories for federal statistics on
the basis of the racial and ethnic distribution of the general
population. She also said that reallocating multiracial people on
this basis would tend to misrepresent the number of individuals in
each of the five categories because the multiracial population does
not have the composition of the racial and ethnic distribution of the
general population.
--------------------
\3 This change in policy was brought about because it brought a more
uniform approach to tabulating the increased incidence of interracial
births and nonmarital births than did the necessarily arbitrary
combination of parental races. In addition, the model birth
certificate was modified in 1989 to include more questions directly
associated with the mother's health and health behaviors.
COLLECTION OF RACE AND
ETHNICITY DATA FOR
EDUCATIONAL PURPOSES
------------------------------------------------ Chapter STATEMENT:3.3
NCES is the primary federal organization for collecting, analyzing,
and reporting data related to education. NCES collects data through
a variety of means, including periodic surveys and data reported by
states from data contained in administrative records, including
school enrollment records. In 1994, NCES updated its 1974 national
standards for student data to establish current and consistent terms,
definitions, and classification codes to maintain, collect, report,
and exchange comparable information about students.\4
Its 1994 handbook, intended as a reference document, includes the
types of information that could be collected about individual
students and maintained in records, and discusses race/ethnicity
classifications and definitions. The racial and ethnic
classifications and definitions are those contained in OMB's
Statistical Directive No. 15--American Indian or Alaskan Native,
Asian or Pacific Islander, Black, Hispanic, and White. The handbook
recommends that Blacks and Whites be separate from Hispanic. That
is, students should be classified as Hispanic, Black (not Hispanic)
or White (Not Hispanic).
In the spring of 1995, NCES sponsored a survey as part of the
research being conducted by OMB to review the current categories in
its Statistical Directive. NCES' March 1996 report summarized its
findings with regard to racial and ethnic classifications used by
public schools.\5 In summary, 73 percent of 926 public schools
responded that they used only the five standard federal categories to
classify the race and ethnicity of students. Of the remaining 27
percent of the schools, 10 percent responded that they used "other"
or "undesignated," with space for indicating a specific race or
ethnicity; 5 percent used "other," without space for specification; 7
percent used additional racial and ethnic categories, such as
"Filipino"; 5 percent used a general multiracial category; and the
remaining 2 percent used specific combinations of the five standard
federal categories or used an "unknown" category (such as
"Black/White," or an "unknown" category).\6
Additional differences were found in how racial and ethnic data were
aggregated into the five federal categories before reporting the data
to the federal government, as well as differences in who identified
the race and ethnicity of the children. For example, about half of
the 27 percent of schools that used other classifications reported
that the central district office handled the task of aggregating this
information, while many of the remaining schools reported that
students were allocated by the school among the five standard federal
categories based on which ones the school considered most
appropriate. With respect to who identified a student's race or
ethnicity, 73 percent of the schools reported that they asked
parents. At 22 percent of the remaining schools, respondents
reported that teachers or administrators assigned students to
categories based on observation; while the remaining 5 percent
reporting using some other method. Most respondents reported that
revisions to Directive No. 15 were not an issue or were only a minor
issue in terms of their applicability to students enrolled in their
schools. However, between 3 and 12 percent of schools reported that
issues such as adding a general "multiracial" category, adding an
"other" category, or changing the terminology used in the racial
categories was significant.
The 5 states that have already enacted laws requiring schools to add
a multiracial category were included in the 12 states and the
District of Columbia whose education departments we contacted. Of
these five states, three--Ohio, Georgia, and Indiana--said they have
implemented requirements that the multiracial option be included as a
category when selecting options for race and ethnicity. All of the
eight local school systems we contacted in these three states (three
in Ohio, three in Georgia, and two in Indiana) said they were using a
multiracial category for school enrollment. According to education
department officials in Ohio and Georgia, the legislation was the
result of the efforts of a small group of individuals in each of
their respective states.
According to an Indiana Department of Education official, the schools
in Indiana are implementing the requirement in a variety of ways.
For example, in some instances school administrators may classify
students into different race and ethnic categories, while in other
schools or school systems, parents may classify their children. We
contacted two schools in Indiana and found that schools in that state
were inconsistent in who classified students into race and ethnic
categories. The Indiana Department of Education official said that
when reporting data to the federal government, the state Department
of Education allocates those students who are classified as
multiracial into the Black, White, American Indian or Alaskan Native,
or Asian or Pacific Islander category on the basis of the percentage
of these races in the state as a whole. The remaining two
states--Michigan and Illinois--have yet to implement their
legislation. According to Michigan Department of Education
officials, Michigan was allowed to use up its stock of forms prior to
collecting the new data. Currently, forms for collecting racial and
ethnic data are being revised for the 1997-98 school year. According
to an Illinois State Board of Education official, Illinois is not
implementing its legislation until OMB reaches a final decision on
whether, and if so, how, Statistical Directive 15 is to be revised.
A few states, such as North Carolina and Florida, have
administratively decided to collect multiracial data. According to a
North Carolina Department of Public Instruction official, the
administrative order has been in effect for the past 1 to 2 years and
gives parents who wish to classify their child as multiracial the
option of doing so. School registration forms we obtained from three
North Carolina school systems listed a multiracial category. Two
North Carolina local school systems provided space on their
registration forms for race or ethnicity to be written in; however,
one of these local system's guidance instructed school administrators
to code this information as Black, White, Hispanic, American Indian,
Asian, or Other. A North Carolina education official said that
currently, fewer than 200 of the 1.2 million students in that state's
schools were estimated to be using the multiracial classification.
This official stated that schools within North Carolina could be
using different methods to classify these students as another race
when aggregating race and ethnic data for federal reporting purposes.
Several other states are considering including a multiracial category
for educational reporting purposes. For example, a California
Department of Education official said that a ballot measure is
currently being considered that would mandate the use of a
multiracial category when collecting data on students' race and
ethnicity. He said that the state uses OMB Statistical Directive 15
to obtain data from the schools on race and ethnicity. The state
also collects data on individuals of Philippine origin due to a state
mandate. This official said that the state would need to resolve
several issues before adopting a multiracial category. These issues
include the cost of requiring that data be collected and the
usefulness of the data to the Department. For example, this official
was not certain what utility the multiracial category provided
educators, or what educators could do with these data if they had
them.
A Texas Education Agency official said that Texas is currently
considering a proposed bill that would include a multiracial
category. Currently, the state is looking at the practical
implications of implementing the proposed legislation. That is, the
state is assessing the potential cost of requiring these data be
collected and ways that the new data could be linked to data
previously collected. According to this official, even though the
state uses OMB's Statistical Directive No. 15 for federal reporting
purposes, some schools in the state collect additional data and break
out the racial or ethnic classifications in more detail than
required.
The varying situations in the Washington D.C. metropolitan area with
regard to identifying race and ethnicity at the time of school
enrollment provide a good illustration of the complexity and
controversial nature of this issue. The school registration form
used in the District of Columbia provides a space for the parent or
guardian of the child to write in a race or ethnicity. Categories
are not prelisted on the form. The school system, however, tabulates
data based on the five federal categories.
Virginia's Department of Education has provided guidance to the
state's school systems directing them to use the five categories
specified in Statistical Directive No. 15 when reporting data to the
department, and informing them that it can only accept data in these
five categories. However, an education department official said that
schools can collect any detailed data they want on students' race and
ethnicity. The city of Alexandria uses OMB's five categories for
collecting data on school enrollment, and school system
representatives told us that the lack of a multiracial category has
not been a major issue. In contrast, Fairfax County has added a
multiracial category to its school registration form. According to a
school system representative, this category was added in 1994 as a
result of concerns expressed by a number of county residents. The
school system allocates multiracial students to the other five
categories when aggregating data and reporting it to others.
Arlington County schools also use the five categories in OMB's
directive. However, in 1993, the county added an "other" category
out of concern that a significant number of students did not fit into
the five categories and so a student or parent could use a
classification other than those listed on the school forms. However,
Arlington County, after receiving a notice from the state that it
would not accept data reported in any categories other than the five
specified by OMB, discontinued using the "other" category in 1996.
Montgomery County, Maryland, schools prelist the five categories on
its school enrollment forms. An official of the Montgomery County
schools said that some parents have expressed concern about there
being only five categories. The official said that the school board
discussed the issue, but has decided to wait to see what changes, if
any, the federal government makes. According to the official we
spoke with, the issue is very much a concern in Montgomery County
schools because the county is home to many multiracial families. In
Prince George's County, a school official told us that the school
system uses only the five standard categories but has been approached
about the issue of more categories on some occasions. However, the
school system is also under a federal court desegregation order. The
order requires the school system to report race and ethnic data in a
certain format to the court annually.
California provides another example of the diverse way in which
racial and ethnic data can be collected. California state education
officials said that even though the state reports racial and ethnic
data to the federal government in accordance with OMB's guidance,
schools in that state are asked to break out race and ethnicity data
into seven categories. These categories include OMB's
classifications of White, Black, Hispanic, American Indian or Alaskan
Native, but the classification for Asian is separated from Pacific
Islander. As noted earlier, the state also separates information on
those of Philippine ancestry. However, these officials said that
schools may collect more detailed data to determine the
representation in their school districts. According to one state
official, the state does not dictate or control the amount or type of
information schools collect because such matters are believed best
left to local control. San Diego, for example, lists 19 racial and
ethnic categories plus a multiracial category on its school
registration form. The instructions on the form state that one of
the 19 listed categories is to be selected, but that a person can
also select the multiracial category. If this latter category is
chosen, additional categories are to be written on the form. San
Francisco's city schools use nine categories and do not use a
multiracial category. Long Beach collects only the seven state
required categories of racial and ethnic information, but also
collects over 50 different language categories on each student.
Because the state may receive additional categories of racial and
ethnic data from some schools, the state, at times, has to aggregate
the data to conform to OMB's five classifications. A state education
official said that this has not caused the state any problems.
--------------------
\4 National Center for Education Statistics: Student Data Handbook
for Early Childhood, Elementary, and Secondary Education, U.S.
Department of Education, Office of Educational Research and
Improvement (NCES-94-303, June 1994).
\5 National Center for Education Statistics: Racial and Ethnic
Classifications Used by Public Schools, U.S. Department of
Education, Office of Educational Research and Improvement (NCES
96-092, March 1996).
\6 Percentages do not sum to 100 because respondents could select any
and all categories that applied to their schools.
------------------------------------------------ Chapter STATEMENT:3.4
Mr. Chairman, this concludes my prepared statement. I would be
pleased to respond to questions from you or other Members of the
Subcommittee.
EXAMPLES OF FEDERAL LAWS REQUIRING
THE COLLECTION OF DATA ON RACE AND
ETHNICITY
=========================================================== Appendix I
Agency or program requiring Description of data
race and ethnic data requirement Legal authority
---------------------------- -------------------- ----------------------------
Federal Home Loan Mortgage The corporation is 12 U.S.C. µ1456(e)(1)(A)
Corporation required to collect
data relating to its
mortgages on housing
consisting of 1 to 4
dwelling units.
These data shall
include the income,
census tract
location, race, and
gender of
mortgagors.
Government National Mortgage The corporations 12 U.S.C. µ1723a.(m)(1)(A)
Association and Federal shall collect data
National Mortgage relating to their
Association mortgages on housing
consisting of one to
four dwelling units.
These data shall
include the income,
census tract
location, race, and
gender of
mortgagors.
Community Development The Fund shall 12 U.S.C. µ4714(b)
Financial Institutions Fund require each
community
development
financial
institution or other
organization
receiving Fund
assistance to
compile such data,
as is determined to
be appropriate by
the Fund, on the
gender, race,
ethnicity, national
origin, or other
pertinent
information
concerning
individuals that
utilize the services
of the assisted
institution.
Department of Education, All statistics and 20 U.S.C. µ6011(j)(4)
Office of Educational other data collected
Research and Improvement and reported by the
Office shall be
collected, cross-
tabulated, analyzed,
and reported by sex
within race or
ethnicity and
economic status
whenever feasible.
Department of Education Shall carry out an 20 U.S.C. µ6491(c)(1)(e)
ongoing evaluation
of programs designed
to help
disadvantaged
children meet high
standards. The
evaluation shall,
when feasible,
collect, cross-
tabulate, and report
data by sex within
race or ethnicity
and socioeconomic
status.
Department of Education, Shall collect and 20 U.S.C. µ9010(b)(1)(C)
Office of Educational report data for the
Research and Improvement National Assessment
of Educational
Progress. The data
shall include
information on
special groups,
including, whenever
feasible,
information
collected, cross-
tabulated, analyzed,
and reported by sex,
race or ethnicity,
and socio-economic
status.
Department of Labor, Requires an annual 29 U.S.C. µ712
Rehabilitation Services report on vocational
Administration rehabilitation and
other rehabilitation
services from
information
collected on each
client. The
information shall
set forth a complete
count of such cases
in a manner
permitting the
greatest possible
cross-
classification of
data. The data
elements shall
include, but not be
limited to, age,
sex, race,
ethnicity, etc.
Veterans Administration, Requires social and 38 U.S.C. µ317(d)(5)
Center for Minority Veterans demographic research
on the needs of
veterans who are
minorities and the
extent to which
veterans' programs
meet the needs of
minority veterans,
without regard to
any law concerning
the collection of
information from the
public.
Department of Commerce Requires a survey to 42 U.S.C. µ2000f.
compile registration
and voting
statistics. The
survey and
compilation shall
only include a count
of persons of voting
age by race, color,
and national origin.
The law also has a
proviso that no
person shall be
compelled to
disclose his race,
color, or national
origin and that
every person
interrogated orally,
by written survey or
questionnaire, or by
any other means
shall be fully
advised about his
right to refuse to
furnish such
information.
Department of Housing and Requires studies 42 U.S.C. µ3608(e)
Urban Development with respect to the
nature and extent of
discriminatory
housing practices
and requires
information on the
race, color,
religion, sex,
national origin,
age, handicap, and
family
characteristics of
persons and
households who are
applicants for,
participants in, or
potential
beneficiaries of,
programs
administered by the
Department of
Housing and Urban
Development. The
Department shall
collect the
information relating
to those
characteristics the
Department
determines necessary
and appropriate
without regard to
any other provision
of law.
--------------------------------------------------------------------------------
Source: GAO analysis of selected legislation.
STATES WITH LAWS REQUIRING A
MULTIRACIAL CATEGORY
========================================================== Appendix II
Law As of April 1997 is law
State enacted being implemented? Agencies covered
-------- -------- ------------------------ ------------------------
Georgia 1994 Yes All state agencies
Illinois 1994 No Schools
Indiana 1995 Yes All state agencies
Michigan 1995 No All state agencies
Ohio 1992 Yes Schools
----------------------------------------------------------------------
Source: State laws and state officials.
Copyright © 1998 The Multiracial Activist. All rights reserved.