Guidelines for using Racial and Ethnic Groups in Data Analyses
Updated: July 2003

Purpose

Background and Context

Definition of Minimum Categories

Guidelines

Implementation Schedule
Data Collection
Data Tabulation
Data Presentation

Additional Information

Special Issues

Estimates of "More than One Race" in Washington State
Bridging Ratios for Trend Analysis

Bridging Methods
Sample Use of Statistical Methods
Recommended Method
Additional Information

Converting Multiple to Single Race: Washington State Department of Health Discussion Paper

Guidelines For Using Racial and Ethnic Groups in Data Analyses (Word Document)

>>> Next >>>

Purpose

The Assessment Operations Group in the Washington State Department of Health is coordinating the development of guidelines related to data development and use in order to promote good professional practice among staff involved in assessment activities within the Washington State Department of Health and in Local Health Jurisdictions in Washington. While the guidelines are intended for an audience of differing levels of training related to data development and use, they assume a basic knowledge of epidemiology and biostatistics. They are not intended to recreate basic texts and other sources of information related to the topics covered by the guidelines, but rather they focus on common issues encountered in public health practice and where applicable, on issues unique to Washington state.

Background and Context

Epidemiology is the study of the distribution and determinants of disease frequency in the human population. Epidemiologists often examine this distribution by race or ethnic group. The concepts of race and ethnic group and the meaning assigned to these concepts have changed considerably over time. Current research on the human genome shows that genetic variation within a particular "race" group is much larger than the variation between groups. (See, for example, Rosenberg NA, et al. Science, 2002, 298: 2381-2384; Garte S. Public Health Reports, 2002117 421-425; Kaplan MS and Bennett T. JAMA, 2003, 289;2709-2716.) Thus, while sickle-cell anemia is often thought of as an "African-specific" genetic condition, people from certain areas of Greece, Italy and the Arabian Peninsula, generally classified as "white," are also affected. Similarly, malignant melanoma is most common in "whites," but affects people of all "races" and skin types.

Most scientists do not believe race is a valid biological construct. (See discussion in 'The Meaning of Race in Science - Considerations for Cancer Research,' Report of the President's Cancer Panel, National Institutes of Health, April 7, 1997.) Researchers, such as Camara Jones, propose that "race is only a rough proxy for socioeconomic status, culture, and genes, but it precisely captures the social classification of people in a race-conscious society such as the United States." … That is, the variable 'race' is not a biological construct that reflects innate differences, but a social construct that precisely captures the impacts of racism." (Jones CP, AJPH, 2000, 90:1212-1215). 

Thus, in most public health assessment, race and ethnic group should be viewed as capturing the effects of complex social, cultural, economic and political factors on human health, and these factors must be addressed in interpreting health data. For example, good birth outcomes among Mexican-American women are thought to be related to socio-cultural practices supportive of healthy lifestyle choices during pregnancy. Discrimination and racism may affect the quality of medical care, leading to poorer health outcomes among African Americans and other race and ethnic groups. (See, for example, Smedley B et al. Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care, Institute of Medicine, 2003.) Additionally, race may sometimes reflect measures of socioeconomic factors. For example, sometimes people "belonging to" a specific racial and ethnic groups may, as a group, have fewer material resources than those in other groups. Differences in health status caused by lack of access to material goods may appear as differences in health status among racial and ethnic groups, although the root cause is not race or ethnicity. Epidemiologic analysis can be used to assess these relationships and guide interventions. 

The distribution of health care in the community may also be a reason for measuring health data by race and ethnic group. Some health care providers and other social and health service organizations serve people primarily from one or several racial or ethnic groups. These providers and organizations want to know the health status of the groups they serve and one method of delineating this is to analyze health data by race and ethnic group. Concepts of race and ethnic group have changed over time. During the early 1990's, the United States Office of Management and Budget (OMB) embarked on a nationwide review of the federal guidelines for reporting race and ethnicity that had been in effect since 1977 (Statistical Policy Directive No. 15). As a result of that review, OMB issued a revised standard in 1997 that was used in the United States Census in 2000, in birth and fetal death certificates in 2003, and in death certificates in 2004. The three major changes in the 1997 standard are:

• People can be identified by more than one racial category
• Pacific Islanders will no longer be classified with Asians
• The question on Hispanic/Latino ethnicity will be asked before the race question.

Definition of Minimum Categories

The minimum categories established in the 1997 OMB Standard for Federal Data on Race and Ethnicity are:

• Race
  • American Indian or Alaska Native (AIAN): A person having origins in any of the original peoples of North and South America (including Central America), and who maintains tribal affiliation or community attachment.
  • Asian: A person having origins in any of the original peoples of the Far East, Southeast Asia, or the Indian subcontinent, including, for example, Cambodia, China, India, Japan, Korea, Malaysia, Pakistan, the Philippine Islands, Thailand and Vietnam.
  • Black or African American: A person having origins in any of the black racial groups of Africa. Terms such as "Haitian" or "Negro" can be used in addition to "Black or African American."
  • Native Hawaiian or Other Pacific Islander (NHOPI): A person having origins in any of the original peoples of Hawaii, Guam, Samoa, or other Pacific Islands.
  • White: A person having origins in any of the original peoples of Europe, the Middle East, or North Africa.
• Ethnic Group
  • Hispanic or Latino: A person of Cuban, Mexican, Puerto Rican, South or Central American or other Spanish culture or origin, regardless of race. The term "Spanish origin" can be used in addition to "Hispanic or Latino."

Guidelines: Implementation Schedule

• Continue to use the OMB 1977 Standards for Race and Ethnicity for all data pertaining to years through 1999, regardless of date of publication
• Adopt the OMB 1997 Standards for Race and Ethnicity for collecting data by January 1, 2005.
• Adopt the OMB 1997 Standards for Race and Ethnicity for tabulating any data collected using the 1997 standard.
• January 1, 2000 - December 31, 2004 is called the transition period. Some programs will use the 1997 standard and some programs will use the 1977 standard. Bridging estimates may have to be used for data comparability.

Guidelines: Data Collection

• Respondents and informants may identify with more than one race. Data on multiple race should be collected as multiple responses to a single question rather than from a separate "multiracial" category. Recommended question wording is "Mark one or more...," "Select one or more....," or "Race - enter one or more."
• For ethnic group, data should be collected on whether or not a person is of Hispanic or Latino culture or origin, but the OMB standards do not permit multiple responses (e.g., both Hispanic and non-Hispanic heritage). To help provide more complete race data for Hispanic respondents, an instruction to answer both the Hispanic or Latino question and the race question may be useful, especially for mail surveys or self-administered questionnaires.
• Data should be collected separately for race and ethnicity, with ethnicity collected first.
• If a combined race/ethnicity format must be used, a person should be allowed to select more than one racial/ethnic category.
• The specific terminology for racial and ethnic categories as described above should be used whenever the categories are listed.
  • Data collection is not limited to the categories described above. In fact, the collection of subgroup detail is encouraged.
  • Subgroup detail could be collected through write-in entries or follow-up questions asked by the interviewer. 
  • If more detailed categories are used, then the additional categories must be organized so that they can be aggregated into the minimum categories described above. 
  • Additional categories should be mutually exclusive 
  • Additional categories should be consistent with available denominator data if one intends to calculate rates. 
  • Additional categories need to be meaningful to the populations about whom data are being collected. If possible, involve the communities in developing additional categories to help assure this.
• Mode of administration should be considered when designing questions and instructions.
  • For face-to-face surveys, a flashcard with the categories may be useful.
  • For telephone surveys, fewer response categories could be used, with follow-up questions to provide more detail. The way response options are read is important. To avoid confusion, pause between categories, e.g., White (pause) Black or African American (pause) etc so that the respondents don't think they have to choose between Black and African American.
  • For self-administered forms with a check box format, definitions for the minimum race categories may be needed.
• Use self-reporting rather than observer identification whenever possible. If self-reporting is not possible (e.g., for a deceased person), attempt to obtain proxy responses from family or friends before using observer identification. 
• Use translated data collection forms to ensure inclusion of people from diverse backgrounds whenever possible.

DATA COLLECTION EXAMPLE FOR SELF-ADMINISTERED QUESTIONNAIRE

NOTE: Please answer Both questions 5 and 6.

5. Are you Spanish/Hispanic/Latino? Mark [X] the "No" box if not Spanish/Hispanic/Latino.

• No, not Spanish/Hispanic/Latino
• Yes, Mexican, Mexican Am, Chicano
• Yes, Puerto Rican
• Yes, Cuban
• Yes, other Spanish/Hispanic/Latino – Print group

6. What is your race? Mark [X] one or more races to indicate what this person considers himself/herself to be.

• White
• Black, African Am., or Negro
• American Indian or Alaska Native – Print name of enrolled or principal tribe
• Asian Indian
• Chinese
• Filipino
• Japanese
• Korean
• Vietnamese
• Native Hawaiian
• Guamanian or Chamorro
• Samoan
• Other Asian – Print race
• Other Pacific Islander – Print race
• Some other race – Print race

Guidelines: Data Tabulation

• Tabulate data so as to accurately represent each person's choice of racial and ethnic identity.
• Include as much detail on race and ethnicity as possible without compromising data quality or confidentiality. More detail can usually be published for population totals than for attributes (e.g., income, education, or health outcomes).
• Aggregate data by racial/ethnic group only when such an aggregation provides meaningful categories (e.g., "nonwhite" or "other" is not usually a meaningful category).
• In addition to the minimum racial and ethnic categories (and subgroup detail where possible), data tables may include three other categories:
  • Other Race: Use for responses that do not match any of the standard racial categories.
  • Race Not Reported: Use when information on race was not provided. If data are available, this category can be subdivided according to the reason that information on race was not obtained: refusal, don't know, and not ascertained.
  • Not Tabulated Above: Use to aggregate responses for any racial categories (e.g., a single race or responses to more than one category) that do not contain enough people to be published separately because of data quality or confidentiality concerns.

The following tables show some options for presenting data. On all tables the five race categories (plus 'Other') are listed alphabetically.

• Follow this example when sample sizes or population counts do not permit greater detail.

Table 1. Minimum Presentation of Data on Race

Race	Number	Percent
Total
AIAN1
Asian
Black
NHOPI1
Other
White
More than one race
Race Not Reported

¹AIAN=American Indian or Alaska Native; NHOPI=Native Hawaiian or Other Pacific Islander

If sample or population sizes for any of the racial groups are too small to meet data quality or confidentiality standards, then combine these racial groups into a category labeled 'Not Tabulated Above' and present data only for the larger racial categories.

• Follow this example to report data containing the minimum racial groupings plus additional detail (e.g., under "Asian" and under "More than one race").

  Table 2. Detailed Presentation of Data on Race

  Race	Number	Percent
  Total
  AIAN1
  Asian
  Asian Indian
  Chinese
  Filipino
  Japanese
  Korean
  Vietnamese
  Black
  NHOPI1
  Other
  White
  More than one race
  AIAN/White
  Asian/White
  Black/White
  Other/White
  Race Not Reported

  ¹AIAN=American Indian or Alaska Native; NHOPI=Native Hawaiian or Other Pacific Islander

• Report as much detail under the category "More than one race" as needed to represent the responses provided by individuals (without violating confidentiality or data analysis standards) and such that the minimum set of racial categories can be recreated. The detail reported under 'More than one race' in this table represents the most frequently reported combinations of racial categories based on National Health Interview Survey data. For other populations, the most frequent combinations may be different.
• When data systems collect more detail on subgroups of the main categories, some persons may indicate that they belong to more than one subgroup. For example, in the Asian category, respondents might indicate both Chinese and Japanese heritage. These respondents should be included in the single race total for Asians, not in the "More than one race" category. If sample size permits, they can be tabulated separately as an Asian subgroup "More than one Asian race."
• · Follow this example to reflect the complexity of reporting race and to present inclusive categories. (Note: persons may be counted in more than one category in the table below.)

  Table 3. Detailed Presentation of Data on Race and the All Inclusive Distributions

  Race	Number	Percent
  Total
  AIAN1
  Asian
  Asian Indian
  Chinese
  Filipino
  Japanese
  Korean
  Vietnamese
  Black
  NHOPI1
  Other
  White
  More than one race
  AIAN/White
  Asian/White
  Black/White
  Race Not Reported
  AIAN all inclusive
  AIAN and other race(s)
  Asian all inclusive
  Asian and other race(s)
  (etc for other categories)

  ¹AIAN=American Indian or Alaska Native; NHOPI=Native Hawaiian or Other Pacific Islander

• Use the 'all inclusive' headings to represent persons who report a particular race either alone or in combination with other race(s). The distributions of individuals under the all-inclusive category may provide information on groups that are not of sufficient size in the sample or population to be included in basic tabulations.
• Present data on Hispanic or Latino ethnicity using the format shown in Table 4, similar to the race tabulations, but no provision is made for reporting more than one subgroup or both Hispanic and non-Hispanic ethnicity. The subgroups used will be a function of the sample size and the population composition where the data are collected.

  Table 4. Hispanic or Latino Ethnicity with Detail

  Ethnicity	Number	Percent
  Total
  Hispanic/Latino
  Cuban
  Mexican
  Puerto Rican
  Not Hispanic/Latino
  Ethnicity Not Reported

  
  
• Report race by ethnicity whenever possible without violating data quality or confidentiality standards. Thus, Tables 1-3 could be further subdivided by tabulating data in each category for "Hispanic or Latino," "Not Hispanic or Latino," and "Ethnicity Not Reported." Table 5 shows Table 2 subdivided in this way. Follow this example to report race by ethnicity.

  Table 5. Detailed Presentation of Data on Race and Hispanic or Latino Ethnicity

  Race	Number	Percent
  Total
  Hispanic or Latino
  AIAN1
  Asian
  Black
  NHOPI1
  Other
  White
  More than one race
  Race Not Reported
  Not Hispanic or Latino
  AIAN
  Asian
  Asian Indian
  Chinese
  Filipino
  Japanese
  Korean
  Vietnamese
  Black
  NHOPI
  Other
  White
  More than one race
  AIAN/White
  Asian/White
  Black/White
  Race Not Reported
  Ethnicity Not Reported
  White
  Race Not Reported

  ¹AIAN=American Indian or Alaska Native; NHOPI=Native Hawaiian or Other Pacific Islander

  Note: Not all categories are included due to small cell sizes. Values for these cells are included in the total category and appropriate subcategories; therefore, subcategories may not add to total.

If possible, collect data under both the 1977 and the 1997 OMB standards for a sufficient period of time to create an estimate of the characteristics of the population under consideration under both methods for reporting race and ethnicity. Use these data to calculate a bridging ratio for use in trend analysis.

• If it is not possible to calculate a bridging ratio that is directly applicable to a specific data collection system, then use a bridging ratio available from another source that most closely approximates the population under consideration (e.g., state or county survey).
• If it is impractical to use a bridging ratio or this approach is not suitable, then clearly indicate that there is a break in the data series between data collected before and after the implementation of the 1997 OMB standard on race and ethnicity. Depending on the nature of the analysis, use appropriate means for conveying the change in the data series, including separate tables or graphs, clearly demarcated breaks in trend lines, footnotes, technical notes, and explanations within the text.

Guideline: Data Presentation

We recommend following the suggested guidelines described in "Use of Race and Ethnicity in Biomedical Publication" (Kaplan MS and Bennett T. JAMA, 2003, 289:2709-2716) when presenting data by race and ethnicity.

• Due to the complexity of the issues, analysts should consider why they are analyzing health data by race and ethnic group and should articulate their reasons in their reports and presentations.
• Analysts should specify how race and ethnicity were collected (e.g., self-report using check boxes, assigned by someone else using an open-ended question.) (See Guidelines: Data Collection for specific recommendations on collecting data on race and ethnic group.)
• Race and ethnicity should not be used as a proxy for genetic variation in the absence of firmly grounded genetic evidence. Similarly, in discussing differences among racial and ethnic groups, analyst should avoid discussing differences as due to inherent underlying traits without clear evidence of such.
• In interpreting differences among race and ethnic groups, analysts should consider all conceptually relevant factors, such as socioeconomic factors, racism, and discrimination. Since lack of adjustment for socioeconomic status may be an important source of bias, analysts should try to adjust for these factors.

>>> Next >>>