Documentation for Bridged-Race Vintage 2006
(July 1, 2000 - July 1, 2006)
Postcensal Population Estimates
for Calculating Vital Rates
On August 16, 2007, the National Center for Health Statistics released the Bridged-race Vintage 2007 postcensal population file.
The bridged-race Vintage 2006 postcensal population estimates file contains estimates of the resident population of the United States as of July 1, 2000 to July 1, 2006, by county, single-year of age (0, 1, 2,..., 85 years and over), bridged-race category (White, Black or African American, American Indian or Alaska Native, Asian or Pacific Islander), Hispanic origin (not Hispanic or Latino, Hispanic or Latino), and sex (1). The estimates on this file are based on Census 2000 and result from bridging the Vintage 2006 postcensal estimates with 31 race groups (the 31 race groups used in Census 2000 in accordance with the 1997 Office of Management and Budget (OMB) standards for the collection of data on race and ethnicity) to the four race categories specified under the 1977 OMB standards. The bridged-race postcensal estimates were produced by the Population Estimates Program of the U.S. Census Bureau in collaboration with the National Center for Health Statistics (NCHS). This file was released by NCHS on August 16, 2007.
Background
Race bridging refers to making data collected using one set of race categories consistent with data collected using a different set of race categories, to permit estimation and comparison of race-specific statistics at a point in time or over time. More specifically, race bridging is a method used to make multiple-race and single-race data collection systems sufficiently comparable to permit estimation and analysis of race-specific statistics.
In 1997, OMB issued "Revisions to the Standards for the Classification of Federal Data on Race and Ethnicity," which supersedes the 1977 "Statistical Policy Directive 15, Race and Ethnic Standards for Federal Statistics and Administrative Reporting" (2,3). Both documents specify rules for the collection, tabulation, and presentation of race and ethnicity data within the Federal statistical system. The race categories specified in both standards represent a social-political construct and are not anthropologically or biologically based. The revised standards increased the minimum number of race categories to be used by Federal agencies from four (White, Black, American Indian or Alaska Native, and Asian or Pacific Islander) to five (White, Black or African American, American Indian or Alaska Native, Asian, and Native Hawaiian or Other Pacific Islander). In addition, the revised standards require Federal data collection programs to allow respondents to select more than one race category when responding to a query on their racial identity. This provision means that under the revised standards there are potentially 31 race groups (5 single-race and 26 multiple-race), depending on whether an individual selects one, two, three, four, or all five of the race categories.
During the transition to full implementation of the 1997 standards, these two different standards for the collection of race and ethnicity data are both being used, creating incomparability across data systems. Further, within a given data system, the change in the race standards results in incomparability across time, thus making it difficult to perform trend analyses. The OMB recognized that race-bridging approaches would be needed to make race data collected under the 1997 standards comparable to race data collected under the 1977 standards. Therefore, the OMB issued "Provisional Guidance on the Implementation of the 1997 Standards for Federal Data on Race and Ethnicity" (4). The guidance document contains a detailed discussion of various bridging methods.
Vital rates are based on information obtained from vital records collected through the state-based Vital Statistics Cooperative Program (numerators) and population estimates derived from the U.S. Census (denominators). Because of differences in the timing of implementation of the 1997 standards, beginning with the 2000 data year, the numerators and denominators of vital rates have incompatible race data. The question on race on the 2000 census was based on the revised OMB standards and so allowed respondents to select more than one race category. As a result, population estimates for 2000 and beyond have five single-race categories and up to 26 multiple-race categories. Implementation of the 1997 standards within the Vital Statistics Cooperative System started in 2003, on an individual state basis, and is expected to proceed slowly as states implement revised birth and death certificates which incorporate the revised OMB standards. As of 2006, most states had not revised the race question on their birth or death certificates and were still collecting race data using the 1977 race categories. Therefore, at this time, the calculation of post-2000 race-specific birth and death rates requires population estimates with the 1977 race categories. When a sufficient number of states have adopted the revised birth and death certificates, rates will be presented using population estimates that comply with the 1997 standards.
NCHS Regression Bridging Method
Bridging methodology developed by NCHS bridges the multiple-race group population counts to single-race categories (5, 6). Information from the pooled 1997-2000 National Health Interview Survey was used to develop the bridging methodology. Regression models with person-level and county-level covariates were used to generate the probability of selecting each single-race category possible for a multiple-race group. The probabilities generated from the fitted regression models are referred to as the NHIS bridging proportions. The Census Bureau applied the NHIS bridging proportions generated by NCHS to the Census 2000 Modified Race Data Summary file (7). This application resulted in a bridged population count for each of the four single-race categories (White, Black or African American, American Indian or Alaska Native, and Asian or Pacific Islander) by county, single-year of age, Hispanic origin group, and sex, for April 1, 2000. The bridging methodology is described in detail in the report, "United States Census 2000 Population with Bridged Race Categories" (5).
Bridged-race Postcensal Population Estimates
Postcensal population estimates are estimates made for the years following a census, before the next census has been taken. Postcensal estimates are derived by updating the resident population enumerated in the decennial census using various measures of population change. The components of population change used in the derivation of the postcensal estimates include: births to U.S. resident women, deaths to U.S. residents, net international immigration, net movement of U.S. Armed Forces and civilian citizens of the U.S, and migration within the U.S. The Census Bureau annually produces a series of postcensal estimates that includes estimates for the current data year and revised estimates for earlier years. Estimates for earlier years in a given series are revised to reflect changes in the components of population change data sets (for example, a preliminary natality file is replaced with a final natality file). The last year in a series is used to name the series. For example, the Vintage 2002 postcensal series has estimates for July 1, 2000, July 1, 2001, and July 1, 2002. The Vintage 2003 series has estimates for July 1, 2000, July 1, 2001, July 1, 2002, and July 1, 2003. The July 1, 2000, July 1, 2001, and July 1, 2002 estimates from the Vintage 2002 and Vintage 2003 series differ.
To date, the Census Bureau has produced the Vintage 2001, Vintage 2002, Vintage 2003, Vintage 2004, Vintage 2005 and Vintage 2006 series of postcensal estimates of the July 1 resident population of the United States using the Census 2000 Modified Race Data Summary File as the base data for the series (7). These series initially had estimates for 31 race groups, in accordance with the 1997 race and ethnicity standards (2). Under a collaborative arrangement with NCHS, the Population Estimates Program of the U.S. Census Bureau applied the NHIS bridging proportions to the 31-race postcensal population estimates to produce bridged-race postcensal estimates (estimates for the four single-race categories: White, Black or African American, American Indian or Alaska Native, and Asian or Pacific Islander).
Variance of Bridged-race Population Estimates
Population estimates generally are assumed to be fixed and do not contribute to the variance of rates. However, this is not true for bridged-race population estimates. Methodology to compute variances for bridged-race population estimates has been developed (8).
Geographic Code Changes for Vintage 2006 Postcensal Estimates
The Vintage 2006 files have the same State and county FIPS codes as the Vintage 2005 files. There are two differences in the county codes that appear on the Vintage 2005 and Vintage 2006 files and those that appear on earlier vintages.
- There are estimates for Broomfield County, Colorado (FIPS code=08014) on the bridged-race Vintage 2005 and Vintage 2006 postcensal files. Broomfield County, Colorado was created effective November 15, 2001 from parts of four Colorado counties: Adams, Boulder, Jefferson, and Weld. The vintages prior to Vintage 2005 did not contain estimates for Broomfield. Note that data for Broomfield County does not appear on NCHS birth or mortality files until data year 2003.
- There are no estimates for Clifton Forge County, Virginia (FIPS code=51560) on the bridged-race Vintage 2005 and Vintage 2006 postcensal files. On July 1, 2001, Clifton Forge city, Virginia, formerly an independent city, merged with Alleghany county (FIPS code=51005). The vintages prior to Vintage 2005 did contain estimates for Clifton Forge. Note that estimates for Clifton Forge city appear on NCHS birth and mortality files prior to data year 2003; beginning with the 2003 data year, no data for Clifton Forge city appear on the birth and death files.
Release of Bridged-race Population Estimates
In response to the need for bridged estimates by a wide range of users, NCHS is making the bridged-race population estimates available for download from the Population Estimates web site (see http://www.cdc.gov/nchs/about/major/dvs/popbridge/popbridge.htm, look under "Datasets and Documentation"). The report detailing the bridging methodology is available for download from this site (see United States Census 2000 Population with Bridged Race Categories under "Methodology").
NCHS is currently using the bridged-race postcensal population estimates to calculate birth and death rates. Previously published reports that used 1990-based postcensal population estimates to calculate rates for 2001 have been re-issued in whole or in part; new reports use the bridged-race estimates).
Although efforts were made to use the best available data and methods to produce these estimates, the modeling process introduces error into the estimates. The potential for error will be greatest for the smallest population groups, particularly the smaller race groups and county level estimates. The postcensal estimates are updated annually as additional data become available, for use in the components of change model. In addition, the bridged-race estimates may be revised periodically to reflect changes made to the bridging process.
NCHS would appreciate receiving feedback on the usefulness of the estimates as well as notification of any problems that have been identified. Please provide comments via e-mail to: PopEst@cdc.gov.
Suggested citation
National Center for Health Statistics. National Center for Health Statistics. Estimates of the July 1, 2000-July 1, 2006, United States resident population from the Vintage 2006 postcensal series by year, county, age, sex, race, and Hispanic origin, prepared under a collaborative arrangement with the U.S. Census Bureau. Available on the Internet from: http://www.cdc.gov/nchs/about/major/dvs/popbridge/popbridge.htm. August 16, 2007.
References
- National Center for Health Statistics. Estimates of the July 1, 2000-July 1, 2005, United States resident population from the Vintage 2005 postcensal series by year, county, age, sex, race, and Hispanic origin, prepared under a collaborative arrangement with the U.S. Census Bureau. Available from: http://www.cdc.gov/nchs/about/major/dvs/popbridge/popbridge.htm. August 16, 2006.
- Office of Management and Budget. Revisions to the standards for the classification of Federal data on race and ethnicity. Federal Register 62FR58781-58790, October 30, 1997. (Available here.)
- Office of Management and Budget. Race and ethnic standards for Federal statistics and administrative reporting. Statistical Policy Directive 15, May 12,1977. (Available here.)
- Office of Management and Budget. Provisional guidance on the implementation of the 1997 standards for the collection of Federal data on race and ethnicity. December 15, 2000. http://www.Whitehouse.gov/omb/inforeg/r&e_guidance2000update.pdf. (Also available here.)
- Ingram DD, Parker JD, Schenker N, Weed JA, Hamilton B, Arias E, Madans JH. United States Census 2000 population with bridged race categories. Vital Health Stat 2 (135). Hyattsville, Maryland: National Center for Health Statistics. 2003. (Available here.)
- Parker JD, Schenker N, Ingram DD, Weed JA, Heck KE, Madans JH. Bridging between two standards for collecting information on race and ethnicity: an application to Census 2000 and vital rates. Public Health Reports 119 (2), p. 192-205. 2004. (Available here.)
- U.S. Census Bureau. Census 2000 Modified Race Data [MR(31)-CO.txt], prepared by the U.S. Census Bureau, 2002. Available on the internet at: http://eire.census.gov/popest/estimates_dataset.php#mrd. (Also available here.)
- Schenker, N (2003). "Assessing variability due to race bridging: application to Census counts and vital rates for the Year 2000," J American Statistical Association, 98, 818-828.
- Ventura SJ, Hamilton BE, Sutton PD. Revised birth and fertility rates for the United States, 2000 and 2001. National vital statistics reports; vol. 51 no. 4. Hyattsville, Maryland: National Center for Health Statistics. 2003. (Available here.)
File layout for the Bridged-Race Vintage 2006 Postcensal Files, 2000-2006
There is one file for the full Vintage 2006 postcensal series with data for all seven years in the series (July 1, 2000 to July 1, 2006) and two files with only the Vintage 2006 estimates for July 1, 2006. The files contain bridged-race postcensal estimates of the July 1 resident population of the United States by year, county, single-year of age (0, 1,..., 85 years and over), bridged-race category (White, Black or African American, American Indian or Alaska Native, Asian or Pacific Islander), Hispanic origin (not Hispanic or Latino, Hispanic or Latino), and sex. There is a record on the files for each combination of county, age, race, sex, and Hispanic origin. The population estimates on the files were derived by the U.S. Census Bureau using the Census 2000 Modified Race Data Summary File as the base file (9). The files were released by NCHS on August 16, 2007.
Control totals for Vintage 2006 data files:
Control total File name Number of records* Month, Year Population count pcen_v2006.txt 4,323,392 July 1, 2000 282,216,952 July 1, 2001 285,226,284 July 1, 2002 288,125,973 July 1, 2003 290,796,023 July 1, 2004 293,638,158 July 1, 2005 296,507,061 July 1, 2006 299,398,484 pcen_v2006_y06.txt 4,323,392 July 1, 2006 299,398,484 pcen_v2006_y06.sas7bdat 4,323,392 July 1, 2006 299,398,484* One record for each county, race, sex, Hispanic origin, and age combination.
File Layout for pcen_v2006.txt:
Location Field size Item and Code Outline Format 1-4 4 Series vintage (2006) Numeric 5-6 2 FIPS State code Numeric 7-9 3 FIPS county code Numeric 10-11 2 Age Numeric (0, 1, 2,..., 85 years and over) 12 1 Race-sex Numeric 1=White male 2=White female 3=Black male 4=Black female 5=American Indian or Alaska Native male 6=American Indian or Alaska Native female 7=Asian or Pacific Islander male 8=Asian or Pacific Islander female 13 1 Hispanic origin Numeric 1=not Hispanic or Latino 2=Hispanic or Latino 14-21 8 Population count July 1, 2000 Numeric 22-29 8 Population count July 1, 2001 Numeric 30-37 8 Population count July 1, 2003 Numeric 46-53 8 Population count July 1, 2004 Numeric 54-61 8 Population count July 1, 2005 Numeric 62-69 8 Population count July 1, 2006 Numeric
File Layout for pcen_v2006_y06.txt:
Location Field size Item and Code Outline Format 1-4 4 Series vintage (2005) Numeric 5-6 2 FIPS State code Numeric 7-9 3 FIPS county code Numeric 10-11 2 Age Numeric (0, 1, 2,..., 85 years and over) 12 1 Race-sex Numeric 1=White male 2=White female 3=Black male 4=Black female 5=American Indian or Alaska Native male 6=American Indian or Alaska Native female 7=Asian or Pacific Islander male 8=Asian or Pacific Islander female 13 1 Hispanic origin Numeric 1=not Hispanic or Latino 2=Hispanic or Latino 14-21 8 July 1, 2006 Population count Numeric
The SAS data set, pcen_v2006_y06.sas7bdat
Variable Item and Code Outline Format VINTAGE Series Vintage (2005) Numeric ST_FIPS State FIPS code Numeric CO_FIPS County FIPS code Numeric AGE Age Numeric (0, 1, 2, ..., 85 years and over) RACESEX Race Numeric 1=White male 2=White female 3=Black male 4=Black female 5=American Indian or Alaska Native male 6=American Indian or Alaska Native female 7=Asian or Pacific Islander male 8=Asian or Pacific Islander female SEX Sex Numeric 1=male 2=female HISP Hispanic origin Numeric 1=not Hispanic or Latino 2=Hispanic or Latino Pop July 1, 2006 resident population Numeric
Source:
Documentation for bridged-race postcensal Vintage 2006 population estimates for July 1, 2000-July 1, 2006, which was released on August 16, 2007 is on the internet at ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/datasets/nvss/bridgepop/DocumentationBridgedPostcenV2006.doc.