Notes on Using Current Population Survey (CPS) Subnational Data

1. Sampling error of estimates. CPS estimates are based on a sample of the population, rather than a complete count. Therefore, they may differ from the figures that would have been obtained if it had been possible to take a complete census using the same questionnaire and procedures. This is particularly true for estimates of small groups or in states or areas with small sub-samples. (The national sample consists of about 60,500 eligible households, but state sub-samples, the size of which depends on a number of factors, including the state’s population, range from about 600 to 4,000 households.) The sampling error tables in the appendix of the Geographic Profile of Employment and Unemployment publications provide approximations of sampling errors for labor force, employment, and unemployment estimates for each area for the relevant year. Tables 1, 12, and 24 (table 23 for years before 1999) display sampling errors for unemployment rates. For more details, including the error tables, see www.bls.gov/opub/gp/laugp.htm.



2. Differences between CPS and LAUS estimates. Due to differing methodologies, CPS estimates may differ, perhaps substantially, from annual average employment and unemployment estimates produced in the Local Area Unemployment Statistics (LAUS) program. This is especially true for metropolitan areas and cities. However, both the CPS and the LAUS program use the same concepts and definitions, including measuring persons, rather than jobs, by place of residence. The CPS can be used to develop labor force estimates only where the sample is present and sufficient in size to ensure a given level of reliability. (The reliability of metropolitan area and city data is substantially less than the statewide reliability.) Currently, CPS substate data are published for only 50 large metropolitan areas and 17 large cities, and only on an annual average basis. The official LAUS data, on the other hand, are available on both monthly and annual average bases, and are produced utilizing widely available data inputs, for about 7,200 areas nationwide. These areas include all counties, cities with a population of at least 25,000, and all New England cities and towns, in addition to regions, states, and metropolitan areas (including those for which CPS estimates are published). The LAUS data, which are used by Federal programs that allocate funds or determine program eligibility, constitute a very large set of information that allows comparisons across thousands of diverse areas throughout the Nation.



3. Absence of “level” estimates for substate areas. CPS estimates for metropolitan areas and cities for 1992 and later include only rates and ratios—such as labor force participation rates, employment/population ratios, and unemployment rates, as well as percent distributions—but not levels (that is, numbers of persons). The absence of levels is a result of the lack of area-specific independent population controls, which itself follows from the inadequate reliability of substate estimates of the 1990 Census undercount. National and statewide data were adjusted for the estimated census undercount based on the Post-Enumeration Survey, but that survey did not yield sufficiently reliable data at the area level other than for Los Angeles County and New York City.



4. Lack of subnational population controls for demographic groups. Apparent differences in estimates between and among demographic groups—particularly in estimates of levels (that is, numbers of persons)—may not reflect real differences, especially prior to 2003. Instead, they may result solely from differences or changes in the demographic composition of the sample due to chance, combined with the interaction of the national and state population controls used in the estimation procedure. Population controlling for subnational areas incorporates several sets of independent demographic population estimates at the national and state levels. This approach can result in a possible source of error in subnational demographic components beyond the sampling error. On balance, the introduction in 2003 of independent population estimates for some demographic groups at the state level, though more limited and generally less reliable than those available at the national level, has improved the reliability and comparability of demographic population and labor force estimates relative to earlier years, when controls were available for only the state total population.



5. Updating of population controls. CPS subnational detailed data are generally based on the annual average population estimates that were available at the end of the year to which the data pertain—and not any subsequent revisions. In particular, the detailed data in the hardcopy publications of Geographic Profile of Employment and Unemployment are not updated, even though the CPS and official LAUS total labor force, employment, and unemployment annual average levels for states typically are reestimated for the previous three to five years when updated population estimates become available from the Census Bureau. Differences in CPS levels as a result of new population controls are generally less than one percent in the first revision, but may grow with subsequent re-controlling. Application of new population controls generally does not affect ratios or proportions, such as the unemployment rate, because all levels for a given state are scaled by the same factor, representing the degree of revision in the state’s population estimate.



6. Changes in metropolitan area definitions. Estimates for many metropolitan areas are affected by changes in the geographic definitions issued by the Office of Management and Budget. Metropolitan area definitions based on the 1980 Census were first reflected in CPS data published for 1986 and were used through 1994. Geographic definitions based on the 1990 Census were used for CPS data published beginning with 1995. Thus, data that span either 1985-86 or 1994-95 can change simply as a result of increases or decreases in the geographic scope of the area. (For the history of metropolitan area definitions, see www.census.gov/population/www/estimates/metrodef.html. Similar information for the affected areas is available in appendix table C-2 of the 1986 and 1995 editions of Geographic Profile of Employment and Unemployment.) Geographic definitions based on the 2000 Census will be implemented with the publication of 2005 data.



7. Impact of CPS redesign. Due to a major questionnaire redesign and a fundamental change in the data collection mode, introduced in January 1994, estimates for years prior to 1994 may not be fully comparable with those for later years. See www.bls.gov/ore/pdf/ec950090.pdf for an analysis of these differences, including which data items were most affected by the redesign.



8. Regional CPS data vs. the sum of statewide data. For 1999 and earlier years, summing the CPS estimates for states that comprise a census region or division may not yield exactly the published estimates for the region or division, due to differential weighting and rounding. CPS region and division estimates were created separately, and the weights for those large areas were not precisely the sum of the state weights. Differences in estimates between the larger geographic areas and the sum of component states, if present, were quite small, generally no more than a few thousand persons.



9. Additional information. Users of subnational CPS data are encouraged to read the preface and appendixes in the Geographic Profile of Employment and Unemployment publication.

Last Modified Date: September 8, 2006