With the publication of the July 2014 first preliminary estimates, the Current Employment Statistics (CES) program began implementing new sample units into production on a quarterly basis. Previously, CES had implemented new sample units into production on an annual basis.
Each year the Bureau of Labor Statistics (BLS) draws the private-sector probability sample for the CES survey from the BLS Longitudinal Database (LDB) of employer records. The probability sample is made up of continuing business units (units that were enrolled in previous years) and new business units (units that have yet to be enrolled). For business units in previous samples but not the current sample, BLS stops collecting payroll data. For newly selected business units, data collection specialists begin enrollment efforts immediately after a sample is selected and continue to enroll sample units throughout the year. The collection of payroll data generally begins in the first month after enrollment. However, new sample units are enrolled over a one-year period to minimize costs. With this approach at least a full year would pass between the sample draw and sample implementation.
CES implemented new sample units into production for all industries once a year; new sample implementation timing differed slightly for national and state and metropolitan area estimates. For national estimates, the new sample was used for the first time to estimate January first preliminary estimates, December second preliminary estimates, and November final estimates. For state and metropolitan area estimates, the new sample was implemented with January preliminary estimates, but no prior months estimates were impacted. For example, the private probability sample selected from the 2012 LDB was enrolled during 2013 and was implemented into production with the January 2014 first preliminary estimates.
The implementation of new sample for all industries simultaneously meant that newly enrolled respondents that started reporting payroll data immediately after the sample draw had provided useful data for almost a year before the data are used to produce CES estimates.
The annual implementation schedule also contributed in part to revisions in national CES estimates between the November second preliminary and final releases and between the December first and second preliminary estimates.
In the past, implementation of new sample units into the CES survey took a large amount of resources and time. CES has been updating processes for several years to improve the efficiency of sample updates and researching the effects of this change on the estimates. Implementation of new sample units every quarter means that units not in the new sample can be dropped quarterly as well. BLS expects this quarterly implementation and drop process to increase monthly response rates, decrease the impact of sample attrition, and reduce respondent burden.
Quarterly sample implementation has not changed the sample design or sample enrollment process. The full sample will continue to be drawn once a year, with a third quarter birth update, for all industries. A portion of the new sample will be enrolled each month. The change is to the timing of incorporating new sample units into monthly estimate production. More information about the sample design can be found in the sample design section of the CES Technical Notes at www.bls.gov/web/empsit/cestn.htm#section1a.
CES implemented new sample units into the estimates on a quarterly basis with the publication of July 2014 first preliminary estimates. Under the new sample implementation plan all industries are classified into four groups that began enrollment and collection during specific quarters after the sample was drawn for the year. Enrollment and data collection procedures began during a specific quarter for each group of industries, and in the first reference month of the following quarter, the newly enrolled sample units are used in estimation. For example, Group 1 was enrolled and sample collection started during the first quarter, January through March; then, for the April reference month, Group 1 sample was used for first preliminary estimates. Each reference month was estimated using the same assigned sample units from the estimation of the first preliminary estimates through the final estimates. The number of actual sample receipts will increase from first to second to final estimates as is currently the case.
Each quarterly sample implementation group is described below.
|Group||Timing||Major Industry Sector|
|Enrollment||Used in CES Estimates|
|Group 1||First Quarter||April Preliminary (1)||Mining and logging|
|Transportation and warehousing|
|Group 2||Second Quarter||July Preliminary (1)||Construction|
|Leisure and hospitality|
|Group 3||Third Quarter||October Preliminary||Information|
|Professional and business services|
|Group 4||Fourth Quarter||January Preliminary||Manufacturing (durable and nondurable goods)|
|Education and health services|
|Birth units for all private industries sampled from the third quarter of the LDB that did not exist on the first quarter of the LDB.|
Back to Top
Last Modified Date: August 1, 2014