Current Employment Statistics Preliminary Benchmark (National) Technical Note

Technical Note

The preliminary March revision amounts presented in this 
release are intended to give a preview of the revisions 
to the Bureau of Labor Statistics Current Employment 
Statistics (CES) data released in February of each year.
Currently published CES data have not been adjusted to new 
levels based on these preliminary revision amounts. 

The preliminary March revisions are calculated by 
comparing not seasonally adjusted CES employment for March 
of the benchmark year to universe employment counts derived 
mostly from Quarterly Census of Employment and Wage (QCEW) 
data. Additional employment not covered by the QCEW is 
calculated from several sources including County Business 
Patterns and Annual Survey of Public Employment & Payroll 
data from the Census Bureau, state employment data, and 
data from the Railroad Retirement Board. This noncovered  
employment is added to QCEW employment to arrive at a universe 
employment total. The preliminary March CES revision amounts are 
the universe employment minus the CES estimated employment. 
More information about CES benchmarking is available in 
the Handbook of Methods in the Calculations section on 
Benchmarking, www.bls.gov/opub/hom/ces/calculation.htm#benchmark. 
More information about noncovered employment is available 
in the Handbook of Methods in the Calculations 
section on Noncovered employment, 
www.bls.gov/opub/hom/ces/calculation.htm#noncovered-employment.

As with all surveys, the CES survey is subject to several types
of error including sampling and nonsampling error. Sampling error 
measures how different the sample is from the population. 
Nonsampling error includes frame error (how different the 
population is now from when the sample was drawn), non-response 
error (how different those units who responded to the survey are 
from those units that did not respond), and response error (how 
different responses were to the survey versus the population for 
units that responded to both). Nonsampling errors can come from 
either the survey data or the population data. In the case of the 
CES survey, the benchmark revision is the difference between two 
independently derived employment counts, each subject to its own 
error sources.

Universe employment for March is calculated and compared to the 
CES estimates at the basic cell level and then aggregated up to 
the major industry sector, total private, and total nonfarm levels. 
More information about aggregating CES data is available in the 
Handbook of Methods in the Calculations section on Aggregation 
procedures, www.bls.gov/opub/hom/ces/calculation.htm#aggregation-
procedures.

If you are deaf, hard of hearing, or have a speech disability, 
please dial 7-1-1 to access telecommunications relay services.