Note to Reviewer

September 28, 2007

MEMORANDUM FOR: Reviewer of 1220-0011

FROM: Chris Manning

National Benchmark Branch Chief

Division of Current Employment Statistics

Office of Employment and Unemployment Statistics (OEUS)

Bureau of Labor Statistics

SUBJECT: Information Correction Worksheet for the Current Employment Statistics Response Analysis Survey (RAS)

Attached is the questionnaire we would like to field to a small group of active Current Employment Statistics (CES) respondents. We would ask respondents these additional questions in order to investigate reported employment differences between the CES and Quarterly Census of Employment and Wages (QCEW). We are planning to ask current CES respondents these questions between December 2007 and March 2008.

Current OMB approval of the CES survey is scheduled to expire July 31, 2008.

If you have any questions about this request, please contact Chris Manning at (202) 691-6538 or e-mail at manning.chris@bls.gov.

Note to Reviewer of 1220-0011

I.Introduction and Purpose

The Current Employment Statistics (CES) program is one of the largest surveys conducted by BLS. CES collects basic employment, payroll and hours information from a sample of over 300,000 business establishments each month. CES is a cooperative program involving the BLS and the States.

CES is closely linked to the Quarterly Census of Employment and Wages (QCEW) though a process called benchmarking. Because the CES program produces estimates based on a sample of employers, the survey is susceptible to both sampling error and nonsampling error. To correct for these errors, CES estimates are adjusted annually to full population counts derived primarily from QCEW employment counts. This procedure, referred to as benchmarking, replaces the March CES sample-based estimate with QCEW-based employment counts for March. Because the two programs are so closely linked, problems arise when businesses report different employment counts to each program.

The Response Analysis Survey (RAS) will assess the quality of the data reported to the two programs and study the prevalence, magnitude, and causes of seasonal differences between the CES and QCEW data. Findings from the RAS could result in proposals to change CES and /or QCEW program operations to account for these differences. Active CES respondents will be asked questions about their reporting patterns, sources for completing the QCEW and CES reports, and other questions related to their reporting.

The major collection medium used for this test will be the CATI instrument that will be used by our data collection center in Niceville, Florida.

Other elements that will take place during the RAS are:

• Case management

• Respondent materials (introduction letters and data sheet)

• Transmission of data to BLS

II.Respondents

There will be an estimated 3,000 reporting units that will be contacted for the Response Analysis Survey (RAS). Each respondent will be asked to complete a telephone interview that will follow the attached RAS questionnaire (which will be programmed into a CATI instrument; Attachment A). Each of the 3,000 reporting units were sampled for the RAS based on reported employment discrepancies between the QCEW and the CES. It is expected that each interview will take 15 minutes to complete per CES respondent. In approximately 25 percent of the cases, the interviewer may need to speak with another individual from the firm in order to complete the interview. For the cases where this occurs, it will add another 5 minutes to the total interview time per reporting unit.

We anticipate an overall response rate of 80 percent based on pre-testing of this current RAS and previously implemented Response Analysis Surveys.

The following table shows expected workload for the RAS between December 2007 and March 2008.

	Interview	Follow-up Interview
Number of respondents	3,000	750
Number of responses	3,000	750
Minutes needed for interview	15	5
Total Minutes	45,000	3,750
Total Minutes for all interviews	48,750
Total hours	812.50

Total Hours Response Burden is = 812.50

The estimate of costs to respondents based on burden hours to participate in this survey is $11,570. This estimate is based on an hourly pay rate of $14.24, which is the 2006 annual average of the median hourly earnings of "Human Resources Assistants, Except Payroll and Timekeeping" from the CPS survey, and was multiplied by the 812.50 burden hours.

All respondents will receive an introduction letter prior to the interview informing them of the expected time burden. The Attachment B includes a copy of the advance letter. Confidentiality assurances will be relayed to the respondent through the introduction script at the beginning of the interview.

III.Payments to Respondents

There are no payments made to the respondents for this survey.