Green Jobs OMB Supporting Statement Part B

11/18/2010

Supporting Statement for Request for OMB Approval

For Occupational Employment Statistics Related to Green Practices Pre-testing

B. COLLECTION OF DATA EMPLOYING STATISTICAL METHODS

1a. Universe

BLS will carry out field testing on a sample of establishments included on the Bureau of Labor Statistics’ Quarterly Census of Employment and Wages (QCEW). The scope of this field testing will be all industries in the 50 States plus the District of Columbia. The QCEW will serve as the base frame and will be supplemented by a targeted sub-population list obtained primarily from the GenGreen, or similar, database.

The source of QCEW data is the Quarterly Contribution Reports submitted to State Workforce Agencies (SWAs) by employers subject to State Unemployment Insurance (UI) laws. The QCEW universe will consist of about 6.5 million private sector establishments. The QCEW data, which are compiled for each calendar quarter, provide information on business name and address along with employment and wages at the six-digit NAICS level for employers subject to State UI laws.

GenGreen is a private trade organization that provides business listings of environmentally friendly businesses in North America. The GenGreen database is expected to contain about 75,000 businesses. The GenGreen list will be further supplemented by a listing of about 1,000 environmentally friendly manufacturing businesses obtained through a targeted internet search.

BLS will attempt to match establishments in the targeted sub-population lists to the QCEW to obtain industry and employment size information and minimize the likelihood of the same business being selected and contacted more than once. If the match rates are very high, BLS will use the targeted sub-population lists to produce a single frame with refined stratification. If the match rates are lower, BLS will use the targeted sub-population lists to implement a multiple-frame sample design.

1b. Sample size

As the measurement of employment within establishments related to green practices is breaking new ground, BLS expects that multiple iterations of the basic questionnaire will need to be field-tested. The initial instrument will be tested, and then based on the findings it will be modified for retesting. The sample size for the field-testing of the different form designs will be in 4 panels covering a period up to 8 months. Larger samples will be selected at the beginning and end of the field testing. The first panel needs to begin the examination of issues. The fourth panel needs a certain amount of statistical power to finalize the form. The second and third panels will be used to modify and fine-tune the survey’s questions. In addition, a web collection instrument will be tested with a fifth panel.

Sample Sizes for Field-Testing
Panels	Total Sample	Estimated Respondents (70%)
First	715	500
Second	571	400
Third	428	300
Fourth	715	500
Fifth	714	500

BLS anticipates that the first panel will have approximately 75 percent of the sample allocated to the GenGreen sub-population and 25 percent to other units in the QCEW. Allocation of sample for subsequent panels will be contingent on information obtained from the first panel, including information on the prevalence of green job activity in the QCEW.

Approximately 50 establishments in each panel will also be contacted for a follow-up interview. Establishments selected previously for the Green Goods and Services feasibility field-testing samples will be excluded from these samples.

Field-testing will identify issues related to data collection of employment and wages for jobs related to environmentally friendly production processes within the establishment. The detailed objectives are given in Part A. Non-response prompting and edit reconciliation interviews will be conducted on each panel. In addition, follow-up interviews of respondents and non-respondents will also be conducted on each panel to gain feedback on the collection form used.

BLS expects that about 5 percent of the sample will be out-of-business or out-of-scope and that an adjusted response rate by mail and follow up will be 70 percent (the out-of-business and out-of-scope establishments will be removed from the denominator in the adjusted response rate calculation). While similar research conducted under clearance number 1220-0181 for Green Goods and Services Sector Pretesting yielded lower response rates than the expected 70 percent rate for this research, BLS believes through the results of its cognitive research conducted under OMB Clearance number 1220-0141 that the targeted population selected from the GenGreen frame will produce higher response rates than the general public due to the nature of the survey.

2a. Sample Design

The QCEW frame will be stratified by industry and employment size (5 size classes listed below). All industries except private households are in-scope for this research. The five size classes are determined on the establishment’s maximum employment over the last 12 months--employment 1-19 (size 1); 20-49 (size 2); 50-99 (size 3); 100-499 (size 4); 500 + (size 5). Any given fixed sample size (e.g., 428, 571, or 715) will be divided by the number of strata (major industry sectors times 5 size classes). In addition, per the description in Section B.1.a, information from the GenGreen list and the BLS composed list may be used to provide further refinement of the QCEW frame; or may provide the basis for multiple frame sampling. Establishments in each stratum will be selected randomly. This algorithm may be modified for panels 2, 3, 4, and 5 based on the findings from previous panels.

Samples will be selected to provide sufficient information to finalize the appropriate questionnaire regarding data collection on green processes and practices within establishments. This field testing study is not designed to provide reliable estimates of green occupational employment and wages.

2b. Estimation Procedure

Analysis will be performed using the frequency distributions for each question and tabulation of aggregated totals of the information collected for each micro record. Unit and item non-response analysis will also be performed. A comparison of respondent and non-respondent groups on certain establishment characteristics (e.g., size, industry, location, and occupations covered) may also be conducted along with related logistic regression analyses. The analysis will examine what type of information is available; who at the establishment is the most likely source for these data; patterns of response and non-response; what can be done to improve the questionnaire, response rates, and reduce respondent burden; and any special issues that need to be addressed.

2c. Reliability

As mentioned above, the primary purpose of this field testing is to develop a questionnaire and web instrument that can be used to collect occupation and wage information on green jobs at the establishment. The field testing is not designed to produce reliable estimates of green occupational employment and wages.

2d. Revisions

The questionnaire/form and follow-up interview script will require revisions after each testing. At this time, the magnitude of revisions to questionnaire/forms is unknown.

2e. Specialized Procedures

None.

2f. Data Collection Cycles

The forms design and panel testing will be conducted on a sample of up to 3,143 establishments, divided into five panels (715, 571, 428, 715, and 714) and questionnaires/forms will be tested and refined based on the responses from each panel. The refined forms will be sent to the new panel and iterative procedures will be used for finalizing the forms. The process will take up to 8 months with 5 panels in FY2010 and 2011. The final panel of 714 establishments will be testing a web collection instrument for impact on response rates and respondent burden. Total respondent burden will be about 858 hours in FY2010 and FY2011 with an assumption that the forms completion interviews will each take an average of 20 minutes, and the follow-up interviews will each take about 30 minutes to complete (see table in Section 12 of Part A).

3. Methods to Maximize Response Rates

BLS will provide the employer with a pledge of confidentiality, an explanation of the importance of the survey research, and the need for voluntary cooperation for maximizing the response rate. Non-response will be minimized through non-response prompting which will be developed along with the forms. The unit and item response rate will be maximized by follow-up telephone calls.

4. Tests

The survey’s questionnaires will be developed using cognitive design techniques. A contractor will conduct 200 cognitive interviews to better understand the collection environment. OMB clearance for this work was requested and approved separately under the BLS Cognitive and Psychological Research Control Number 1220-0141. These interviews will focus on learning what information is available from firms and how best to formulate questions to collect information that firms maintain on their environmentally friendly activities and employees involved in those activities. BLS will use this information to guide the initial design of the survey questionnaires and minimize the burden on potential respondents in the future surveys. The tests will be conducted in accordance with OMB guidelines.

5. Statistical and Analytical Responsibility

Ms. Shail Butani, Chief, Statistical Methods Division of the Office of Employment and Unemployment Statistics, and Mr. George Stamas, Chief, Division of Occupational Employment Statistics, Office of Employment and Unemployment Statistics, are responsible for the statistical and analytical aspects of the OES program respectively. Ms. Butani can be reached on 202-691-6347. Mr. Stamas can be reached on 202-691- 6350. BLS will conduct this survey using the help of an outside contractor.

Attachments

1. Draft solicitation letter

2. Draft data collection form for panel testing

3. Draft Script for non-response prompting

4. Draft Non-response prompting postcard

5. Draft RAS script for non-respondents

6. Draft RAS script for respondents

4