M3DODSupportingStatementB.wpd

SUPPORTING STATEMENT - PART B

U.S. Department of Commerce

U.S. Census Bureau

Manufacturers’ Shipments, Inventories, and Orders to the Department of Defense

Form (M-3DOD)

(OMB Control No. 0607-xxxx)

B. Collection of Information Employing Statistical Methods

1. Description of Universe, Respondent Selection, and Response Rates

There are 16 specific NAICS (6 digit level) industries defined as being in-scope in this survey. The target population for this survey will be comprised of all active establishments with 5 or more employees classified within these industries. The U.S. Census Bureau will identify the set of eligible establishments using the U.S. Census Bureau’s final 2006 Business Register.

A probability sample of approximately 5,000 establishments will be selected. The individual establishment probabilities will be assigned based upon the establishment’s relative importance within the industry and will be independent of its inclusion in other surveys conducted by the U.S. Census Bureau. Larger establishments within an industry will have a higher probability of selection for the survey than will smaller establishments.

A similar survey, the shipments to Federal Agencies Benchmark Survey, conducted in 1992 as part of the Economic Census, achieved a final response rate of 90 percent. As a stand-alone survey with mandatory authority, the U.S. Census Bureau expects at least an 80 percent response rate with this survey.

2. Sampling Methodology and Estimating Procedures

The set of eligible establishments will be identified in the final 2006 Business Register. A probability-proportionate-to-size sample methodology will be used. Each establishment’s probability of selection will be based on a 2006-based measure-size (MOS). For establishments included in the 2006 Annual Survey of Manufactures (ASM), the “total value of shipments” from the ASM will be defined as the MOS. For the remaining establishments, a 2006 “total value of shipments” will be estimated using final 2002 Census of Manufactures micro-data or imputed using industry-specific shipment/payroll parameters.

An independent sample will be selected within each of the 16 NAICS (6 digit level) industries. Coefficient of variation targets will be set to yield an expected sample of approximately 5,000 establishments. A fixed sample size selection technique will be used to assure that the realized sample size is equal to the expected sample size.

Each selected establishment will be assigned a sample weight that is equal to the inverse of its probability of selection. For example, a selected establishment with a probability of 0.5 would have a sample weight of 2.00. Estimates will be generated by multiplying the individual establishment response data by the appropriate sample weight and summing the weighted response data across the set of respondents within an industry. The estimates will be adjusted for unit non-response.

Documentation of the statistical methods will be prepared in accordance with U.S. Census Bureau and OMB standards.

3. Efforts to Maximize Response

Companies that do not respond to the survey will receive one follow-up letter and be contacted by telephone if still delinquent. A duplicate form is sent out with the first follow-up letter. Subsequently, the U.S. Census Bureau will conduct telephone follow-up during the second and third months following mailout of the original report form, to obtain the data by telephone, and/or urge the respondent to return the form with the data, or to document why the form was not completed. All report forms will be reviewed and evaluated for any comments and suggestions made by the respondents to improve the survey results.

4. Tests of Procedures or Methods

The M3DOD survey will consist of the creation of a new form design, layout, and instructions. In order to improve usability and minimize respondents’ error, the U.S. Census Bureau conducted unstructured interviews with fourteen manufacturing establishments who are impacted by the M3DOD survey. This cognitive testing enabled the U.S. Census Bureau to understand the best approach to effectively collect the data variable within the M3DOD survey. Interviews concluded on April 30, 2008. More information on cognitive testing of the M3DOD survey form can be found in Appendix C.

5. Contacts for Statistical Aspects and Data Collection

Paul Hsen, Assistant Division Chief

Research and Methodology Branch

Manufacturing and Construction Division

U.S. Census Bureau

(301) 763-4586

Chris Savage, Branch Chief

Manufacturers’ Shipment, Inventories, and Orders Branch

Manufacturing and Construction Division

U.S. Census Bureau

(301) 763-4832

Appendices for Supporting Statement

Appendix A

• Manufacturers’ Shipments, Inventories, and Orders to the Department of Defense

Supplement: 2006-2007 form (M-3DOD) with Instructions

• Major Manufacturing Activities List (NAICS Classification Codes)

Appendix B

• Introductory Letter

• Reminder Letter

Appendix C

• Cognitive testing results