Part B OMB.wpd

0

PART B. COLLECTION OF INFORMATION EMPLOYING STATISTICAL METHODS

1. Universe and Respondent Selection

Sampling Frame: The respondents to this survey consist of all businesses in the MRTS sample and an additional panel of non-certainty businesses. The ARTS sample is a probability sample selected from retail employers (NAICS Major Groups 441- 454, 721, 722) on the U.S. Census Bureau's Business Register, which covers all employers who made social security payments for their employees under the Federal Insurance Contributions Act (FICA). The sample is updated quarterly by use of a two phase selection procedure (OMB Approval #0607-0189). Employer Identification numbers (EIN) added to the survey in the following year represent recent employer births while administrative records from the Internal Revenue Service provide estimates of retail non-employers. There are approximately 22,000 firms included in the survey. Of this number about 3,650 are large firms and were selected with certainty (probability equal to one). The remaining 18,350 firms are smaller and selected with a probability less than one.

Dollar volume response rates for data items collected in the 2006 ARTS were 94 percent for sales, 94 percent for end-of-year inventories, 75 percent for purchases, and 74 percent for accounts receivables.

For non-reporting firms, data are imputed using survey data and administrative records. It is expected that the non-response by data item for future years will be about the same as the 2006 results.

2. Procedures for Collecting Information

Information, via questionnaires, will be requested from a sample of firms operating retail and accommodation and food services establishments in the United States, with probability of selection based on their sales or inventory size. The sample will be selected from retail and accommodation and food services firms (NAICS Major Groups 441- 454, 721, 722) contained in the U.S. Census Bureau's Business Register that covers all employers who made social security payments for their employees under the Federal Insurance Contributions Act (FICA).

Before the sampling frame was available, a study was conducted of an universe consisting of 1,681,806 employer retail and accommodation food services businesses, using files from the 2002 Economic Census. This study determined the stratification of the sampling units based on 2002 sales and group and also determined the optimal allocation of the sample necessary to meet specified sampling variability objectives for sales estimates for different industry groups. The primary status boundary determined in the study was the certainty cutoff for each industry.

The sampling units consist of both large companies and EINs. If a company had sales above the corresponding certainty cutoff for its industry, the company was selected with certainty. All companies not selected with certainty were sampled on an EIN basis. If a company was comprised of more than one EIN,each EIN was treated as a separate sampling unit. The EINs were stratified according to their major industry and their estimated sales. Within each stratum, a simple random sample of EINs was selected without replacement. The maximum sampling weight for an EIN selected for the ARTS was 250.

For U.S. totals, the estimated coefficient of variation in 2006 was 0.6 percent for annual sales, 0.5 percent for end-of year inventories, and 0.7 percent for purchases. The 2006 coefficient of variation for accounts receivables was 1.7 percent.

3. Methods to Maximize Response

Firms not responding to the initial survey request are sent two follow-up letters and report forms. If they do not report by mail, the NPC in Jeffersonville, Indiana, will contact the firm via telephone in an effort to obtain the necessary data.

4. Testing of Procedures

We systematically monitor data collection procedures and methods to identify ways of reducing burden and streamlining data processing.

5. Contacts for Statistical Aspects of Data Collection

Direct questions regarding the sample design and statistical methodology used for this survey to

Carol King, Chief of the Statistical Methods Branch, U.S. Census Bureau, (301) 763-2675 or William C. Davie, Jr. ,

(301) 763-7182 , Chief, Program Research and Development Branch. Planning and implementation of this survey are under the direction of John Trimble , Chief, Annual Retail and Wholesale Indicators Branch, U.S. Census Bureau, (301) 763-7223 .

List of Attachments

1 - BEA Support Letter

2 - Copy of Report Forms

3- Sample Cover Letter