Dominic A Smith

Census of Retail Trade Sampling Frame

For the Census of Retail Trade I have included descriptions of the sampling frame for each year below.

• 1977 - "Two basic sources were used to obtain the data for the census of retail trade. Questionaires were mailed, as a general rule, to firms with four or more paid employees and to a sample of smaller employer firms. The mail portion of the census utilized 21 different retail trade questionaires to allow for the many types of businesses covered. Federal administrative records were used the the Census Bureau to obtain information from the more than one million smaller business firms, including all retail businesses without employees. Data from the retail trade census are provided for about 100 kind-of-business classifications within the 10 major groups listed below." (pg. 13)
• 1982 - "Two basic sources were used to obtain the census data: Forty-three different questionnaires, allowing for the many types of businesses covered, were mailed to firms above a predetermined size (generally equivalent to four or more paid employees) and to a sample of smaller employer firms; information for the balance was obtained from Federal administrative records. The monthly and annual retail trade surveys are based on samples of firms with employees found in the census, supplemented with area samples of nonemployers and of recent " births" identified from administrative records during the years between censuses. Most of the data are collected by.mail. Data from the census are provided for about 100 kind-ofbusiness classifications within the SIC major groups listed below... The 1982 census tables show columns for "all establishments" as well as "establishments with payroll." Unfortunately, the Federal administrative records that were to have been used to obtain data for establishments without payroll were found to be misclassified. Therefore, data in the "all establishments" columns are limited to geographic area totals only." (pg. 17)
• 1987 - "For the 1987 Census of Retail Trade, data were gathered through a combination of questionnaires, which were sent to all but the smallest firms, and the administrative records of other Federal agencies. There were 36 variants of the census questionnaire, permitting each kind of business to answer in terms of categories relevant to its operations. Questionnaires were mailed to all firms above a certain size cutoff (generally equivalent to four or more paid employees), and to a sample of smaller firms with paid employees. Information for firms with no employees and for small firms not in the sample were obtained from Federal administrative records. as discussed in chapter 1. In addition to the regular census forms, a small sample of establishments were sent an additional questionnaire for 1987 covering operating expenses, capital expenditures, depreciable assets, and value added. The results are published as part of the census in the subject series." (pg. 25)
• 1992 - "The Bureau sent over 1 million questionnaires and classification forms to a sample of single-establishment companies and to all establishments of multiestablishment companies. For 1992, The Bureau selected a stratified sample by kind of business and geography in the mail canvass of single-establishment companies to provide estimates of merchandise-line sales and other industry specific items for the Nation as a whole and for States and metropolitan areas. The agency used administrative records to obtain information, as for 1982 and 1987, for small single-establishment companies and for all nonemployers not included in the mail sample. This census collected basic statistics such as number of establishments, sales, payroll, number of employees, and organizational status (legal form of organization). Establishments, classified in approximately 100 kind-ofbusiness categories, all sold merchandise to the general public for personal or household consumption and rendered services incidental to the sale of goods." (pg. 7)
  For the 1987 and 1982 censuses, Census used a 10-percent sample in the mall canvass of establishments below the cutoff to provide estimates of merchandise-line sales and other Industry-specific items for the Nation as a whole and for States and metropolitan areas. (pg. 7)
• 1997 - "The 1997 Economic Census initial mailout plans called for inclusion of all multiestablishment retail firms (approximately 530,000). An additional 320,000 single-establishment companies for which the Census Bureau had insufficient classification information would be sent the classification form. The remaining small singleestablishment retail trade companies were sampled at a rate of 13.8 percent, creating a sample mail file of 121,309 firms. These were sent one of 37 retail trade long questionnaires." (pg. 62)
• 2002 - "The initial mailout plans for the 2002 Economic Census called for mailing questionnaires to all multiestablishment retail, accommodations, and food services establishments. The Census Bureau mailed forms to 566,451 MU companies and 289,000 certainty case SU businesses and to a stratified sample—a further 156,000 addresses—of the remaining in-scope, but noncertainty, SUs; and to 131,000 single-establishment companies for which the Census Bureau had insufficient classification information." (pg. 75)
  "Certainty SUs: The largest SUs, with a 2002 annualized payroll greater than an industry specific cutoff, were included in the sample ‘‘with certainty.’’ The cutoffs for this certainty sample were developed for each eight-digit NAICS industry code12 within specified areas."" (pg. 75)
• 2007 The Census Bureau obtained millions of electronic records for nonemployers (i.e., establishments with no employees for whom payroll taxes are filed) and very small employer companies6 from the Internal Revenue Service, Social Security Administration new business (“birth”) records, and Bureau of Labor Statistics records during 2007 Economic Census operations.
  Generally, very small firms were single-establishment companies that had less than the specified minimum annual payroll levels. The payroll cutoff varied by industry, typically corresponding to 1 to 3 paid employees in retail firms. (pg. 48)
• 2012 - Firms in the economic census and related programs are divided into those sent report forms and those not sent report forms. For most economic census sectors and programs, all large- and medium-size firms and all multi-establishment firms were sent report forms to be completed for each of their establishments and returned to the Census Bureau. For most sectors and programs, report forms were also mailed to a sample of small employers (those with paid employees). This sample consists of single-establishment firms with payroll below a specified cut off. This cut off and the sampling rate, varies by economic census sector, industry and geography. However, for most very small firms, data from existing administrative records of other federal agencies are used. These records provide basic information for the business, including data on sales, payroll, number of employees, legal form of organization and other statistics. Other industry-specific census statistics are estimated only from those establishments that receive and return a report form.