You decided to use attributes sampling in your testing of Frederics Clothing’s sales procedures for the year ended July 31, 198X. The client has a straightforward system that in previous years has yielded excellent results with conventional testing. Of course, the fact that July and August were the only two months ever tested could account for part of the excellent results. Recently, however, the gross margin for Frederics has somewhat declined. Although a cause has not yet been determined, there is concern that some shipments are unintentionally not being billed. For every duplicate invoice, a shipping document is attached. In addition, there is a separate file of prenumbered shipping documents filed in numerical order. There is a formal price list that can be deviated from only with an authorized signature from Bill Newshirt. George Sloppyslacks authorizes credit by initialing the sales document before shipment. There are 2,340 sales invoices for the year, beginning with invoice number 18721 and ending with 21060. There are 3,590 shipping documents, because some shipments are combined for billing. The shipping documents begin with 2893 and end with 6482.
Required:

a. Define the sampling unit to test for unbilled shipments. Why did you use that sampling unit ?

b. Assume you decide to test for validity a sample of sales invoices.
(i) Can you use the same random numbers ?
(ii) How many “attributes sampling data sheets” do you need—one or two ?
(iii) How many worksheets do you need—one or two ?
(iv) How can you organize this problem to minimize the effort and still do an adequate job ?