Gemayel et al. (2011) demonstrated that RSS can be an effective way to reduce auditing costs when assessing the true value of an account is time consuming (and, consequently, expensive) by allowing for smaller sample sizes than would be necessary for a SRS approach with the same precision. They point out that this is commonly the case for accounts such as inventory; accounts receivable; property, plant, and equipment; and accounts payable. In these settings, the auditor will draw a sample from a set of accounts through their subsidiary ledgers, and then proceed with on-site inspections, recalculations, confirmations, and other auditing procedures when necessary.

Tackett (2012) provided simulated sales invoice records data for an electrical/plumbing distribution center, constructed in such a way that 15% of the recorded sales invoices are fraudulent, with stated book values larger than the true audited values for the materials sold in those transactions. Table 15.1 in the NSM Third Edition R Package (sample subset included here) provides the stated book values and the true audited values for a population of 12,557 such sales invoice records, with 15% (1884) of them being fraudulently recorded. (In practice, of course, all that would be available for the auditor would be the stated book values, and the true audited values would be obtained only for those accounts that were selected for inclusion in the RSS. The auditor would not need to find the audited values for the entire population. However, we have included the true audited values in Table 15.1 for all 12,557 sales invoices so that the data set can be used to illustrate the application of RSS in such a setting.)

Select an SRS of 20 invoices from the population of 15% overstated book value accounting data given in Table 15.1. Be explicit about how you choose the invoices to include in your SRS.