Statistical Sampling in Sales and Use Tax Audits – Defining Sample Population

​The department continues to improve its statistical sampling program to promote efficiency and accuracy, as reported in Wisconsin Tax Bulletin 196 (January 2017). An important part of statistical sampling is defining the sample population. This includes pre-screening, considering possible refund items, and evaluating alternative review methods.


The auditor and computer audit specialist (CAS) pre-screen the data to identify transactions that may be excluded from the sample population. For example, they might identify certain accounts or vendors that are unlikely to have taxable sales; or transactions that did not take place in Wisconsin. This generally results in a smaller sample size, less time needed to examine the source documentation, and a shorter audit.

Communication between the auditor, CAS, and taxpayer during the pre-screening process will likely result in a more efficient sample that can shorten the duration of the audit.

Considering Possible Refund Items

While the auditor and CAS will try to limit the sample population as much as possible, the taxpayer may request to include certain types of transactions because items could be taxed in error. The auditor and CAS will work with the taxpayer to include the transactions in the audit as appropriate. Options include:

  • keeping the transactions in the sample population
  • using a separate sampling class, or
  • 100% review

It is important for the auditor, CAS, and taxpayer to discuss possible refund items as early in the process as possible. If a refund claim is filed late in the audit, the auditor may not be able to include the refund claim in the audit. Instead, the taxpayer would need to appeal the audit determination.

Alternatives to Statistical Sampling

The auditor and CAS will consider alternate review methods. Generally, the statistical sample is the group of transactions for which the auditor will review source documentation. However, some types of transactions can be more efficiently reviewed by looking at other supporting detail. This may include utilities with an exempt percentage, intercompany transactions, or allocated or ratio-computed exemptions.

For example, the auditor notes that thousands of charges appear in the Stores account. The taxpayer did a detailed study five years ago of the Stores disbursements and is self-assessing use tax based on the results of the study. Instead of including the invoices from the Stores account in the statistical sample, the auditor decides to examine several months of Stores disbursements to determine if the percentages in the study need to be updated.

Transactions examined using an alternative method are typically excluded from the sample population.

The most efficient audits are those where the auditor, CAS, and taxpayer have a continuing dialog throughout the sample selection process.

For more information about statistical sampling, see Publication 516, Statistical Sampling.

July 1, 2017