Stratification and key items in the audit file
Stratifying a population means splitting it into groups so that each group is more uniform than the whole. In a dollar-based split the largest items form a stratum the auditor tests in full (key items), and smaller items are sampled. AU-C 500 describes selecting all items when the population is made up of a small number of large items or when an item is individually significant, and AU-C 530 describes stratification as a way to improve efficiency and to direct the sample toward the larger or riskier items.
Coverage is reported two ways. Item coverage is the share of items you will touch. Dollar coverage is the share of the population value in the selected items, which is usually the number a reviewer asks about. Each stratum other than the top one is drawn at random without replacement from a seed you can record. Results from a sample must be projected to the stratum it came from, and the key-item stratum is tested with no sampling risk. How many to take from each stratum is your judgment or comes from your sampling plan.
Common questions
What is stratified sampling in an audit?
The population is divided into strata, here by dollar amount, and each stratum is tested differently. Typically all of the largest items are examined and the smaller strata are sampled, which gives high dollar coverage from a manageable number of items.
Why test every item above a threshold?
Large items carry the most risk of material misstatement and are cheap to cover by item count. Examining them all removes sampling risk for that part of the population, so the sample only has to represent the rest.
How many items should I take from each stratum?
This tool does not set sample sizes. Use your firm's sampling plan, or the audit sample size and monetary unit sampling calculators, and enter the number you decided for each stratum.
How are credits and negative amounts treated?
By absolute value, so a credit counts toward its stratum and the dollar total instead of netting against debits. Check the strata against how your population was defined.
Do I project results from each stratum separately?
Yes. Misstatements found in a sampled stratum are projected to the population of that stratum, and the projections are added to the misstatements found in the key items. The total is then compared with tolerable misstatement.
Runs in your browser. The figures you type are not sent to CORAA or stored. Check the result against your firm's methodology before relying on it.