Inside Assurance Toolkit #10 — Audit Sampling Working Paper

Audit Sampling Working Paper: Document the Decision Behind the Selection

A list of selected transactions is not a sampling rationale. A useful working paper explains the objective, the population, why those items were selected, what was tested and how far the conclusion can reach.

Inside Assurance Toolkit #10 presents an eight-page carousel with a reusable fillable working-paper page. Its fictional stock-adjustment case demonstrates documentation decisions, not a recommended sample size or a completed audit of a real business.

Start with the question and the population

The fictional question is whether selected manual stock adjustments show the required approval before posting. For learning, assume an export containing 80 adjustments for one defined review period. Before using a real export, reconcile record counts, identifiers and relevant totals to independent source controls; inspect exclusions, duplicates and date boundaries.

That approval question is different from asking whether adjustments were complete, accurately valued or supported by a physical count. A well-documented sample for one question does not answer all the others.

Explain the selection and its limits

The example deliberately selects eight manual weekend or override-related items for exploratory review. This is a targeted selection, not a representative statistical sample. Eight is a storytelling choice, not an assurance benchmark. Its rationale is to examine specified risk features; it does not justify a confidence level or a population-wide error estimate.

If the engagement objective requires a conclusion about the population, design an appropriate sampling approach and sample size under the applicable methodology. Record the population, sampling unit, risk considerations, expected and tolerable deviation or misstatement where relevant, and the basis for evaluation. Do not retrofit a representativeness claim after seeing the results.

Record what actually happened to each selected item

The fictional log contains eight selected IDs. Six procedures are completed: four show no exception for the stated attribute and two show no approval evidence before posting. Two procedures remain pending because the relevant evidence has not yet been obtained.

The two missing-approval results are exceptions to the stated test in the fictional records. They do not automatically establish financial misstatement, fraud or the absence of all authorization. Investigate alternative evidence, timing, cause and significance. Pending procedures are not passes, and replacing a difficult item without recording why can distort the work.

The arithmetic is intentionally transparent:

• Eight items selected equals six completed plus two pending.
• Six completed equals four with no observed exception plus two exceptions.
• Two out of six completed tests equals 33.3% within those completed targeted procedures.

That descriptive fraction is not an estimate of the deviation rate among all 80 records. The selection is targeted, two tests are incomplete and no inferential design has been established. Do not multiply it by 80 or describe the remaining population as error-free.

Investigate before concluding

For each exception, record the criterion, observed condition, evidence reference, possible cause, significance and the next procedure. For pending items, document the missing evidence and how the limitation affects the conclusion. Follow the applicable engagement methodology for alternative procedures, additional work and escalation.

A defensible interim conclusion might read: “In six completed targeted tests, two records lacked the expected pre-posting approval evidence. Two selected items remain untested. The work identifies issues for follow-up in the selected records and does not support a population-wide deviation estimate.”

The conclusion should also consider the quality of the population and the design of the procedure. Testing every recorded entry would still not, by itself, establish that unrecorded entries do not exist.

Use the fillable page

Page 7 provides fields for objective and criterion, population and completeness, selection method and rationale, procedure and evidence, exceptions and limitations, and conclusion and reviewer. It is an interactive PDF form for a concise sampling summary; detailed item evidence should remain in a linked audit working-paper system.

Download the PDF and use a PDF reader that supports forms. Some browser previews display form fields without supporting saved edits. Verify the saved values after editing and retain supporting records outside the form. The page does not calculate a sample size, statistical confidence or projected misstatement.

Professional context

PCAOB AS 2315 provides a public example of how an auditing standard connects objectives, population characteristics, selection and evaluation. It is a U.S. external-audit standard; it is not presented here as a legal requirement for every Indonesian business or internal audit engagement. Apply the standards and methodology relevant to your own assignment.

The case, diagrams and fillable page are original Inside Assurance material. Their purpose is to make the reasoning reviewable: what did you select, why, what evidence did you obtain, and what remains unanswered?

Inside Assurance — Beyond the numbers. Inside the business.