Audit sampling documentation: what the working paper must record

An audit sampling working paper must let another experienced auditor understand and re-perform the sample: population, parameters, size, selection method with its starting point, items tested, misstatements and conclusion. ISA 530 has no documentation paragraph of its own; the requirements come from ISA 230, applied to each step of the sample.

Updated 10 November 2026

What ISA 230 requires

  • Paragraph 8. Documentation must be enough for an experienced auditor, with no previous connection with the audit, to understand the nature, timing and extent of the procedures, their results and the conclusions.
  • Paragraph 9(a). Record the identifying characteristics of the specific items tested.
  • Paragraphs 9(b) and 9(c). Who performed the work and when, and who reviewed it, when and to what extent.
  • Paragraph A12. For a systematic sample, the auditor may identify the items by recording "the source, the starting point and the sampling interval". The standard's example selects every 125th shipping report from a given starting number.

A12 is the paragraph that matters most for a sample: without the random start (or the generator's seed) and the interval, nobody can check that the sample wasn't picked for convenience. The PCAOB standard in the US asks for the same (AS 1215, paragraph 10 and its note).

What to document for each ISA 530 step

ParagraphStepWhat should be on file
6Sample designObjective of the test, population (what it includes, total agreed to the ledger), sampling unit and what counts as a misstatement or deviation
7Sample sizeMethod (monetary unit, attribute, stratified or non-statistical), confidence, tolerable and expected misstatement or deviation rate, factors and resulting size
8SelectionSelection method, random start or seed, interval and the list of items selected
9-11ProceduresWhat was done on each item, replacement items and why, and items that could not be tested (treated as a misstatement or deviation)
12-13Deviations and misstatementsNature and cause of each; if any was treated as an anomaly, why it is not representative, with a high degree of certainty
14ProjectionMisstatement projected to the population (tests of details)
15Evaluation and conclusionUpper limit against tolerable misstatement or deviation rate, and the conclusion on whether sampling gave a reasonable basis

ISA 530 is not in the ISA 230 appendix that lists documentation requirements in other ISAs either: everything above is documented because paragraph 8 needs it.

Example: a monetary unit sampling working paper

With the figures from the monetary unit sampling Excel template, the working paper would record at least:

SectionContent
ObjectiveOverstatement of trade receivables at 31 December
Population12 invoices, 144,600, agreed to the ledger; no negative or zero items
Parameters95 % confidence, tolerable misstatement 70,000, expected misstatement 5,000; factors 3.00 and 1.6 (AICPA)
Size and interval7 items; interval 20,657.14
SelectionSystematic monetary unit selection, random start 7,413; 7 points, 6 invoices (INV-005 with two points; INV-005 and INV-011 in the top stratum)
ProcedureAgreed to subsequent receipt or confirmation; reference to each invoice's working paper
MisstatementsINV-002, 378 (2 % tainting, pricing error); INV-011, 300 (top stratum); INV-007, understatement of 350. Cause of each
EvaluationBasic precision 61,971.43, projected overstatement 713.14, incremental allowance 309.86, upper limit 62,994.43
ConclusionThe upper limit does not exceed tolerable misstatement: at 95 % confidence the balance is not materially overstated
Sign-offPrepared by and date; reviewed by, date and extent of review

The causes of the misstatements are illustrative: on a real engagement they are investigated (paragraph 12) and can change the conclusion even when the limit is below tolerable.

Common working paper gaps

  • No starting point. The interval is recorded, but not the start or the seed, so the selection can't be re-performed.
  • Population not agreed. Nothing shows that the sampled total matches the balance, or which items (negative, zero, different nature) were taken out and how they were tested.
  • Parameters not supported. Confidence or tolerable misstatement aren't linked to the engagement's risk assessment and materiality.
  • Deviations without a cause. They are counted, but nobody investigates why they happened.
  • A conclusion that doesn't follow from the evaluation. The paper concludes without comparing the upper limit with tolerable misstatement, or without projecting the misstatements found.

Frequently asked questions

Which standard governs audit sampling documentation?

ISA 230, which applies to the whole audit. ISA 530 has no documentation paragraph of its own: what is documented for a sample comes from ISA 230 paragraphs 8 and 9 applied to each step of the sample.

What should be recorded for a systematic selection?

The source of the documents, the starting point and the sampling interval, as ISA 230 paragraph A12 says. With those three, another auditor can re-perform the selection.

Do I need to keep the list of items selected?

You must be able to identify the items tested (ISA 230 paragraph 9). For a systematic sample, the source, start and interval are enough; in practice the list is attached because it makes review easier.

What should the conclusion of a sample say?

Whether the upper misstatement limit or upper deviation rate exceeds the tolerable amount or rate, and whether sampling gave a reasonable basis for conclusions about the population (ISA 530 paragraph 15), taking into account the cause of the misstatements found.

Who should sign the sampling working paper?

The preparer and the reviewer, each with the date, and the extent of the review, under ISA 230 paragraph 9.

The working paper, produced for you

AuditSampler Pro puts the parameters, factors, random start, selected items and evaluation in one PDF, for monetary unit and attribute samples.