ACCA AA: When Audit Sampling Isn't the Right Answer

Ask an ACCA Audit and Assurance (AA) candidate how an auditor should test a population of transactions or balances, and the reflex answer is almost always "take a sample." It is an understandable habit — audit sampling gets the most classroom airtime, has its own dedicated standard, and feels like the technically sophisticated answer. But ISA 500 Audit Evidence is explicit that audit sampling is only one of several means available to an auditor for selecting items to test, and in a good number of exam scenarios it is not the appropriate one. Candidates who default to sampling regardless of what the scenario actually describes routinely lose marks in AA questions that were testing something much simpler: whether you can match the testing approach to the population in front of you.

This matters because the examining team designs these scenarios deliberately. A question describing a small number of material non-current asset additions, or a handful of individually significant receivable balances, is not an accident — it is testing whether you recognise that sampling would be the wrong tool entirely. Getting the underlying logic straight is the fastest way to stop losing these marks, and it pairs well with the mechanics covered in our separate piece on how auditors actually select test samples, which focuses on the sampling method itself once sampling genuinely is the right choice.

Three ways to select items for testing, not one

ISA 500 sets out that when designing tests of controls or substantive procedures, the auditor has three broad means of selecting items for testing:

• Selecting all items (100% examination) — testing every item in the population.
• Selecting specific items — targeting particular items based on characteristics such as high value, unusual nature, or known risk indicators, without covering the rest of the population.
• Audit sampling — selecting a subset of items in a way that gives every sampling unit a chance of selection, governed in detail by ISA 530 Audit Sampling, so that the results can be projected onto the population as a whole.

Only the third of these is audit sampling in the technical sense. The first two are legitimate, standard-compliant testing approaches in their own right — they are not compromises or shortcuts taken when sampling “isn't practical.” An AA scenario that describes a population where one of the first two approaches clearly fits is testing exactly this distinction, and writing an answer built entirely around sample sizes, sampling risk, and selection methods when the scenario calls for 100% testing will not earn the marks available.

When 100% testing is the right call

Testing every item in a population is usually the most appropriate approach when the population is small enough that testing it in full is not disproportionately time-consuming, or when the items in it are individually significant enough that missing a misstatement in even one of them would matter. Typical AA scenario triggers include a small number of material non-current asset additions or disposals in the year, a short list of individually material receivable or payable balances, or a population where the risk of material misstatement is assessed as high and the auditor needs certainty rather than a projection. In these cases, sampling would actually introduce unnecessary sampling risk — the risk that a sample happens not to be representative — where none needs to exist, because testing everything is entirely feasible.

When selecting specific items is the right call

Selecting specific items sits between the other two approaches. It is used to target the high-value, unusual, or higher-risk items within a population directly, rather than testing everything or drawing a representative sample of the whole. A common AA scenario is a large population of trade receivables where a handful of balances are individually material and the remainder is not — the auditor might select every balance above a materiality-linked threshold for full testing, then apply a separate sampling or analytical approach to the smaller, more homogeneous remainder. The key exam-technique point is that testing specific items only gives the auditor evidence about the items actually selected — it says nothing about the rest of the population, and the results cannot be extrapolated the way a properly designed sample's results can. Candidates who select specific items but then write about “projecting the misstatement to the whole population” are mixing up two different approaches and will lose marks for the inconsistency.

When audit sampling is genuinely the right call

Audit sampling earns its place when the population is large, reasonably homogeneous, and made up of items that are not individually significant — a large volume of low-value expense transactions, a big population of payroll entries, or thousands of similar inventory lines, for example. Here, testing every item would be inefficient and disproportionate, and no small number of individual items carries enough risk on its own to justify targeted selection. Sampling lets the auditor draw a conclusion about the whole population from testing a representative subset, provided the sample is designed properly under ISA 530 — every item in the population needs a chance of selection, and the auditor needs to consider sampling risk, tolerable and expected misstatement, and how any misstatements found in the sample will be evaluated and projected.

A practical decision framework for the exam

When an AA scenario describes a population to be tested, work through three questions in order before deciding how to write up the testing approach:

• How big is the population, and are the items individually significant? A small population, or one made up of individually material items, points towards 100% testing.
• Are there a few high-risk or high-value items sitting inside a larger, less significant population? That combination points towards selecting those specific items for full testing, potentially alongside a separate approach for the remainder.
• Is what's left large, homogeneous, and made up of low-risk, non-material individual items? That is the profile that genuinely calls for audit sampling.

Working through the population characteristics in this order — size, then individual significance, then homogeneity and residual risk — stops the reflex jump to sampling and forces an answer that actually matches the scenario given. It is also worth remembering that these approaches are not mutually exclusive within a single audit area: it is entirely normal, and often the technically correct answer, to select specific high-value items for 100% testing and then sample the remaining, less significant balance of the same population.

This same discipline — reading the scenario's specific risk and population characteristics rather than reaching for a memorised standard procedure — comes up again when assessing going concern indicators, which we cover in our article on identifying material uncertainty related to going concern. In both cases, the examining team is rewarding candidates who apply judgement to the facts given rather than reproducing a standard technique on autopilot.

Building this habit also pays off beyond the exam: recognising which testing approach fits which population is a core part of real audit planning, referenced throughout the ACCA Audit and Assurance syllabus and directly examinable at the planning and risk-response stage of a question.

FAQ

Is audit sampling ever wrong to mention at all in an AA answer?

No — sampling is frequently the correct answer, particularly for large, homogeneous populations of low-value items. The mistake is not mentioning sampling; it is defaulting to it without first checking whether the population described actually fits the profile for 100% testing or specific-item selection instead.

Can an auditor combine more than one of these approaches for the same balance?

Yes, and doing so is often the strongest answer. A common approach is to select all individually material items within a population for 100% testing, then apply audit sampling to the remaining, less significant items, giving the auditor evidence covering the whole balance without disproportionate effort.

Does selecting specific items reduce audit risk the same way sampling does?

Not in the same way. Selecting specific items gives assurance only over the items actually tested — it cannot be used to draw a statistical conclusion about the untested remainder of the population. If the untested portion is significant, the auditor still needs a separate procedure, such as sampling or analytical review, to address it.