When a sample finds errors
What an audit team does when sampled items are misstated: errors versus anomalies, projecting to the population, extending the test, and recording it.
By Achekam · · 5 min read · For Accountants & audit firms in South Africa, Nigeria, Kenya
The sample of purchase invoices comes back with some items charged to the wrong period and another with VAT claimed on a supplier who is not registered. The senior adjusts those items, writes "errors corrected" on the paper and moves on. The reviewer sends it back. The errors in the sample were corrected, but the sample was only ever a stand-in for the population, and nobody has asked what those errors say about the items that were not tested.
ISA 530 requires the auditor to investigate the nature and cause of every deviation or misstatement found in a sample, to project misstatements found to the population, and to evaluate whether the sample has provided a reasonable basis for conclusions about the population. Correcting the sampled items is necessary; it is never enough on its own.
Errors versus anomalies
The first step is understanding each misstatement: what happened, why, and whether it could have happened elsewhere in the population.
ISA 530 allows a misstatement to be treated as an anomaly, and therefore not projected, only in extremely rare circumstances: when the auditor obtains a high degree of certainty that it is not representative of the population. That certainty comes from additional procedures, not from the explanation alone. A one-off system failure on a single day, confirmed by the system's own logs and by testing other transactions from that day, might qualify. "The clerk was new" does not, because other transactions processed by the clerk may have the same problem.
Look at the cause before deciding anything else. Misstatements that share a cause, such as a supplier whose invoices are always posted on payment rather than receipt, suggest a systematic problem that projection may understate. A misstatement that looks deliberate is a fraud question, and the response moves beyond sampling.
Projecting the misstatement
For tests of details, the misstatements found in the sample are projected to the population, so the auditor can compare the likely total with tolerable misstatement. The method should match how the sample was drawn:
- Ratio approach: the misstatement in the sample as a proportion of the sample's value, applied to the population's value. Suits samples where misstatements are roughly proportional to item size.
- Difference approach: the average misstatement per sampled item, applied to the number of items in the population. Suits misstatements that are similar in amount regardless of item size.
- Monetary unit sampling: each misstatement is expressed as a proportion of the recorded amount of the item, its tainting, and projected to the sampling interval it represents.
Where the population was stratified, project within each stratum and add the results. Items tested as specific items, such as every invoice above a threshold, are not projected; their misstatements count in full, but only for themselves.
A worked example in words: the team samples the purchases population and finds the invoices with cut-off errors. Because the errors share a cause, an invoice-on-payment practice at one supplier, the team projects them and also tests every invoice from that supplier near year end. The projected misstatement for the rest of the population is added to the actual misstatement found in the supplier's invoices, and the total is compared with tolerable misstatement.
Extending the test or changing the approach
If the projected misstatement plus any anomalous misstatement exceeds tolerable misstatement, or comes close to it, the sample has not provided a reasonable basis for concluding on the population. ISA 530 then expects the auditor to:
- ask management to investigate the misstatements identified and the potential for further misstatements, and to make any necessary adjustments;
- tailor the nature, timing and extent of further procedures, for example by extending the sample or testing an alternative control;
- consider whether the risk assessment needs revising.
Extending a sample is only useful if the extension is drawn by the same method from the same population. In Achekam's Audit Intelligence, a reviewer who did not prepare the plan approves the sample before it is drawn, and every draw can be reproduced from its stored seed. An extension therefore goes through the same plan-and-approve step, and the original draw remains reproducible beside it. How to choose the method in the first place is covered in random, systematic or monetary unit sampling.
Recording findings on the paper
In Achekam, findings can be raised from a working paper and closed or withdrawn, with Findings and Exceptions tabs on each paper. Staff can record an audit adjustment on a lead schedule line with a reason, and each figure on the lead schedule is traceable to its trial balance lines.
On the work item, the sampling step is marked done only with a written response and its documents, or set aside with a reason; steps are never deleted. The response should show the population and its value, the method and sample size, each misstatement with its cause, whether any was treated as an anomaly and the evidence for that, the projection method and result, and the conclusion against tolerable misstatement.
Where projections go while the register stays empty
Projected misstatements are part of the total that must be evaluated at completion under ISA 450, together with factual and judgemental misstatements. In Achekam, that total cannot yet be assembled automatically: the completion Summary of Uncorrected Misstatements stays empty, because no screen or route writes to the misstatement register it reads. Server code to record or project a misstatement exists, but no route calls it.
Until that changes, keep a misstatement schedule on a working paper of its own, and carry each sample's projected misstatement into it with a reference back to the sampling paper. Projected amounts are not usually booked as adjustments, because they are estimates; they still count in the evaluation. How that schedule should be built is covered in the uncorrected misstatements schedule. The ACCA's article on evaluation of misstatements explains ISA 450 on evaluating misstatements identified during an audit, including the effect of uncorrected misstatements, with worked examples.
A review point worth raising
When reviewing any sampling paper with exceptions, look first for the word "projected". If it is not there, send the paper back before reading anything else.
General guidance only, not advice on your own circumstances. Tax and payroll rules change: check the current position with the authority or your adviser.
