Audit sampling glossary

Audit sampling means applying an audit procedure to less than 100% of the items in a population, giving every item a chance of selection, so you can conclude about the whole population. This glossary defines, in one to three sentences each, the terms you meet when planning, selecting and evaluating a sample under ISA 530 (AU-C 530 in the US), with a worked figure where it helps.

Updated 10 November 2026

Definitions marked "(ISA 530)" follow the standard's definitions, explained in our own words. Worked figures use the case from the monetary unit sampling guide: a population of 4,812,350, tolerable misstatement 120,000, expected misstatement 15,000 and 95% confidence. Each term links to the guide that covers it in depth.

A

Anomaly
A misstatement or deviation that the auditor can show is not representative of the rest of the population. It takes a high degree of certainty and evidence on file; it is left out of the projection, but its effect, if uncorrected, is still considered. (ISA 530) See the guide: Evaluating MUS results
Attribute sampling
Sampling for tests of controls that estimates the rate of deviation from a control in the population. Its result is an upper deviation limit compared with the tolerable rate. See the guide: Attribute sampling
Audit sampling
Applying an audit procedure to less than 100% of the items in a population, in a way that gives every sampling unit a chance of selection, so the auditor has a reasonable basis for conclusions about the whole population. (ISA 530) See the guide: ISA 530

B

Basic precision
In MUS, the part of the upper misstatement limit that covers sampling risk even when no misstatement is found: the zero-error confidence factor times the sampling interval. At 95% with an interval of 32,000, 3.00 × 32,000 = 96,000. See the guide: Evaluating MUS results
Benford's law
The expected distribution of first digits in many sets of real amounts, P(d) = log10(1 + 1/d): 30.1% start with 1 and 4.6% with 9. In an audit it flags groups of amounts worth examining; it does not prove fraud. See the guide: Benford's law
Block selection
Choosing one or more blocks of contiguous items, such as every March invoice. It rarely works for sampling, because items in a sequence tend to resemble each other and the block does not represent the rest of the year. See the guide: Sample selection methods
Book value
The recorded amount of an item or of the whole population. In MUS, population book value divided by the sampling interval gives the sample size. See the guide: Monetary unit sampling

C

Confidence factor
The Poisson factor in the MUS formula, set by the risk of incorrect acceptance and the number of misstatements: 2.31, 3.00 and 4.61 at 90%, 95% and 99% with zero misstatements, from the AICPA Audit Sampling Guide. See the guide: Monetary unit sampling
Confidence level
The complement of the sampling risk you accept: 95% confidence is a 5% risk of the wrong conclusion (incorrect acceptance in tests of details, overreliance in tests of controls). See the guide: Sampling risk

D

Deviation
A case in a test of controls where the control was not applied as designed, such as an invoice paid without approval. Deviations are counted, not valued. See the guide: Attribute sampling

E

Expansion factor
In MUS planning, the factor that inflates expected misstatement to allow for sampling error on it: 1.5, 1.6 and 1.9 at 90%, 95% and 99% (AICPA). See the guide: Monetary unit sampling
Expected misstatement
The misstatement the auditor expects to find in the population, based on prior years or knowledge of the client. The closer it gets to tolerable misstatement, the larger the sample; once expected misstatement × expansion factor reaches it, no sample size works. See the guide: Audit sample size
Expected rate of deviation
The deviation rate the auditor expects in a test of controls. With 5% tolerable, 1% expected and 95% confidence, the sample is 93 items with one deviation allowed (AICPA Table A-1). See the guide: Attribute sampling

H

Haphazard selection
Picking items without a structured technique while avoiding conscious bias and predictability. Fine for non-statistical sampling, not for statistical sampling, because it gives no unit a known probability of selection. See the guide: Sample selection methods

I

Incremental allowance
In MUS evaluation, the amount added to the upper limit for each misstatement found: (increase in confidence factor − 1) × projected misstatement, ranking taintings from largest to smallest. At 95%, the first misstatement adds (4.75 − 3.00 − 1) = 0.75 times its projection. See the guide: Evaluating MUS results
Individually significant item
An item examined 100% because of its size or risk, outside the sample. In MUS, any item equal to or larger than the sampling interval is selected with certainty. See the guide: Stratified sampling

J

Journal entry testing
Testing required by ISA 240 on every audit in response to the risk of management override: get the complete journal, filter it with risk criteria and examine the entries selected. It is targeted selection, not sampling. See the guide: Journal entry testing

M

MAD (mean absolute deviation)
The average absolute difference between actual and expected digit proportions in a Benford test. It does not depend on the number of records; for the first digit, below 0.006 is close conformity and above 0.015 nonconformity (Nigrini). See the guide: Benford's law
Management override of controls
Management's ability to manipulate the accounts by bypassing controls that work for everyone else. ISA 240 treats it as a risk present in every entity. See the guide: Journal entry testing
Misstatement
A difference between the amount, classification, presentation or disclosure of an item and what the financial reporting framework requires. It can arise from error or fraud. See the guide: Evaluating MUS results
Monetary unit sampling (MUS)
Statistical sampling in which the sampling unit is each monetary unit of the population, so larger items are more likely to be selected. Used in tests of details of balances at risk of overstatement. See the guide: Monetary unit sampling

N

Non-sampling risk
The risk that the auditor reaches a wrong conclusion for any reason other than sampling, such as using an unsuitable procedure or misreading evidence. Planning, supervision and review reduce it; a larger sample does not. (ISA 530) See the guide: Sampling risk
Non-statistical sampling
Sampling that lacks one of the two features of statistical sampling: random selection and evaluation using probability theory. ISA 530 allows it, but sampling risk is judged, not measured. See the guide: Statistical vs non-statistical sampling

P

Performance materiality
An amount below materiality for the financial statements as a whole, set to reduce to a low level the chance that uncorrected and undetected misstatements exceed it. Tolerable misstatement for a sample is usually set at or below it. See the guide: Audit sample size
Population
The entire set of data from which a sample is drawn and about which the auditor wants to conclude. It must be complete, reconciled to the ledger, and suitable for the assertion being tested. (ISA 530) See the guide: Audit sample size
Projected misstatement
The misstatement in the sample extrapolated to the population. In MUS each item below the interval contributes its tainting × the interval; items above it contribute their actual difference. See the guide: Evaluating MUS results

R

Random selection
Selection using computer-generated random numbers or tables, giving each unit the same probability of selection. It is the basis of statistical sampling. See the guide: Sample selection methods
Risk of incorrect acceptance
In tests of details, the risk of concluding that a balance is not materially misstated when it is. It affects audit effectiveness and is the risk the confidence level controls. See the guide: Sampling risk
Risk of incorrect rejection
In tests of details, the risk of concluding that a balance is materially misstated when it is not. It affects efficiency: it usually leads to extra work that ends up correcting the conclusion. See the guide: Sampling risk
Risk of overreliance
In tests of controls, the risk of concluding that a control is more effective than it really is. It is the counterpart of the risk of incorrect acceptance and affects audit effectiveness. See the guide: Attribute sampling

S

Sample size
The number of items to examine to reduce sampling risk to an acceptably low level. With a population of 4,812,350, tolerable misstatement 120,000, expected misstatement 15,000 and 95% confidence, MUS gives 151 items. See the guide: Audit sample size
Sampling interval
In systematic selection and in MUS, the distance between one selection and the next on the cumulative listing. In the MUS example, (120,000 − 15,000 × 1.6) ÷ 3.00 = 32,000, and every item of 32,000 or more is selected. Many tools, AuditSampler Pro included, use book value divided by the rounded-up sample size instead (4,812,350 ÷ 151 = 31,869.87): a slightly smaller, and so slightly more conservative, interval. See the guide: Monetary unit sampling
Sampling risk
The risk that the conclusion drawn from a sample differs from the one you would reach by examining the whole population. It can lead to accepting a misstated balance or an ineffective control, or to rejecting a sound one. (ISA 530) See the guide: Sampling risk
Sampling unit
Each of the individual items that make up a population: invoices, lines on a listing or, in MUS, each monetary unit. The right one depends on the test: for receivables existence, each customer balance or each invoice. (ISA 530) See the guide: Monetary unit sampling
Selecting specific items
Choosing items for their size or risk, such as everything above 50,000 or all closing entries. It is not sampling: the result cannot be projected to the rest of the population. See the guide: Sample selection methods
Statistical sampling
Sampling with random selection of the sample and evaluation of results using probability theory, including measuring sampling risk. Without both features, sampling is non-statistical. (ISA 530) See the guide: Statistical vs non-statistical sampling
Stratification
Dividing a population into sub-populations (strata), each a group of units with similar characteristics, often similar monetary value. It allows smaller samples and puts effort on the large items. (ISA 530) See the guide: Stratified sampling
Systematic selection
Divide the population by the sample size to get an interval, pick a random start within the first interval, and take one item in each interval. Watch out for patterns in the population that line up with the interval. See the guide: Sample selection methods

T

Tainting
In MUS, the misstatement in an item divided by its book value. An invoice of 10,000 audited at 8,000 has a 20% tainting; with an interval of 32,000 it projects 6,400. See the guide: Evaluating MUS results
Test of controls
A procedure that evaluates whether a control operated effectively in preventing or detecting material misstatements. When it is sampled, attribute sampling is used. See the guide: Attribute sampling
Test of details
A substantive procedure on individual items in a balance or class of transactions, such as vouching customer invoices to support. When it is sampled, MUS or variables sampling is used. See the guide: Monetary unit sampling
Tolerable misstatement
An amount set by the auditor as the most misstatement acceptable in the population, seeking appropriate assurance that actual misstatement does not exceed it. It is usually at or below performance materiality. (ISA 530) See the guide: Audit sample size
Tolerable rate of deviation
A rate of deviation from a control set by the auditor as the most acceptable, seeking appropriate assurance that the actual rate does not exceed it. The more you rely on the control, the lower it should be. (ISA 530) See the guide: Attribute sampling
Top stratum
The stratum of items above a cut-off amount, examined 100%. The rest of the population is sampled, and only the rest is projected. See the guide: Stratified sampling

U

Upper deviation limit
The highest deviation rate in the population at the chosen confidence, worked out from the deviations in the sample. With 93 items and one deviation at 95% it is 5.0%: if it does not exceed the tolerable rate, the control can be relied on. See the guide: Attribute sampling
Upper misstatement limit
In MUS, projected misstatement + basic precision + incremental allowance. With one 20% tainting and an interval of 32,000: 6,400 + 96,000 + 4,800 = 107,200, below tolerable misstatement of 120,000, so the balance is accepted. See the guide: Evaluating MUS results

V

Variables sampling
Sampling for tests of details that estimates an amount (audited value or total misstatement) instead of a rate. It includes mean-per-unit, difference and ratio estimation. See the guide: Attribute vs variable sampling

Frequently asked questions

What is audit sampling in simple words?

Testing part of a population instead of all of it, choosing the items so that every one could have been picked, and using the results to conclude about the whole population. ISA 530 sets the requirements.

What is the sampling unit in audit sampling?

Each of the individual items in the population that can be selected: an invoice, a line on a listing, a customer balance or, in monetary unit sampling, each monetary unit.

What is the audit sampling interval?

The distance between one selection and the next in systematic selection and in monetary unit sampling: population value divided by sample size. Any item equal to or larger than the interval is always selected.

What is the difference between sampling risk and non-sampling risk?

Sampling risk is the chance that a sample leads to a different conclusion from testing the whole population; larger samples reduce it. Non-sampling risk comes from anything else, such as an unsuitable procedure or misreading the evidence.

What is tolerable misstatement?

The most misstatement the auditor will accept in a population, seeking appropriate assurance that the actual misstatement does not exceed it. It is usually set at or below performance materiality.

  • IAASB, ISA 530 Audit Sampling
  • AICPA, AU-C 530 Audit Sampling, and Audit Guide: Audit Sampling
  • ICAC, NIA-ES 530 Muestreo de auditoría (Spain)
  • IAASB, ISA 240 The Auditor's Responsibilities Relating to Fraud in an Audit of Financial Statements
  • M. J. Nigrini, Benford's Law: Applications for Forensic Accounting, Auditing, and Fraud Detection, Wiley, 2012

From the calculation to the working paper

AuditSampler Pro plans, selects and evaluates monetary unit, attribute and stratified samples under ISA 530, and produces the working paper as a PDF. Client data stays on your computer.