Audit Materiality and Sampling: Why Auditors Test Selectively

Audit materiality is the threshold at which a misstatement, individually or in combination with others, could reasonably be expected to influence the decisions of users who rely on the financial statements. If an error or omission is below that threshold, it is considered immaterial, meaning it would not change a reasonable person’s assessment of the financial statements. Auditors establish this threshold at the start of every engagement so they can focus their work on what genuinely matters and design efficient, risk-focused procedures.

Because most organizations have thousands or millions of transactions, auditors cannot and do not test every one. Instead, they rely on audit sampling, applying procedures to a representative subset of a population and projecting those results to the whole. Together, materiality and sampling are what make a financial statement audit both practical and rigorous.

What Is Audit Materiality and How Is It Determined?

Materiality is not a fixed dollar amount imposed by a regulator. It is a matter of professional judgment, informed by quantitative benchmarks and qualitative context, applied to the specific facts of each engagement.

The Governing Standards

For audits of private companies and nonprofits, the AICPA’s AU-C Section 320, Materiality in Planning and Performing an Audit, sets the framework. Auditors of public companies follow PCAOB AS 2105, Consideration of Materiality in Planning and Performing an Audit. Both standards share the same conceptual foundation: a misstatement is material if there is a substantial likelihood it would influence the judgment of a reasonable investor or financial statement user.

Neither standard prescribes a fixed percentage. Instead, both require auditors to choose an appropriate benchmark, apply professional judgment, and document their rationale.

Choosing a Benchmark

Auditors typically anchor materiality to a financial statement figure that reflects the primary interest of users. Common benchmarks include:

• Pre-tax income from continuing operations: Often used for for-profit entities; a commonly cited starting point is 5% of pre-tax income.
• Total revenues: Often used when income is volatile or negative; typical starting ranges are 0.5% to 1% of revenues.
• Total assets: Useful for asset-heavy businesses such as financial institutions; a common range is 0.5% to 1%.
• Net assets or total equity: Sometimes appropriate for nonprofits or capital-intensive entities.

These percentages are starting points, not bright lines. The SEC’s Staff Accounting Bulletin No. 99 (SAB 99) explicitly warns that mechanically applying any single percentage “has no basis in the accounting literature or the law.” A misstatement that appears small by the numbers can still be material if it, for example, turns a reported profit into a loss, conceals a regulatory violation, or masks a trend that investors would find significant.

Qualitative Factors That Override the Math

Even a misstatement well below the quantitative threshold can be material when qualitative circumstances amplify its significance. Auditors weigh factors such as:

• Whether the misstatement masks a change in earnings direction
• Whether it relates to a segment that has been publicly highlighted
• Whether it involves related-party transactions or fraud indicators
• Whether it affects compliance with debt covenants or regulatory requirements
• Whether it affects management compensation tied to reported figures

Combining a quantitative threshold with a qualitative overlay is what prevents “managing to the number” and keeps materiality judgments useful to financial statement users.

Performance Materiality: The Working Buffer

Once an auditor sets overall materiality for the financial statements as a whole, they establish a second, lower figure: performance materiality. Performance materiality is defined in AU-C 320 as the amount, or amounts, set below overall materiality to reduce to an appropriately low level the probability that the aggregate of uncorrected and undetected misstatements exceeds overall materiality. When performance materiality is applied to a specific sampling procedure, the resulting threshold is called tolerable misstatement, which under AU-C 530 may equal performance materiality or be set lower when the sampled population is smaller than the account balance.

In practice, most firms set performance materiality at 50% to 75% of overall materiality. The lower end of that range applies when:

• Prior-year audits revealed significant audit adjustments
• The control environment is weak or untested
• It is a first-year engagement with limited historical data
• The auditor expects a higher volume of small errors that could accumulate

A well-controlled, repeat-engagement client with a clean history might justify performance materiality closer to 75% of overall materiality. The buffer exists because auditors are testing samples, not populations, so they need a cushion to ensure that errors they do not detect, combined with errors they find but management does not correct, will still not cross the overall materiality line.

Performance materiality drives the design of every substantive test: it determines how large a sample needs to be, how much misstatement the auditor can tolerate in any given account, and how aggressively errors found in testing need to be projected.

How Audit Sampling Works

With materiality set, auditors move to designing their procedures. Because testing every transaction in a large population is impractical, they select a sample and draw conclusions about the population as a whole. This is governed by AICPA AU-C Section 530, Audit Sampling, and for public company audits, by PCAOB AS 2315, Audit Sampling.

Statistical vs. Non-Statistical Sampling

Statistical sampling uses probability theory at both the selection stage and the evaluation stage. Every item in the population has a known, non-zero chance of being selected, and the auditor can quantify sampling risk mathematically. Common statistical techniques include:

• Monetary unit sampling (MUS): Probability of selection is proportional to dollar amount. Larger balances are more likely to be selected, which concentrates testing where misstatement would matter most.
• Classical variables sampling: Uses traditional statistical methods to estimate the total population value or error.
• Attribute sampling: Used for tests of controls; estimates the rate at which a control deviation occurs in the population.

Non-statistical (judgmental) sampling relies on the auditor’s judgment for both selection and evaluation. It does not produce a mathematically measurable sampling risk, but it remains fully acceptable under GAAS and PCAOB standards when applied with appropriate rigor. Many auditors use non-statistical sampling for smaller populations or when specific items warrant focused scrutiny rather than random selection.

Neither approach is inherently superior. The choice depends on cost, population characteristics, and what provides sufficient, appropriate audit evidence in the circumstances.

What Drives Sample Size

Sample size is not arbitrary. The auditor considers several inputs simultaneously:

1. Tolerable misstatement: The lower the threshold, the larger the sample needed to detect errors at that level.
2. Expected misstatement in the population: If prior experience suggests the population contains errors, a larger sample is required to bound the risk.
3. Risk of incorrect acceptance: How likely is the auditor willing to accept that the account is fairly stated when it actually is not? Lower acceptable risk requires larger samples.
4. Population size: For large populations, sample size grows much more slowly than population size, which is why statistical sampling can be efficient even for millions of transactions.
5. Stratification: Dividing a population into homogeneous subgroups (by dollar size, transaction type, or timing) and sampling within each stratum can reduce required sample sizes while improving the precision of conclusions.

High-Value and 100% Testing

Not everything is sampled. Auditors routinely identify items that are individually significant, meaning their dollar value alone exceeds or approaches overall materiality, and test 100% of those items. The remaining population is then sampled separately. This is not a departure from sampling theory; it is a straightforward risk response, directing effort exactly where a single error could matter.

Projecting Sample Results

When a sample reveals misstatements, auditors do not simply total what they found. They project the error to the population. If a sample representing 10% of a population by dollar value contains $50,000 in misstatements, the auditor projects $500,000 of likely misstatement across the full population and compares that projection to tolerable misstatement. Under AU-C 530, all misstatements found in a sample are generally projected to the population, leaving no room to dismiss findings as isolated.

If projected misstatement approaches or exceeds tolerable misstatement, the auditor expands testing, requests corrections, or evaluates the impact on the audit opinion.

Materiality, Sampling, and the Audit Opinion

Materiality and sampling connect directly to the audit report. The standard unqualified (clean) opinion states that the financial statements are presented fairly, in all material respects, in conformity with the applicable financial reporting framework. That phrase “in all material respects” is not hedging language. It is the auditor’s acknowledgment that the audit was designed to detect material misstatements, not every conceivable error, and that the procedures performed provide reasonable, not absolute, assurance.

When uncorrected misstatements accumulate to near the materiality threshold, the auditor evaluates whether the financial statements, taken as a whole, are still fairly presented. If they are not, the auditor must either require correction or modify the opinion. This is why management and auditors often review and resolve audit adjustments before the report is issued: those adjustments are not administrative formalities but the mechanism by which material misstatements are eliminated before the financial statements go to users.

For clients subject to heightened oversight, such as nonprofit organizations with federal funding above $1,000,000 in expenditures (the Single Audit threshold for fiscal years beginning on or after October 1, 2024) or employee benefit plans requiring ERISA audits, specific compliance thresholds layer on top of the financial statement materiality analysis, requiring auditors to maintain separate, often lower, materiality levels for program-specific testing.

For organizations looking to understand what a financial statement audit covers, the AICPA’s audit standards resources provide the definitive framework.

How AI-Native Audit Practices Approach Materiality and Sampling

The conceptual framework for materiality and sampling has not changed, but the mechanics of applying it are evolving. Firms that have integrated AI-assisted audit tools can analyze full populations of transactions electronically before designing samples, which allows materiality judgments to be grounded in actual data rather than estimates. When all journal entries, accounts payable transactions, or revenue items can be screened for anomalies in advance, stratification becomes more precise, expected misstatement estimates improve, and sample sizes can be calibrated with greater confidence.

The result is not a different audit. It is the same risk-based, judgment-driven process, applied with better information and less manual effort. Clients in financial statement audit engagements increasingly benefit from this because source-linked workpapers make it straightforward to see exactly which items were tested and why.

Related reading: Generally Accepted Auditing Standards: GAAP vs. GAAS Explained

Frequently Asked Questions

What is materiality in an audit?

Materiality in an audit is the threshold above which a misstatement, either individually or in combination with other misstatements, could reasonably be expected to influence the economic decisions of financial statement users. Auditors set this threshold at the start of the engagement to focus their work on what genuinely matters. Both the AICPA (AU-C 320) and PCAOB (AS 2105) require auditors to determine materiality based on professional judgment applied to the specific circumstances of the entity, not a fixed rule.

How do auditors calculate materiality?

Auditors select a relevant benchmark, such as pre-tax income, total revenues, or total assets, and apply a percentage to arrive at a preliminary materiality figure. A commonly cited starting point is 5% of pre-tax income or 0.5% to 1% of total revenues, but these are guidelines, not requirements. The SEC’s SAB 99 framework requires auditors to also consider qualitative factors that could make a smaller amount material in context.

What is performance materiality and how does it differ from overall materiality?

Overall materiality is the threshold for the financial statements as a whole. Performance materiality is a lower working figure, typically 50% to 75% of overall materiality, set to provide a buffer that accounts for the aggregate of errors the auditor might not detect. It is what auditors actually use when designing individual procedures and sample sizes. If errors accumulate to near performance materiality in a given account, the auditor investigates further.

Why do auditors sample instead of testing everything?

Testing 100% of a large population is generally impractical and not proportionate to the additional assurance it provides. Audit sampling, governed by AU-C 530 and PCAOB AS 2315, allows auditors to draw reasonable conclusions about an entire population by testing a representative subset. The sample is designed to detect misstatements above tolerable misstatement with a quantifiable level of confidence. Items that individually exceed overall materiality are typically tested in full outside of the sampling process.

What happens when auditors find errors in a sample?

When a sample contains misstatements, auditors project those misstatements to the entire population and compare the projected total to tolerable misstatement. Under current AICPA standards, all misstatements found are projected, with no ability to dismiss findings as anomalies. If projected misstatement approaches or exceeds the materiality threshold, the auditor will expand testing, request that management correct the financial statements, or modify the audit opinion if corrections are not made.

Does materiality apply to nonprofit or government audits?

Yes, but with additional layers. For nonprofits that expend $1,000,000 or more in federal awards in a fiscal year (for years beginning on or after October 1, 2024), the Single Audit under Uniform Guidance requires a separate materiality assessment for compliance testing at the major program level. Government audits following the GAO Yellow Book also apply entity-level financial statement materiality alongside program-specific thresholds. The interplay between these different materiality levels makes nonprofit and governmental audits more complex than standard financial statement audits.