Audit sampling is a core AUD topic that tests your understanding of how auditors draw conclusions about entire populations by examining a subset of items. The CPA exam covers both the conceptual framework for sampling and the specific techniques used in practice, including attribute sampling for tests of controls and variables sampling for substantive testing. This guide walks through the key concepts, terminology, and decision-making processes you need to know for the exam.
What Is Audit Sampling?
Audit sampling is the application of audit procedures to less than 100% of items within an account balance or class of transactions. The purpose is to evaluate some characteristic of the balance or class and project the results to the entire population. Sampling is necessary because auditing every single transaction would be prohibitively expensive and time-consuming for most accounts.
Key principle: Any time the auditor applies a procedure to less than 100% of items and uses the results to draw a conclusion about the entire population, they are sampling. If the auditor examines only specific items (such as all transactions over a certain dollar amount), that is not sampling; it is targeted selection and the results cannot be projected to the population.
Statistical vs. Non-Statistical Sampling
The CPA exam tests the distinction between statistical and non-statistical sampling approaches.
Statistical Sampling
Statistical sampling uses mathematical probability theory to select sample items and evaluate results. It allows the auditor to quantify sampling risk and provides a statistically valid basis for projecting results. Statistical sampling requires random selection of sample items and the use of probability theory to evaluate results.
Advantages of statistical sampling:
- Provides an objective, measurable basis for quantifying sampling risk.
- Allows the auditor to design the most efficient sample for a given level of confidence.
- Results can be defended objectively because they are based on mathematical probability.
Non-Statistical Sampling
Non-statistical sampling (also called judgmental sampling) relies on the auditor's professional judgment to select sample items and evaluate results. The auditor uses their knowledge of the entity, the account, and the risks to determine sample size and selection methods.
Exam point: Both statistical and non-statistical sampling can provide sufficient appropriate audit evidence if properly applied. Neither approach is inherently superior. However, only statistical sampling allows the auditor to mathematically quantify sampling risk. The exam frequently tests this distinction.
Sampling Risk
Sampling risk is the risk that the auditor's conclusion based on a sample differs from the conclusion they would reach if they examined the entire population. Sampling risk exists whenever the auditor does not examine 100% of the population.
There are two types of sampling risk for tests of controls:
- Risk of assessing control risk too low (risk of overreliance): The sample results suggest the control is effective when it is actually not. This risk affects audit effectiveness because the auditor may reduce substantive testing based on an incorrect assessment.
- Risk of assessing control risk too high (risk of underreliance): The sample results suggest the control is not effective when it actually is. This risk affects audit efficiency because the auditor performs more substantive testing than necessary.
For substantive testing, the two types of sampling risk are:
- Risk of incorrect acceptance: The sample results suggest an account balance is not materially misstated when it actually is. This is the more dangerous risk because it affects audit effectiveness.
- Risk of incorrect rejection: The sample results suggest an account balance is materially misstated when it is not. This affects audit efficiency, leading to unnecessary additional work.
Exam tip: The risks that affect audit effectiveness (risk of overreliance and risk of incorrect acceptance) are the primary concerns because they can lead to an inappropriate audit opinion. The risks that affect only efficiency cause extra work but do not compromise audit quality.
Attribute Sampling
Attribute sampling is used primarily for tests of controls. It tests the rate of deviation from a prescribed control procedure in a population. The result is expressed as a percentage, called the sample deviation rate, which is projected to the population.
Key terms for attribute sampling:
- Expected population deviation rate: The auditor's estimate of the deviation rate in the population before sampling. This affects sample size.
- Tolerable deviation rate: The maximum rate of deviations the auditor is willing to accept while still concluding the control is effective. This is set before sampling.
- Allowance for sampling risk: The difference between the tolerable rate and the expected rate. A larger allowance means a smaller required sample.
- Sample deviation rate: The actual rate of deviations found in the sample.
- Upper deviation rate: The sample deviation rate plus an allowance for sampling risk. If the upper deviation rate exceeds the tolerable rate, the auditor concludes the control is not effective.
Factors that increase sample size in attribute sampling:
- Higher confidence level desired (lower risk of overreliance)
- Higher expected population deviation rate
- Lower tolerable deviation rate
- Larger population size (though this has diminishing effect beyond a certain point)
Variables Sampling
Variables sampling is used for substantive testing of account balances. It estimates the dollar amount of misstatement in a population. The three primary variables sampling methods are:
Mean-Per-Unit (MPU) Estimation
MPU calculates the average audited value of items in the sample and multiplies by the number of items in the population to estimate the total audited value. The estimated misstatement is the difference between the estimated audited value and the recorded book value.
Difference Estimation
Difference estimation calculates the average difference between the audited value and the book value for each sample item, then projects this average difference to the entire population. This method is most effective when misstatements are expected and vary in amount.
Ratio Estimation
Ratio estimation calculates the ratio of audited values to book values in the sample and applies this ratio to the total book value of the population. It works best when the size of misstatements is proportional to the book value of the items.
Monetary Unit Sampling (MUS)
Monetary unit sampling, also called probability proportional to size (PPS) sampling, is a commonly used statistical approach for substantive testing. It treats each individual dollar in the population as a potential sampling unit, which means larger account balances have a higher probability of selection.
Characteristics of MUS that the exam tests:
- Automatically stratifies the sample by dollar value, giving larger items greater chance of selection.
- Effective for detecting overstatements. It is not well-suited for testing for understatements because understated items (smaller than they should be) have a lower chance of selection.
- Requires minimal knowledge of the population characteristics to design the sample.
- Assumes zero misstatements in the population when determining sample size, which simplifies planning but can be limiting if many misstatements are expected.
- Uses a table or formula based on reliability factors to calculate sample size.
Sample Size Determination
Several factors influence the required sample size for any sampling application:
- Confidence level: Higher confidence (lower sampling risk) requires a larger sample.
- Tolerable misstatement or deviation rate: A smaller tolerable amount requires a larger sample to detect it.
- Expected misstatement or deviation rate: Higher expected errors require a larger sample.
- Population size: For large populations, this has minimal effect on sample size in statistical sampling. For smaller populations, the finite population correction factor may reduce the required sample.
- Population variability: Greater variability in the population (larger standard deviation) requires a larger sample in variables sampling.
Evaluating Sample Results
After performing the sampling procedures, the auditor evaluates the results:
- Calculate the projected misstatement or deviation rate for the population based on sample results.
- Compare the projected misstatement to the tolerable misstatement (or compare the upper deviation rate to the tolerable deviation rate for attribute sampling).
- Consider the qualitative aspects of identified misstatements. Even if the projected misstatement is below tolerable, the nature of the misstatements may indicate a control problem or fraud risk.
- Determine whether the results support the planned conclusion or whether additional procedures are needed.
Exam point: If the projected misstatement exceeds the tolerable misstatement, the auditor should consider requesting management to examine the population and make corrections, expanding the sample, performing alternative procedures, or considering the effect on the audit opinion.
Common Exam Questions on Sampling
- Which type of sampling risk is most concerning for audit effectiveness?
- What happens to sample size when the tolerable deviation rate decreases?
- What is the primary difference between attribute and variables sampling?
- When is MUS most effective and what is its main limitation?
- How does the auditor evaluate results when the upper deviation rate exceeds the tolerable rate?
Sharpen Your Sampling Knowledge with Think CPA
Audit sampling has its own vocabulary and logic that can feel unfamiliar at first. Think CPA provides targeted practice on attribute sampling, variables sampling, MUS, and result evaluation, helping you build fluency with the terminology and decision-making frameworks the CPA exam expects. If sampling has been a confusing topic in your studies, structured practice with Think CPA can help it click into place.
Final Thoughts
Audit sampling is one of those AUD topics that rewards understanding over memorization. Know the difference between statistical and non-statistical sampling, understand the four types of sampling risk, and be able to apply attribute sampling to tests of controls and variables sampling to substantive tests. Focus on the relationships between sample size factors and the evaluation of results. With this knowledge, sampling questions become approachable and predictable on the CPA exam.