Audit Sampling Interview Questions & Answers

Q: What is audit sampling?

Audit sampling is applying audit procedures to less than 100% of items in a population such that each sampling unit has a chance of selection, to draw a conclusion about the whole population. It lets auditors gather sufficient evidence efficiently when testing every item is impractical, accepting some sampling risk.

Q: What is the difference between statistical and non-statistical sampling?

Statistical sampling uses random selection and probability theory to measure sampling risk and quantitatively project results, giving objective, defensible conclusions. Non-statistical (judgmental) sampling relies on the auditor's judgment for selection and size without measuring sampling risk mathematically. Both must yield representative samples; statistical adds measurability.

Q: What is sampling risk and non-sampling risk?

Sampling risk is the risk that the auditor's conclusion from a sample differs from the conclusion had the whole population been tested (because the sample isn't fully representative). Non-sampling risk is the risk of a wrong conclusion for reasons unrelated to sample size—like using an inappropriate procedure or misinterpreting evidence. Larger samples reduce sampling risk; better performance reduces non-sampling risk.

Q: What are the risks of over-reliance and under-reliance (for tests of controls)?

In tests of controls, the risk of over-reliance (or incorrect acceptance) is concluding controls are effective when they aren't—it affects audit effectiveness (a wrong opinion risk). The risk of under-reliance (incorrect rejection) is concluding controls are ineffective when they are—it affects efficiency (unnecessary extra work). Over-reliance is the more serious concern.

Q: What factors affect sample size?

For substantive tests, sample size increases with higher assessed risk of material misstatement, higher required assurance, and larger expected misstatement, and decreases with a larger tolerable misstatement and reliance on other procedures. For tests of controls, size increases with a lower tolerable deviation rate, higher expected deviation rate, and higher required confidence. Population size has little effect for large populations.

Q: What is tolerable misstatement and tolerable rate of deviation?

Tolerable misstatement is the maximum misstatement in a population the auditor will accept in substantive testing (usually set at or below performance materiality). Tolerable rate of deviation is the maximum rate of control failures the auditor will accept while still relying on the control. Both set the threshold against which sample results are evaluated.

Q: What is monetary unit sampling (MUS) and when is it used?

MUS (probability-proportional-to-size) selects sampling units as individual monetary units, so larger-value items have a higher chance of selection. It's efficient for detecting overstatement in populations with few large errors (e.g., receivables, inventory), automatically emphasizing high-value items, and provides a statistical projection of misstatement. It's less suited to testing understatement or populations with many small errors.

Q: How do you project (extrapolate) misstatements from a sample?

Project the misstatements found in the sample to the whole population—e.g., in classical variables sampling, ratio or difference estimation; in MUS, using the tainting of sampled units and the sampling interval. Add an allowance for sampling risk to get an upper misstatement limit, and compare it to tolerable misstatement to conclude.

Q: What are common sample selection methods?

Methods include random selection (every item equal chance), systematic selection (every n-th item after a random start), monetary unit selection (value-weighted), haphazard selection (no conscious bias, for non-statistical), and block selection (contiguous items—generally discouraged as unrepresentative). The method must give a representative sample.

Q: What is stratification and why is it used?

Stratification divides a population into sub-populations (strata) with similar characteristics—commonly by value—so each is sampled separately. High-value items may be tested 100% while lower-value strata are sampled. This reduces variability within strata, improves efficiency, and focuses effort where misstatement would matter most.

What is audit sampling?

Beginner

Audit sampling is applying audit procedures to less than 100% of items in a population such that each sampling unit has a chance of selection, to draw a conclusion about the whole population. It lets auditors gather sufficient evidence efficiently when testing every item is impractical, accepting some sampling risk.

Real-world example Rather than test all 10,000 invoices, the auditor tests a representative sample and infers the population's error rate.

Common follow-ups: Why sample instead of test everything? | What is a sampling unit?

Audit Evidence & Procedures Audit Risk & Materiality Audit Sampling

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

Beginner

Statistical sampling uses random selection and probability theory to measure sampling risk and quantitatively project results, giving objective, defensible conclusions. Non-statistical (judgmental) sampling relies on the auditor's judgment for selection and size without measuring sampling risk mathematically. Both must yield representative samples; statistical adds measurability.

Real-world example The auditor uses statistical sampling to quantify the projected error and its precision, versus a judgmental pick for a small test.

Common follow-ups: What does statistical sampling let you measure? | Must non-statistical samples be representative?

Audit Evidence & Procedures Audit Risk & Materiality Audit Sampling

What is sampling risk and non-sampling risk?

Beginner

Sampling risk is the risk that the auditor's conclusion from a sample differs from the conclusion had the whole population been tested (because the sample isn't fully representative). Non-sampling risk is the risk of a wrong conclusion for reasons unrelated to sample size—like using an inappropriate procedure or misinterpreting evidence. Larger samples reduce sampling risk; better performance reduces non-sampling risk.

Real-world example A sample happening to miss the errors gives false comfort (sampling risk); misreading a document gives a wrong result (non-sampling risk).

Common follow-ups: How do you reduce each type of risk? | Which risk relates to sample size?

Audit Risk & Materiality Audit Evidence & Procedures Audit Sampling

What are the risks of over-reliance and under-reliance (for tests of controls)?

Intermediate

In tests of controls, the risk of over-reliance (or incorrect acceptance) is concluding controls are effective when they aren't—it affects audit effectiveness (a wrong opinion risk). The risk of under-reliance (incorrect rejection) is concluding controls are ineffective when they are—it affects efficiency (unnecessary extra work). Over-reliance is the more serious concern.

Real-world example Concluding a control works when it doesn't (over-reliance) could let a material misstatement pass—worse than doing extra testing.

Common follow-ups: Which risk threatens effectiveness? | Why is over-reliance the bigger concern?

Internal Controls Evaluation Audit Risk & Materiality Audit Sampling

What factors affect sample size?

Intermediate

For substantive tests, sample size increases with higher assessed risk of material misstatement, higher required assurance, and larger expected misstatement, and decreases with a larger tolerable misstatement and reliance on other procedures. For tests of controls, size increases with a lower tolerable deviation rate, higher expected deviation rate, and higher required confidence. Population size has little effect for large populations.

Substantive: higher risk / lower tolerable error -> larger sample.
Controls: lower tolerable deviation / higher confidence -> larger sample.

Real-world example Because receivables are high risk with a low tolerable misstatement, the auditor selects a larger confirmation sample.

Common follow-ups: Does population size drive sample size? | How does tolerable error affect size?

Audit Risk & Materiality Audit Evidence & Procedures Audit Sampling

What is tolerable misstatement and tolerable rate of deviation?

Intermediate

Tolerable misstatement is the maximum misstatement in a population the auditor will accept in substantive testing (usually set at or below performance materiality). Tolerable rate of deviation is the maximum rate of control failures the auditor will accept while still relying on the control. Both set the threshold against which sample results are evaluated.

Real-world example With tolerable misstatement of $75,000, a projected sample error of $30,000 is acceptable, whereas $90,000 would not be.

Common follow-ups: How does tolerable misstatement relate to materiality? | What does the tolerable deviation rate govern?

Audit Risk & Materiality Internal Controls Evaluation Audit Sampling

What is monetary unit sampling (MUS) and when is it used?

Advanced

MUS (probability-proportional-to-size) selects sampling units as individual monetary units, so larger-value items have a higher chance of selection. It's efficient for detecting overstatement in populations with few large errors (e.g., receivables, inventory), automatically emphasizing high-value items, and provides a statistical projection of misstatement. It's less suited to testing understatement or populations with many small errors.

Sampling interval = Tolerable misstatement / reliability factor
Every n-th monetary unit is selected (large items sampled more).

Real-world example Testing receivables for overstatement, the auditor uses MUS so big balances are almost certain to be selected.

Common follow-ups: Why does MUS suit overstatement testing? | What is the sampling interval?

Audit Risk & Materiality Audit Evidence & Procedures Audit Sampling

How do you project (extrapolate) misstatements from a sample?

Advanced

Project the misstatements found in the sample to the whole population—e.g., in classical variables sampling, ratio or difference estimation; in MUS, using the tainting of sampled units and the sampling interval. Add an allowance for sampling risk to get an upper misstatement limit, and compare it to tolerable misstatement to conclude.

Simple projection (ratio):
  Sample error / Sample value x Population value = projected error.

Real-world example A 2% error rate in the sampled dollars projects to a $40,000 estimated misstatement across the population.

Common follow-ups: What is added for sampling risk? | How do you conclude from the projection?

Audit Risk & Materiality Audit Evidence & Procedures Audit Sampling

What are common sample selection methods?

Beginner

Methods include random selection (every item equal chance), systematic selection (every n-th item after a random start), monetary unit selection (value-weighted), haphazard selection (no conscious bias, for non-statistical), and block selection (contiguous items—generally discouraged as unrepresentative). The method must give a representative sample.

Real-world example The auditor uses systematic selection with a random start to pick every 50th invoice for testing.

Common follow-ups: Why is block selection discouraged? | Which methods suit statistical sampling?

Audit Evidence & Procedures Audit Risk & Materiality Audit Sampling

What is stratification and why is it used?

Intermediate

Stratification divides a population into sub-populations (strata) with similar characteristics—commonly by value—so each is sampled separately. High-value items may be tested 100% while lower-value strata are sampled. This reduces variability within strata, improves efficiency, and focuses effort where misstatement would matter most.

Real-world example The auditor tests all balances over $100k individually and samples the remaining smaller balances separately.

Common follow-ups: How does stratification improve efficiency? | Why test high-value items fully?

Audit Risk & Materiality Audit Evidence & Procedures Audit Sampling