CPA Exam Sampling and Analytical Procedures
Last updated: May 2, 2026
Sampling and Analytical Procedures questions are one of the highest-leverage areas to study for the CPA Exam. This guide breaks down the rule, the elements you need to recognize, the named traps that catch most students, and a memory aid that scales to test day. Read it once, then practice the same sub-topic adaptively in the app.
The rule
Audit sampling (AU-C 530) is the application of an audit procedure to less than 100% of the items in a population so the auditor can draw a conclusion about the entire population. Sampling can be used for tests of controls (attribute sampling) or substantive tests of details (variables sampling), and the auditor must evaluate both sampling risk (the risk the sample is unrepresentative) and non-sampling risk (procedure or judgment errors). Substantive analytical procedures (AU-C 520) develop an independent expectation of a recorded amount, define a tolerable difference, compare the expectation to the recorded amount, and investigate any difference exceeding the threshold. The two are not interchangeable — analytical procedures rely on plausible relationships among data, while sampling produces a projectable result.
Elements breakdown
Attribute Sampling (Tests of Controls)
A statistical sampling plan used to estimate the rate of deviation from a prescribed control in a population.
- Define the control and what constitutes a deviation
- Set tolerable deviation rate (TDR)
- Set expected population deviation rate (EPDR)
- Set acceptable risk of overreliance (low)
- Determine sample size from tables
- Select sample randomly or systematically
- Compute upper deviation rate (UDR)
- Compare UDR to TDR to conclude
Variables Sampling (Tests of Details)
A statistical or non-statistical plan used to estimate a monetary amount or misstatement in a population.
- Define the population and sampling unit
- Set tolerable misstatement (planning materiality based)
- Set expected misstatement
- Set acceptable risk of incorrect acceptance
- Choose method: MUS, classical variables, or non-statistical
- Select sample
- Project misstatement to population
- Compare projected misstatement plus allowance for sampling risk to tolerable misstatement
Common examples:
- Monetary Unit Sampling (MUS) for accounts receivable confirmations
- Classical variables for inventory pricing tests
Monetary Unit Sampling (MUS) Mechanics
A probability-proportional-to-size sampling method that selects each dollar in the population with equal probability.
- Each item's selection probability proportional to its book value
- Automatically stratifies toward larger items
- Effective when few or no misstatements expected
- Overstatements detected efficiently
- Understatements and zero/negative balances handled poorly
- Project misstatement using tainting percentage for partial errors
Substantive Analytical Procedures (AU-C 520)
Substantive procedures that evaluate financial information by studying plausible relationships among financial and non-financial data.
- Assess suitability for the assertion
- Evaluate reliability of data used
- Develop precise independent expectation
- Define amount of difference acceptable without investigation
- Compare expectation to recorded amount
- Investigate differences exceeding threshold
- Corroborate management explanations with evidence
Required Analytical Procedures
Two phases where analytical procedures are mandatory under GAAS.
- Risk-assessment phase during planning (AU-C 315)
- Overall review near the end of the audit (AU-C 520.06)
- Substantive use during fieldwork is optional but common
Sampling Risk vs Non-Sampling Risk
Two distinct risk categories the auditor must address.
- Sampling risk: sample not representative of population
- For controls: risk of overreliance (efficiency) and underreliance (effectiveness)
- For details: risk of incorrect acceptance (effectiveness) and incorrect rejection (efficiency)
- Non-sampling risk: wrong procedure, misinterpreted evidence, missed deviation
- Reduce sampling risk with larger samples; reduce non-sampling risk with training and supervision
Common patterns and traps
The Sampling Risk Direction Trap
Candidates routinely mix up the four sampling-risk outcomes. For tests of controls, the efficiency risk (auditor concludes controls are weaker than they are) is overreliance's opposite — underreliance. For substantive tests, incorrect rejection wastes time but incorrect acceptance lets a material misstatement slip through. The exam often asks which risk affects audit effectiveness versus efficiency. Effectiveness risks (overreliance, incorrect acceptance) are the dangerous ones because they lead to an unmodified opinion on a misstated set of statements.
A choice that says 'the auditor accepted the population when it was materially misstated, which affects audit efficiency' — wrong, that is incorrect acceptance and it affects effectiveness.
The Imprecise Expectation Trap
AU-C 520 requires that the auditor develop an expectation precise enough to identify a misstatement that, individually or combined with others, could be material. A wrong-answer choice will show the auditor 'comparing this year's revenue to last year's revenue' and concluding nothing looks unusual — that is not an analytical procedure performed at the required level of precision. The expectation must be independent, disaggregated where possible, and built from reliable data.
A choice that has the auditor accept the recorded balance because 'it is within 10% of last year' without disaggregation, without testing data reliability, and without a documented threshold.
The MUS Misapplication Trap
MUS efficiently detects overstatement in populations where most items are correctly recorded and there are few zero-or-negative balances. Wrong answers will apply MUS to test for understatement of accounts payable, or to a population dominated by zero-balance items. MUS will under-select small or zero items and produce a misleading conclusion. For completeness testing of liabilities, the auditor needs a different population (e.g., subsequent disbursements) or a different method.
A choice recommending MUS to test the completeness assertion for accounts payable, or to a population with many credit balances.
The 'Investigate Then Accept Management's Word' Trap
When an analytical procedure surfaces a difference exceeding the threshold, the auditor must investigate AND obtain corroborating audit evidence — not simply accept management's explanation. A choice that says 'auditor inquired of the controller, who explained the variance, so the auditor concluded the balance was reasonable' violates AU-C 520.07. Inquiry alone is not sufficient appropriate audit evidence to corroborate a significant fluctuation.
A choice in which the auditor closes a flagged variance with only an inquiry response from a member of management, without examining supporting documentation.
The Sample Size Driver Confusion
Sample size for tests of controls increases with higher expected deviation rate, lower tolerable deviation rate, and lower acceptable risk of overreliance. Population size has only a minor effect once the population is large. Wrong answers reverse one of these or claim that doubling the population doubles the sample. The exam tests whether you can identify which factor moves sample size up or down.
A choice claiming that a decrease in tolerable deviation rate decreases the required sample size.
How it works
Picture you are auditing Reyes Manufacturing's accounts receivable, recorded at $4,200,000 across 1,800 customer accounts. You decide to confirm a sample. Because larger balances carry more risk and MUS picks each dollar with equal probability, MUS will tend to land on the big customers automatically — it stratifies for you. After receiving replies, suppose you find one $50,000 invoice where the customer says only $40,000 is owed. The tainting is $10,000 / $50,000 = 20\%, and you project that taint across the sampling interval to estimate likely misstatement in the population. If your projected misstatement plus an allowance for sampling risk stays under tolerable misstatement, you accept the recorded balance. If you had instead used a substantive analytical procedure — say, estimating bad-debt expense by applying historical loss rates to aged receivables — you would compare your independent expectation to the recorded number and investigate any variance exceeding your threshold. The sample gives you a projectable monetary result; the analytical gives you corroborative evidence based on a relationship. Both are substantive, but they answer different questions.
Worked examples
Which sampling approach is most appropriate for the substantive testing of Liu Industries' accounts receivable balance?
- A Classical variables sampling, because the population is large and the auditor needs to estimate the standard deviation of misstatements.
- B Monetary unit sampling (MUS), because the auditor is testing primarily for overstatement, expects few misstatements, and the population contains no negative balances. ✓ Correct
- C Block sampling of all customer accounts whose balances were outstanding for more than 90 days, because aged balances pose the highest existence risk.
- D Haphazard selection of 25 accounts judged representative by the senior, applied without projecting the results, because the population is homogeneous.
Why B is correct: MUS is most appropriate when the auditor is primarily concerned with overstatement (the existence assertion for receivables), expects few misstatements, and faces a population with no zero or negative balances — all of which match Liu Industries' facts. MUS automatically stratifies because each dollar has equal probability of selection, so larger accounts receive proportionally greater coverage, and the projection mechanics handle low-error populations efficiently under AU-C 530.
Why each wrong choice fails:
- A: Classical variables sampling is more useful when the auditor expects a meaningful number of misstatements and needs to estimate their distribution; here, prior-year adjustments were minimal, making MUS more efficient. Classical variables also typically requires larger samples in low-error populations. (The MUS Misapplication Trap)
- C: Selecting only aged balances is a directed judgmental selection, not a sample that supports projection to the entire population. The conclusion would not address existence for the unselected portion of receivables, leaving most of the balance untested. (The Sampling Risk Direction Trap)
- D: Haphazard selection without projection cannot be evaluated statistically, and AU-C 530 requires the auditor to project misstatements found in the sample to the population. Twenty-five items chosen by feel does not produce a defensible conclusion about an $8.4 million balance. (The Sampling Risk Direction Trap)
What is the most appropriate next step for the auditor under AU-C 520?
- A Document the controller's explanation and conclude that payroll expense is reasonable, because the difference of $150,000 is below the $200,000 investigation threshold. ✓ Correct
- B Increase the precision threshold to $300,000 to reduce the work required, because management has provided a plausible explanation for the variance.
- C Reject the analytical procedure as ineffective and replace it with a 100\% recalculation of payroll for the year.
- D Investigate further by selecting a sample of fourth-quarter payroll registers and overtime authorizations to corroborate the controller's explanation, because management inquiry alone is not sufficient appropriate evidence.
Why A is correct: Under AU-C 520.07, the auditor must investigate differences exceeding the established threshold. Here, the $150,000 difference is below the $200,000 threshold, so further investigation is not required, and the procedure supports the recorded balance. The threshold was set during planning to reflect the difference that, individually or combined with other misstatements, could be material; staying within it permits the auditor to conclude payroll is reasonable.
Why each wrong choice fails:
- B: Adjusting the precision threshold after the fact to avoid investigation undermines the rigor of the analytical procedure and violates the spirit of AU-C 520. Thresholds are set during planning, not retroactively widened to accommodate a variance. (The Imprecise Expectation Trap)
- C: A 100\% recalculation of $14.2 million in payroll is grossly disproportionate when an analytical procedure with a precise expectation, reliable data, and a difference within threshold already provides sufficient evidence. The auditor would be expending effort with no incremental assurance.
- D: Corroborating management's explanation is required only when the variance exceeds the investigation threshold. Here the difference is within threshold, so the trigger for additional substantive procedures has not been reached. (The 'Investigate Then Accept Management's Word' Trap)
What conclusion should the auditor reach regarding reliance on this control?
- A Rely on the control, because the sample deviation rate of 2.6\% (2 ÷ 78) is well below the 6\% tolerable deviation rate.
- B Do not rely on the control as initially planned, because the upper deviation rate of 6.4\% exceeds the tolerable deviation rate of 6\%. ✓ Correct
- C Rely on the control, because the expected population deviation rate of 1\% is below the tolerable rate, and the two missing initials are isolated occurrences.
- D Increase the sample size by 25\% and re-evaluate, because the deviations are too few to draw a statistical conclusion.
Why B is correct: In attribute sampling, the auditor compares the upper deviation rate (UDR) — not the sample deviation rate — to the tolerable deviation rate (TDR). The UDR of 6.4\% incorporates an allowance for sampling risk and exceeds the 6\% TDR, so the auditor cannot rely on the control at the planned level under AU-C 530. The appropriate response is to reduce planned reliance and increase substantive testing over the related assertion.
Why each wrong choice fails:
- A: The sample deviation rate alone ignores sampling risk; AU-C 530 requires comparing the UDR (sample rate plus an allowance for sampling risk) to the TDR. Using the unadjusted sample rate would systematically overstate the auditor's basis for reliance. (The Sampling Risk Direction Trap)
- C: The expected deviation rate is an input used to determine sample size; it is not the figure used to evaluate the result. Treating two missing initials as 'isolated' without investigating their cause and projecting their impact violates AU-C 530's requirement to consider the qualitative nature of deviations. (The Sample Size Driver Confusion)
- D: The sample of 78 was determined to satisfy the auditor's planning parameters and produced an evaluable UDR. Expanding the sample is not the appropriate response when the UDR exceeds the TDR; the appropriate response is to reduce planned reliance.
Memory aid
PERC-I for analytical procedures: Plausibility, Expectation (independent), Reliability of data, Compare, Investigate. For sampling: 'TUE' — Tolerable rate, Upper rate, Evaluate (UDR ≤ TDR ⇒ rely).
Key distinction
Sampling tests a population by examining a subset and projecting results; analytical procedures test a recorded amount by comparing it to an independent expectation built from relationships. Sampling produces a measurable sampling-risk allowance; analytical procedures do not.
Summary
Sampling lets you project a deviation rate or a dollar misstatement from a sample to a population, while substantive analytical procedures corroborate a recorded amount against an independent expectation — pick the tool that matches the assertion and the available data.
Practice sampling and analytical procedures adaptively
Reading the rule is the start. Working CPA Exam-format questions on this sub-topic with adaptive selection, watching your mastery score climb in real time, and seeing the items you missed return on a spaced-repetition schedule — that's where score lift actually happens. Free for seven days. No credit card required.
Start your free 7-day trialFrequently asked questions
What is sampling and analytical procedures on the CPA Exam?
Audit sampling (AU-C 530) is the application of an audit procedure to less than 100% of the items in a population so the auditor can draw a conclusion about the entire population. Sampling can be used for tests of controls (attribute sampling) or substantive tests of details (variables sampling), and the auditor must evaluate both sampling risk (the risk the sample is unrepresentative) and non-sampling risk (procedure or judgment errors). Substantive analytical procedures (AU-C 520) develop an independent expectation of a recorded amount, define a tolerable difference, compare the expectation to the recorded amount, and investigate any difference exceeding the threshold. The two are not interchangeable — analytical procedures rely on plausible relationships among data, while sampling produces a projectable result.
How do I practice sampling and analytical procedures questions?
The fastest way to improve on sampling and analytical procedures is targeted, adaptive practice — working questions that focus on your specific weak spots within this sub-topic, getting immediate feedback, and revisiting items you missed on a spaced-repetition schedule. Neureto's adaptive engine does this automatically across the CPA Exam; start a free 7-day trial to see your sub-topic mastery climb in real time.
What's the most important distinction to remember for sampling and analytical procedures?
Sampling tests a population by examining a subset and projecting results; analytical procedures test a recorded amount by comparing it to an independent expectation built from relationships. Sampling produces a measurable sampling-risk allowance; analytical procedures do not.
Is there a memory aid for sampling and analytical procedures questions?
PERC-I for analytical procedures: Plausibility, Expectation (independent), Reliability of data, Compare, Investigate. For sampling: 'TUE' — Tolerable rate, Upper rate, Evaluate (UDR ≤ TDR ⇒ rely).
What's a common trap on sampling and analytical procedures questions?
Confusing sampling risk with non-sampling risk
What's a common trap on sampling and analytical procedures questions?
Using analytical procedures without an independent, precise expectation
Ready to drill these patterns?
Take a free CPA Exam assessment — about 25 minutes and Neureto will route more sampling and analytical procedures questions your way until your sub-topic mastery score reflects real improvement, not luck. Free for seven days. No credit card required.
Start your free 7-day trial