AUD · Performing Further Procedures and Obtaining Evidence · 7 practice questions
MUS vs. Classical Variables vs. Attribute Sampling
Use attribute sampling for deviation rates, MUS for positive-dollar overstatement testing, and classical variables for two-sided monetary estimates. Below: a side-by-side comparison showing how the objective and receivable balances change the choice.
Try one first
Hint
Focus on whether the sampling method can select credit/negative balances and whether the auditor's objective requires measuring both overstatements and understatements.
Answer C. MUS (probability-proportional-to-size) is oriented toward detecting overstatements and does not appropriately select or weight negative (credit) balances, which can bias an estimate of net misstatement. Because the auditor's objective is to estimate total (net) dollar misstatement and the population includes a meaningful number of credit balances, a variable-sampling approach that accommodates negatives (for example, classical variables sampling) is more appropriate.
Why not A: Periodicity can be a concern for systematic methods, but MUS uses a random start to reduce that risk, and the stem gives no indication of a problematic posting cycle. The dominant issue here is MUS's unsuitability for populations with a meaningful number of credit balances when estimating net misstatement.
Why not B: Skewness and stratification are legitimate concerns, but MUS intentionally gives greater probability to large items and often performs well on skewed populations when the goal is detecting overstatements. The stem's more specific and decisive problem is the presence of negative balances coupled with a net-misstatement objective.
Why not D: Attribute sampling estimates rates of deviations (for example, control failures or occurrence exceptions), not dollar misstatements. Because the auditor's objective is to estimate net dollar misstatement, attribute sampling is not an appropriate substitute.
Side by side
| Attribute sampling | Monetary-unit sampling (MUS/PPS) | Classical variables sampling | |
|---|---|---|---|
| Receivables question | How often are required invoice approvals missing? | Are recorded receivables overstated? | What is the dollar misstatement, considering both directions? |
| Estimate sought | Control deviation rate. | Monetary overstatement. | Dollar misstatement, including overstatements and understatements. |
| What gets sampled | Invoices selected for a yes/no approval check. | Recorded positive dollars; inspect the related invoice or account. | Invoice or account items. |
| Treatment of large balances | Dollar size does not determine selection weight for the approval-rate estimate. | Larger recorded balances have a greater chance of selection. | May stratify or test large balances separately. |
| Best-fit error pattern | Missing approvals measured as a proportion of invoices. | Low-frequency overstatements concentrated in large balances. | Two-sided errors or many small errors unrelated to item size. |
| Zero and credit balances | Dollar sign does not change the yes/no approval check. | Test separately; ordinary positive-dollar MUS does not appropriately cover them. | Can include them in the monetary estimate. |
| Supported conclusion | Estimates the population's missing-approval rate, not dollar misstatement. | Supports a monetary-overstatement conclusion, not an approval deviation-rate estimate. | Estimates population dollar misstatement in either direction. |
Key points
- Remember: Ask 'a rate or dollars?' first; then check whether the dollar population fits MUS.
- Checking approval is yes/no, but counting exceptions in a dollar-weighted sample does not estimate the invoice deviation rate.
- Zero or credit balances can be tested separately if MUS is used only for the positive-balance portion.
How the exam traps you
- Treating exceptions from a MUS sample as an invoice approval deviation rate. Use attribute sampling to estimate the missing-approval rate.
- Choosing MUS for the entire population just because a few receivables dominate the total. For one two-sided dollar estimate including zero and credit balances, use classical variables.
Question 2
Hint
Decide first whether the objective is to detect overstatement or understatement and whether the population contains many small, zero, or negative monetary amounts; then pick the sampling method that can estimate dollar misstatements in that context.
Answer D. Classical variables sampling is used to estimate the dollar amount of misstatement (both overstatements and understatements) and can be applied with stratification to handle many small items and zero/negative balances. MUS/PPS selects items probability‑proportional‑to‑size and is efficient for detecting overstatement in populations of positive balances but does not provide reliable coverage for unrecorded liabilities or populations with many zeros/negatives. Attribute sampling addresses yes/no compliance questions rather than estimating dollar misstatement. Nonstatistical haphazard selection cannot quantify sampling risk and may miss systematic understatement among many small invoices.
Why not A: Tempting because MUS efficiently targets high‑dollar items, but MUS is PPS and is primarily suited to overstatement testing in populations of positive balances; it performs poorly with many zero or negative amounts and is not appropriate when the objective is to detect unrecorded (understated) liabilities.
Why not B: Tempting because auditors often examine large or recent invoices, but judgmental or haphazard selection is not a probability sample, does not quantify sampling risk, and may overlook systematic understatement among many small vendor invoices.
Why not C: Tempting because checking whether invoices are supported sounds directly relevant, but attribute sampling measures compliance (presence/absence) and does not estimate the dollar magnitude of unrecorded liabilities; it won't provide a monetary estimate of understatement.
Question 3
Hint
Decide whether selection probabilities varied by invoice dollar amount and whether the method is intended to estimate monetary misstatement or an attribute (deviation) rate.
Answer A. A method that gives larger invoices a higher chance of selection is monetary‑unit (probability‑proportional‑to‑size) sampling. MUS/PPS is designed to estimate monetary misstatements and overweights large items, so it does not provide an unbiased estimate of an attribute (deviation) rate. For testing authorization (an attribute), equal‑probability statistical attribute sampling is the appropriate approach.
Why not B: This is tempting because attribute sampling is the correct type of sampling to estimate a deviation rate; however it misclassifies the technique actually used, the stem describes dollar‑proportional selection, not equal‑probability attribute sampling.
Why not C: This distractor appeals because focusing on large invoices sounds risk‑targeted, but it is wrong for two reasons: the described method was probability‑based (dollar‑proportional), not judgmental, and a judgmental large‑item sample would still be biased and unsuitable for estimating a deviation rate.
Why not D: This is tempting because systematic equal‑probability sampling is a valid attribute method for estimating deviation rates; however it misclassifies the sampling technique used, the auditors used a dollar‑weighted selection approach, not an equal‑probability systematic attribute sample.
Question 4
Hint
Do not stop at 'positive balances' or 'wide variation.' Focus on what relationship the stem gives between book amounts and likely audited amounts, and whether the auditor wants an overstatement-focused method or a two-sided dollar estimate.
Answer A. Ratio estimation is appropriate when audited amounts are expected to vary proportionally with recorded amounts because it uses the relationship between book and audited amounts to estimate total dollar misstatement. The stem signals proportional misstatements and explicitly requests a two-sided dollar estimate, which points to a classical variables approach such as ratio estimation rather than PPS or mean-per-unit methods. Difference estimation and mean-per-unit are less efficient when misstatements are proportional to recorded amounts.
Why not B: PPS (monetary-unit) sampling does emphasize larger balances and requires positive book values, which makes it tempting here. However, PPS is primarily oriented toward detecting overstatement and is not the most natural choice when the auditor specifically wants a two-sided total-dollar estimate and when misstatements are expected to be proportional, scenarios better served by classical variables methods such as ratio estimation.
Why not C: Difference estimation projects an average dollar difference per item and is appropriate when misstatements tend to be roughly a constant dollar amount per item. The stem indicates misstatements are roughly proportional to recorded amounts, so ratio estimation is more efficient than difference estimation in this case.
Why not D: Mean-per-unit ignores the relationship between recorded and audited amounts and simply projects the average audited amount; it is less efficient when misstatements vary proportionally with recorded amounts. Because the stem signals that recorded amounts still matter, mean-per-unit is not the best choice here.
Question 5
Hint
Focus on what the auditor wants the sample to measure, not on factors that mainly change sample size.
Answer A. The primary determinant of the sampling technique is the audit objective. Attribute sampling is typically used for tests of controls that measure deviation rates, while variables sampling is used in substantive tests to estimate monetary misstatement. Other factors (population size, expected errors, choice of statistical vs nonstatistical) affect sample design or size but do not change this basic method choice.
Why not B: Population size influences how large a sample may need to be but does not determine whether the auditor should use attribute versus variables sampling. The method is chosen based on whether the auditor is assessing deviations (attribute) or estimating dollar misstatements (variables).
Why not C: Choosing between statistical and nonstatistical approaches is a separate design decision and can apply to either attribute or variables sampling. The first decision is whether the test is about deviations (controls) or amounts (substantive), which governs attribute vs variables.
Why not D: Expected error rates affect sample size and risk assessments but do not primarily dictate the sampling technique. The key distinction remains whether the test measures control performance (deviation rate) or estimates monetary misstatement.
Question 6
Hint
Focus on the audit objective: Is the auditor trying to estimate a dollar misstatement or the frequency of a control deviation?
Answer B. Attribute sampling is used when the auditor wants to estimate the rate of occurrence of a control deviation (a yes/no characteristic), such as missing approvals on vouchers. The objective described is to estimate a deviation frequency in a test of controls, not to estimate a monetary misstatement, so attribute sampling is the appropriate technique.
Why not A: Variables sampling is used in substantive testing to estimate the dollar amount of a monetary misstatement (e.g., mean or total), not to estimate a rate of control deviations, so it is not appropriate here.
Why not C: Monetary-unit sampling (MUS) is designed to test for monetary overstatement and gives higher-dollar items a greater chance of selection; it is not used to estimate the frequency of control deviations.
Why not D: Block sampling selects contiguous items (e.g., from a single period or sequence) and is a selection method rather than the standard technique for estimating a population deviation rate; it may not produce a representative estimate of a control deviation rate.
Question 7
Hint
Focus on whether the sampling unit is the customer account or the individual dollar.
Answer A. Monetary-unit sampling (MUS) treats each dollar as a sampling unit (a form of probability-proportional-to-size sampling), so larger recorded balances have a higher probability of selection. That makes MUS well suited when the auditor wants automatic emphasis on higher-dollar items without separately stratifying the population. (A minor limitation: MUS is less effective for populations with many negative or zero balances, but that does not affect the core choice here.)
Why not B: Simple random sampling gives each account an equal chance of selection, not each dollar, so it does not automatically weight selection toward larger balances. It is unbiased but won't prioritize high-dollar accounts as MUS does.
Why not C: Block sampling selects clusters (for example, consecutive items or a specific time block) and is not a value-weighted, probability-proportional-to-size method. Focusing only on one block at month-end may miss large balances recorded at other times and does not systematically increase the chance for larger-dollar items across the population.
Why not D: Haphazard sampling is nonstatistical and unsystematic; it does not ensure that larger-dollar items have a greater probability of selection. It relies on the auditor's informal judgment and therefore fails the value-weighting requirement in the stem.
Common questions
What is the difference between attribute sampling and MUS?
Attribute sampling estimates the rate of invoices missing approval. MUS weights selection by recorded dollars and supports conclusions about monetary overstatement.
When should I use classical variables sampling instead of MUS?
Use classical variables when you need a two-sided dollar estimate or one monetary sample covering positive, zero, and credit balances. Skewness alone does not override those needs.
Is MUS appropriate for a highly skewed population?
Yes, when balances are positive and overstatement risk is concentrated in large items. Larger recorded balances receive a greater chance of selection.
Watch it solved
A full CPA AUD task-based simulation on Sampling Techniques, worked step by step.
AUD Simulation: Attribute Sampling, Deviations and Upper Limit on YouTube
Practice AUD like the real exam
The free ChatCPA simulator: real exam layout, timed testlets, starting with a question on this topic. No account needed to start.
Open the free simulator →More on Sampling Techniques
Questions from the ChatCPA bank of 17,658 CPA exam questions, each with a written reason for every wrong answer. ChatCPA is built by Nicholas Miller, CPA (Oregon #14907). How these pages are made. Spot an error? Tell us.