When would you prefer attribute sampling versus monetary-unit sampling?

When would you prefer attribute sampling versus monetary-unit sampling?

• A. Attribute sampling for tests of controls; monetary-unit sampling for populations with unknown error rates.
• B. Attribute sampling for substantive tests; monetary-unit sampling for deterministic populations.
• C. Attribute sampling only for large populations; monetary-unit sampling only for small populations.
• D. Attribute sampling and monetary-unit sampling are interchangeable.

Answer: (A)

Attribute sampling is used to test a binary condition—whether a control works or not—so it’s the natural choice for evaluating the operating effectiveness of internal controls. Monetary-unit sampling treats each dollar as a sampling unit and uses probability proportional to size to estimate total misstatement, which is particularly efficient when you don’t know the error rate and you want to infer the overall dollar impact. Because of this, you’d select attribute sampling for tests of controls and monetary-unit sampling when the population’s error rate is unknown. The other options mix up how these methods are used—attribute sampling isn’t the standard tool for substantive tests, MUS isn’t limited to deterministic populations, and the methods aren’t interchangeable.

Attribute sampling is used to test a binary condition—whether a control works or not—so it’s the natural choice for evaluating the operating effectiveness of internal controls. Monetary-unit sampling treats each dollar as a sampling unit and uses probability proportional to size to estimate total misstatement, which is particularly efficient when you don’t know the error rate and you want to infer the overall dollar impact. Because of this, you’d select attribute sampling for tests of controls and monetary-unit sampling when the population’s error rate is unknown. The other options mix up how these methods are used—attribute sampling isn’t the standard tool for substantive tests, MUS isn’t limited to deterministic populations, and the methods aren’t interchangeable.