Audit Toolkit — Working Paper

Sampling Calculator

Determine sample sizes for tests of controls (attribute sampling) and substantive tests of details (monetary unit sampling), computed using standard statistical sampling methodology.

Ref: SA 530 / ISA 530 — Audit Sampling

01Attribute Sampling — Tests of Controls
02Monetary Unit Sampling — Substantive Tests

Used to determine how many items to test from a population when checking whether a control operated as designed (e.g. approval, three-way match, reconciliation review). Sample size is computed using exact binomial acceptance sampling — finding the smallest sample where, if the true deviation rate equalled your tolerable rate, the chance of observing this few (or fewer) deviations stays within your stated risk. This is a discrete, exact calculation (not an approximation), and it responds to changes in expected deviation rate as well as tolerable rate and confidence.

Total number of items/transactions in the period. Leave blank if unknown — not required for the calculation unless very small (<250).
Higher reliance planned on the control → use a higher confidence level.
Maximum deviation rate you can accept without concluding the control is ineffective.
Anticipated rate of deviation, based on prior year results or a pilot sample.
Recommended sample size
items
95%Confidence
Achieved risk of overreliance
Expected deviations in sample
Tolerable deviation rate
Precision gap (TDR − EDR)

If actual deviations found in the sample exceed , re-assess planned reliance on the control and consider extending the sample or revising the substantive approach.

Step 2

Select your sample from the population

Upload your completed Sales/Purchase Register template (the standard FinZfun template — columns are auto-detected, no mapping needed). We’ll randomly select items using an unbiased random number generator, with no data leaving your browser.

Used for substantive tests of details on account balances or classes of transactions (e.g. debtors, inventory value, revenue lines), selecting items with probability proportional to their value. Sample size uses a Poisson-based dollar-unit-sampling model that properly accounts for expected misstatement by iterating on the expected number of tainted monetary units — rather than a flat percentage adjustment, which understates the sample needed.

Total book value of the population being tested.
Linked to assessed risk of material misstatement — higher risk → higher confidence.
Usually set at or below performance materiality for this account.
Anticipated misstatement based on prior experience, or 0 if none expected.
Recommended sample size
items
95%Confidence
Sampling interval (₹)
Reliability factor used
Basic precision (₹)
Expected misstatement as % of TM

Select every th rupee interval starting from a random start point, and test the item in which that rupee falls. If tolerable misstatement is close to or below expected misstatement, sampling will not be efficient — consider 100% testing of high-value items instead.

Step 2

Select your sample from the population

Upload your completed Sales/Purchase Register template (columns are auto-detected). We’ll select items using systematic PPS selection at the sampling interval calculated above. Where a row has a Taxable Value filled in, that’s used as the sampling base — since Sales/Purchases are recorded net of GST in the books; rows missing it fall back to Invoice Value, and you’ll see how many rows fell back. No data leaves your browser.

Common Questions

FAQ — Sampling Calculator

Attribute sampling is used for tests of controls — it tells you how many items to check to conclude whether a control (like an approval or reconciliation) is operating as designed, and gives you a pass/fail rate. Monetary unit sampling is used for substantive tests of details on account balances — it selects items with a probability proportional to their rupee value, so higher-value items are more likely to be tested.

For attribute sampling, the reliability factor is derived from the Poisson distribution — the same statistical basis used to build published AICPA/ICAI attribute sampling tables. For monetary unit sampling, the confidence factor comes from standard probability-proportional-to-size (PPS) sampling theory.

The calculator will show a warning if expected exceeds tolerable, since sampling becomes inefficient in that case. When the two are close, the sample size increases sharply — at that point it’s often more efficient to test the population 100%, or to revisit whether the tolerable rate/misstatement set during planning is realistic.

For attribute sampling, if you enter a population under 250 items, the calculator caps the sample size at the population itself (you can’t sample more items than exist). For very small populations, consider whether 100% testing is more practical than statistical sampling.

No. This is a planning aid based on standard statistical sampling methodology under SA 530 / ISA 530. Final sample sizes remain a matter of professional judgment, and should be applied alongside your firm’s audit manual and engagement-specific risk assessment.

Yes. Fill in the Client, Period, and Prepared By fields at the top, run your calculation, then use the “Print / Save as PDF” button under the result to generate a printable working paper.

Scroll to Top