SC QUALITY GUILDAudit sampling plan
SCQUALITYGUILD.COM/TOOLS/AUDIT-SAMPLING-PLAN.HTML
Size an attribute sample of records or transactions for a stated confidence and tolerable error rate, share it across strata, draw the items, and state what the result supports.
ON THE EXAM · CQA II.A.5, V.E.2, V.E.3 · FREE · NOTHING YOU TYPE LEAVES YOUR BROWSER
Q-BotClick for an explanation of this tool.For educational use. This page is a study aid. The method behind it has far more depth and detail than one page and one tool can cover, and what you see here is not everything the method offers or everything this site offers. For full understanding and application, use the governing standards and texts, and seek further education, training or qualified advice. Results depend on the data you enter, so check them before relying on them.
Judgmental and statistical sampling
Auditors rarely check every record, so the conclusion rests on a sample. A judgmental sample is chosen by the auditor's experience: the rush orders, the new supplier, the night shift. It is good at finding problems but supports no statement about the whole population. A statistical sample gives every item a known chance of selection, so the result can be generalized with a stated confidence. Many audits combine the two and report them separately.
The zero-failure sample size
For attribute testing (each item is right or wrong), pick a confidence C and a tolerable error rate p. If the true error rate were p, the chance that n items all pass is (1 − p)n. Setting that equal to 1 − C gives
n = ln(1 − C) / ln(1 − p), rounded up.
At 95 percent confidence and a 5 percent tolerable rate, n = ln(0.05) / ln(0.95) = 58.4, so 59 items. At 95 percent and 10 percent, n = 29. The formula assumes a large population; for a small one the exact (hypergeometric) figure is a little lower, which the tool also shows.
What a clean sample tells you
If n items show no errors, the error rate is no more than 1 − (1 − C)1/n at confidence C. Fifty-nine clean items give an upper bound of about 4.95 percent at 95 percent confidence. It does not show the error rate is zero. A quick check is the rule of three: at 95 percent, the bound is roughly 3/n. If errors are found, the plan has failed; each error is evidence to evaluate, and the sample should not be extended until the rate looks acceptable.
Selecting the items
- Simple random: random numbers against a numbered list. Record the seed.
- Systematic: every k-th item (k = N / n) from a random start. Watch for cycles in the list that match the interval.
- Stratified: split the population into groups that may differ (sites, shifts, months) and sample each, usually in proportion to its size.
On the CQA exam
Expect to tell judgmental from statistical sampling, pick the method that fits a scenario, and explain what a sample result does and does not let the auditor conclude.