Attribute sampling for tests of controls, without a statistics textbook.
What it does
- Enter how confident you want to be, the highest failure rate you could accept, and how many failures you will allow in the sample. It returns the sample size.
- Adjusts for small populations, so you are not testing 60 items out of 70.
- Picks the items for you by systematic selection: a random start, then every k-th item, up to 500 items.
- A quick reference table of common sample sizes.
Good to know
At 95% confidence and a 5% tolerable rate with no failures allowed, it gives 60 items (59 on the exact formula). That matches the standard tables most firms use. Always check against your own firm's methodology.