Acceptance sampling decides whether to accept a lot from a small sample. The acceptance sampling calculator takes the sample size, the acceptance number, the AQL, the LTPD, the incoming defect rate and the lot size, and gives the probability of acceptance at each quality level, the producer’s risk, the consumer’s risk, the average outgoing quality and the average total inspected.
How the calculation works
- A lot is accepted if a random sample of n items contains c or fewer defective items. The number of defects in the sample follows a binomial distribution with the lot’s defect rate p.
- Probability of acceptance Pa = P(defects ≤ c), with n trials and probability p. At the AQL this gives the producer’s risk = 1 − Pa, the chance of rejecting a good lot.
- At the LTPD (the worst quality the customer will tolerate), Pa is the consumer’s risk, the chance of accepting a bad lot.
- With rectifying inspection, rejected lots are fully inspected and defects removed. AOQ = Pa × p × (N − n) ÷ N, and the average total inspected is n + (1 − Pa) × (N − n).
Example
A plan takes a sample of 80 from lots of 2,000 and accepts the lot with 2 or fewer defects. At an AQL of 1 %, Pa is 95.34 %, so the producer’s risk is 4.66 %. At an LTPD of 6 %, Pa is 13.44 %, which is the consumer’s risk. A lot with 3 % defects is accepted 56.81 % of the time, the average outgoing quality is 1.636 %, and on average 909.2 items per lot are inspected.
Why it matters
Choosing a sampling plan by feel can leave you accepting bad lots or rejecting good ones too often. The calculator shows exactly how a plan behaves at the quality levels that matter to you and to your supplier, before you adopt it in a contract.
Good to know
The binomial model is exact for a large lot and a good approximation when the sample is a small part of the lot. For small lots, the hypergeometric distribution is more precise. The calculator treats the AQL and LTPD as defective percentages, and it covers a single sampling plan by attributes, not double or sequential plans.
Related tools
The defect density calculator measures defects per unit of size, the control chart limits calculator monitors a process over time and the binomial calculator gives the underlying probabilities.