Attribute Sampling
How is attribute sampling different from variables sampling?
Attribute sampling classifies each unit as conforming or nonconforming; variables sampling uses measured numerical data such as dimensions or weight.
What is the core lot-disposition rule?
A sample of size n is drawn from a lot, each unit is inspected, and the lot is accepted if the count of nonconforming items is at or below the plan’s acceptance number c.
Why is random sampling important?
Because attribute sampling assumes the sample represents the lot; non-random selection can bias results and understate actual defect rates.
Attribute sampling is a pass/fail, conforming/non-conforming inspection method in which each sampled unit is classified by whether it has a specified characteristic rather than measured on a continuous scale. In acceptance sampling, a lot is accepted or rejected based on the number of nonconforming items found in a random sample, with systems such as ISO 2859 defining attribute sampling plans for lot inspection.
On a manufacturing shop floor, attribute sampling is used to check incoming lots, in-process batches, or finished goods for discrete defects such as missing labels, wrong part numbers, damaged packaging, incomplete paperwork, or absent traceability marks. In warehousing and receiving, it verifies whether sampled pallets, cases, or serialized units meet predefined criteria before put-away or release to production, reducing the need to inspect every unit. For raw material tracking, the method supports compliance checks on certificates of analysis, heat numbers, approval stamps, or lot-code linkage in the inventory system. The operational logic is straightforward: define the attribute, draw a random sample, count nonconforming units, and compare that count with the acceptance rule. This works well when the decision is fast and the characteristic is binary, such as label present or container seal intact.
False acceptance of a bad lot: If sample size is too small, a lot with hidden defects can pass because sampled units happen to be clean; this risk rises when defect distribution is uneven across the lot.
Poor sample integrity at receiving: Haphazard picks, top-layer-only sampling, or skipping damaged containers make the sample unrepresentative, and systematic packaging or transport damage goes undetected.
Traceability breaks on mixed or mislabeled stock: If attribute checks do not verify lot/heat/serial linkage, a visually acceptable item can enter inventory with broken genealogy, causing downstream release, recall, or substitution failures.