Inspection and Testing
Inspection is appraisal - comparing product to a stated requirement by gauge, visual check, or measurement - while testing typically applies a stimulus and measures the response (functional, environmental, destructive). Know the standard inspection points: receiving/incoming, first-article (FAI), in-process, final, and source inspection at the supplier, plus dock audits on already-shipped lots.
Sampling logic dominates this domain. A plan is defined by n (sample size) and c (acceptance number); its OC curve shows the risk of accepting a bad lot (consumer's risk, β) and rejecting a good lot (producer's risk, α). Under ANSI/ASQ Z1.4-style plans, switching moves from normal to tightened after 2 of 5 consecutive lots are rejected, and to reduced after 10 consecutive accepted lots with production steady. The trap: 100% inspection is assumed perfect, yet manual inspection misses a measurable percentage of defects - inspection screens, it does not create quality.
Nonconforming output must be identified, segregated, and dispositioned, typically through an MRB. The classic dispositions are use-as-is (needs engineering justification), rework (restore to print), repair (restore to function, often with a concession), regrade, and scrap or return to supplier. Traceability - lot, date code, and records - is what makes containment or recall possible.
Worked example. A 1,000-piece lot inspected at AQL 1.0% uses n = 80, c = 2. Accepting the lot means two or fewer defectives were found in the sample; it does not certify the lot is 1.0% defective - it bounds the risk at the stated confidence.