Sampling Protocol
Statistical acceptance sampling measuring product batch conformity against defined defect thresholds through predetermined lot sizes and inspection levels defines this standard procedure. Garment buyers and mill quality controllers apply aql inspection to determine whether a production shipment meets contractual standards before release from the manufacturing facility. Visual, dimensional and packaging anomalies are categorised as critical, major or minor based on their effect on marketability or consumer safety.
Mathematical tables derived from ISO 2859-1 or ANSI/ASQ Z1.4 specify exact sample draw sizes alongside maximum allowable defect counts. The process assesses probability rather than total uniformity. Non-sampled units inside an accepted lot remain uninspected.
Defect Classification
Production defects split into non-negotiable legal tiers and cosmetic variations. Broken needle fragments, exposed metal points or mold growth rank as critical non-conformities, demanding a zero acceptance threshold across sampled lots. Major defects comprise structural flaws that impair functional performance or consumer acceptance, including skipped stitching, open seams, fabric holes and out-of-tolerance garment measurements.
Minor defects represent superficial blemishes such as untrimmed thread ends, slight shade variations within a roll match or minor print misalignments. Typical commercial agreements mandate an acceptable quality limit of 0 for critical, 2.5 for major and 4.0 for minor defects under General Inspection Level II.
Inspection Procedure
Random garment sampling takes place once a production order reaches at least eighty percent completion and packaging. Inspectors select cartons using pseudo-random number sequences across distinct production runs, cutting lines and dye lots. Drawn garments undergo inspection on light tables against approved purchase order specifications, sealed golden samples and tech packs.
If the tally of recorded defects equals or remains below the designated acceptance number, the shipment passes commercial release. When defect tallies equal or exceed the rejection number, the entire shipment faces full mill sorting, reprocessing or outright purchase order cancellation.
Statistical Boundary
Sampling methodology delivers mathematical confidence limits rather than complete quality assurance. A passed inspection lot can still contain undetected defective units proportional to the selected acceptance tier. High production variance within uninspected sub-lots undermines statistical assumptions regarding batch homogeneity.
Small production batches under five hundred units yield sample sizes with diminished statistical discrimination, necessitating tightened screening protocols.