Sampling Plans
Meaning ~ Statistical testing schedules that determine sample sizes and reject limits for production batches, protecting buyers from receiving defective goods.

Dock
Industrial quality control uses mathematical models to determine how many units to draw from a production lot and how many defects can be tolerated before rejecting the entire shipment.
On a receiving platform, an inspector pulls eighty test specimens from a ten-thousand-unit shipment of polymer gaskets. Under the inspection protocol, if nonconformities remain at five or fewer, the shipment proceeds to the assembly line; if defects exceed five, the entire lot goes back to the vendor. Standards like ISO 2859-1 establish these boundaries to shield buyers from bad inventory while granting manufacturers a predictable margin for process variation.
How automated in-line computer vision inspection will eventually alter these lot-by-lot statistics in international dispute-resolution forums remains an open question.

Arithmetic
Calculating sample size begins with the total quantity of the submitted lot.

How Is the Sample Size Code Letter Determined?
Applying a standardized attribute plan starts by locating the lot size on a code letter table. These tables group lot quantities into bands, spanning small production runs up to hundreds of thousands of units. The chosen inspection level then determines the code letter.
General Inspection Level II serves as the default baseline, balancing inspection costs against statistical confidence. Level I reduces sample size for lower-risk items, Level III expands it to curb buyer risk, and special levels S-1 through S-4 cover destructive or high-cost testing where only a few specimens can be destroyed.
| Lot Size | Level I | Level II | Level III |
|---|---|---|---|
| 151 to 280 | E | G | H |
| 281 to 500 | F | H | J |
| 501 to 1200 | G | J | K |
| 1201 to 3200 | H | K | L |
| 3201 to 10000 | J | L | M |
A batch of ten thousand machined parts dictates an eighty-unit sample under general inspection level II to maintain standard risk margins.

The Probability Curves of Consumer Risk
Every code letter and AQL combination defines an operating characteristic curve showing the mathematical probability of accepting a lot across various defect rates. Buyers who set loose AQL targets risk accepting shipments with elevated defect levels. Producer’s risk ~ the chance that an acceptable lot gets rejected ~ typically stays around five percent.
Consumer’s risk represents the probability of accepting a lot with a high defect rate, which standard tables calculate at the ten percent probability point. Weak thresholds set on paper inevitably allow higher defect rates off the production line.

Bench
When evaluating drawn specimens, inspectors sort defects into three distinct tiers of severity.

Triage of Mechanical and Aesthetic Defect Categories
Defects carry different commercial consequences, so inspection protocols categorize failures into critical, major, and minor tiers. A critical defect compromises safety or violates regulatory standards. Major defects impede intended function or shorten product life without posing safety risks.
Minor defects involve surface or cosmetic flaws that leave functionality intact but can reduce retail value.
| Code Letter | Sample Size | AQL 1.0 (Ac/Re) | AQL 2.5 (Ac/Re) |
|---|---|---|---|
| G | 32 | 0 / 1 | 1 / 2 |
| H | 50 | 1 / 2 | 2 / 3 |
| J | 80 | 1 / 2 | 3 / 4 |
| K | 125 | 2 / 3 | 5 / 6 |
| L | 200 | 3 / 4 | 7 / 8 |
| Note: The abbreviation Ac designates the acceptance number, while Re indicates the rejection number; the lot is accepted if the defect count does not exceed Ac. | |||

Measurement Scales for Variable Testing Plans
Attribute plans judge units on a pass-fail basis, but variable plans record specific physical measurements like thickness, hardness, or tensile strength. Standard ANSI/ASQ Z1.9 covers variable procedures, which achieve comparable statistical confidence from far smaller sample sizes than attribute tables require. These plans rely on the assumption that the measured trait follows a normal distribution.
Provided process variance remains stable, variable inspection saves time and preserves material, avoiding costly scrap when testing is destructive.
- sampling plans allocate sample sizes based on the critical nature of the tested component.
- sampling plans establish the limits of acceptable lot risk before physical testing starts.
- sampling plans specify the required action when nonconforming units reach the rejection threshold.
Accepting defective lots outside statistical limits leads directly to assembly line disruptions and severe warranty expenses.

Crate
A statistically valid sample requires every unit in the shipment to have an equal chance of being selected.

Lot Integrity and Randomization Protocols
Selection bias frequently creeps in when inspectors pull samples only from the top layer of cartons. To remain statistically valid, a lot must be homogeneous ~ produced under the same conditions, using consistent machinery and raw materials. Mixing batches from separate production runs into one lot invalidates the statistical model, which is why quality engineers require separate sampling for each distinct run so defective output is not hidden by higher-quality units.
ISO 2859-1 governs the transition from normal to tightened inspection when consecutive batches fail the quality threshold.
- Inspectors count the pallets and cartons at the loading dock to verify total batch size against the shipping manifest.
- Sampling protocols require pulling individual units from various depths within randomly selected cartons.
- Technicians transfer the sampled items to the bench, maintaining stable environmental conditions throughout measurement.
- Quality managers enter the final defect tally into the register as soon as testing ends.
Discrepancies often arise when transit vibrations alter the mechanical alignment of units after leaving the packing line.

Ledger
Buyers safeguard capital by writing explicit quantitative thresholds into master supply agreements.

Commercial Allocation of Sorting and Scrap Liabilities
Contract terms specify who absorbs the cost of sorting a failed lot. If a shipment of electronic components fails dockside inspection, the buyer can reject the entire lot or execute a 100 percent sort at the supplier’s expense. Rejection ties up working capital and risks missing delivery windows, so on-site sorting often serves as a practical compromise.
Well-drafted agreements ensure the supplier covers labor and replacement parts whenever defect counts exceed agreed limits.
Destructive testing forces buyers to accept larger statistical risks to preserve salable inventory.
Consider a forty-thousand-unit shipment of industrial gaskets priced at fifty cents each, FOB Shanghai. Under General Level II sampling, a lot this size requires Code Letter N, calling for a five-hundred-unit sample. With an AQL of 1.5 percent for major defects, the acceptance threshold is fourteen and the rejection number is fifteen.
Finding sixteen defective gaskets triggers full lot rejection. Returning forty thousand gaskets overseas carries heavy shipping costs ~ the supplier would owe roughly three thousand dollars in return freight plus two thousand dollars in port demurrage. Instead of shipping the lot back, the supplier might fund an on-site third-party sort.
At forty dollars an hour for eighty hours, sorting labor comes to three thousand two hundred dollars. Here, the financial consequences of setting sampling parameters fall directly on the supplier, who pays for labor and replacement units. Statistical terms in supply contracts directly shape commercial exposure.
- sampling plans limit buyer liability by setting explicit statistical grounds for lot-wide rejection.
- sampling plans define how risk shifts between supplier and buyer under EXW or FOB terms.
- sampling plans provide the documentation needed to issue chargebacks for sorting or rework.
Requiring a full 100 percent supplier-funded sort whenever a lot fails Level II inspection significantly shifts leverage toward the buyer.

Swing
Standard quality procedures adjust inspection frequency based on a manufacturer’s recent performance history.

What Happens When Consecutive Batches Fail?
Standard sampling protocols adapt over time. If two out of five consecutive lots fail initial inspection, rules require a switch from normal to tightened inspection. Tightened mode lowers allowable defect limits or increases sample size, raising the rejection risk for marginal lots and forcing suppliers to address production problems promptly.
If quality fails to improve and five consecutive lots are rejected under tightened inspection, the standard calls for stopping inspection altogether, effectively halting shipment approvals.

The Mechanics of the Switching Score
Returning from tightened to normal inspection requires five consecutive accepted lots on original inspection. Conversely, moving to reduced inspection becomes an option after ten consecutive passes under normal inspection, provided overall quality metrics remain consistently high. Reduced inspection cuts sample size by about sixty percent, lowering costs for both parties while keeping the statistical integrity of the sampling plan intact.





