Defect Classification
Meaning ~ System for grouping product errors into severity levels that define if a batch is accepted, reworked, or scrapped based on risk and contract limits.
Sampling Plan Bias
Meaning
Systematic errors in the selection of units for testing lead to results that do not accurately represent the entire production lot. Sampling plan bias occurs when the method of picking samples favors certain parts of the batch, such as only taking units from the top layer of a pallet. This distortion makes the quality of the shipment appear better or worse than it truly is and the bias undermines the validity of the audit.
Selection Flaw
Choosing units based on convenience rather than true randomness introduces a consistent lean toward a specific outcome. Sampling plan bias often happens in factories where the supplier has pre-arranged the units they want the inspector to see. Random number generators and proper access to all areas of the warehouse are the primary tools used to combat this issue.
Statistical Impact
Distorted data ruins the mathematical probability that the inspection tables rely on for accuracy. When sampling plan bias is present, the operating characteristic curve no longer provides a true picture of the risk levels for the buyer. Correcting this requires a change in the physical process of how samples are pulled from the master lot.
Inaccurate Result
Decisions based on skewed data lead to the acceptance of defective goods or the unnecessary rejection of sound inventory.