STATISTICAL THEORY AND METHODS FOR VALIDATING RESULTS OF SAMPLING INSPECTION BY ATTRIBUTES
Abstract
Simple but robust statistical methods are described and developed for use in validating suppliers' inspection records of attribute sampling data. The methods are essentially two-sample significance tests for homogeneity of discrete variates treated as continuous, and the combination of their probabilities to test the hypothesis of over-all agreement of paired inspection results. The statistical theory and techniques presented in this paper form the basis for DoD Handbook H109, 'Statistical Procedures for Determining Validity of Suppliers' Attributes Inspection.' The procedures of the handbook constitute a system of product verification inspection wherein the consumer's sampling results establish the reliability of the supplier's acceptance sampling records and provide independent estimates of the quality of product submitted for acceptance. The characteristics and application of alternate tests are discussed. Utilization of the power function of the significance tests affords substantial reduction in the amount of product verification inspection.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 16, 1962
- Accession Number
- AD0612676
Entities
People
- Henry Ellner
Organizations
- United States Assistant Secretary of Defense