All manufacturing and development companies perform testing and/or inspections that involve concluding whether or not a product or lot is acceptable vs. design or QC specifications. Such test/inspections may occur during design verification/validation or during incoming or final QC.
The most informative method for analyzing the data that results from such activities is the calculation of the product’s or lot’s “reliability” at a chosen “confidence” level (where “reliability” means “in-specification”). Such a method produces information that is more valuable than simply that the given product or lot “passed” (as is the case when “AQL Attribute Sampling Plans” are used) or a % in-specification statement without any corresponding confidence statement (as is the case with AQL Variables Sampling Plans and with Process Capability calculations).
The output of a “Confidence/Reliability” calculation is a definitive statement that the given product or lot has a specific % in-specification, which conclusion we can state with a specific level of confidence (e.g., 95% confidence of 99% reliability, or 90% confident of 93% reliability”).
Areas to be Covered:
The seminar begins with a discussion of relevant regulatory requirements, as motivation for calculating “confidence/reliability”. Then, some vocabulary and basic concepts are discussed.
Next, detailed descriptions are given for how to calculate confidence/reliability for data that is either pass/fail (i.e., “attribute” data), normally-distributed measurement data, non-normally distributed measurement data that can be transformed into normality, or non-normally distributed measurement data that cannot be transformed into normality. Spreadsheets are shown as examples of how to implement the methods described in the seminar. A final discussion is provided on how to introduce the methods into a company.
Who will Benefit:
Vocabulary and Concepts
Normal Probability Plotting
Non-Normal Data that can be normalized
Reliability Plotting (for data that cannot be normalized)