Almost all design and/or manufacturing companies evaluate product and processes in order to either manage risks, validate processes, establish product/process specifications, QC to such specifications, and/or monitor compliance to such specifications. The various statistical methods used to support such activities can be intimidating. If used incorrectly or inappropriately, statistical methods can result in new products being launched that should have been kept in R&D; or, conversely, they can result in deciding a new product is not ready for launch, but whose product reliability or process capability is actually very good. In QC, mistakenly chosen sample sizes and inappropriate statistical methods may result in product being rejected that should have passed, and vice-versa.
This seminar provides a practical approach to understanding how to interpret and use a standard tool-box of statistical methods, including confidence intervals, t-tests, Normal K-tables, Normality tests, Confidence/Reliability calculations, Tolerance Limits, Reliability Plotting, AQL sampling plans, measurement equipment analysis (including Gage R&R), and Statistical Process Control (including Cpk/Ppk calculations). Without a clear understanding and correct implementation of such methods, a company risks not only significantly increasing its complaint rates, scrap rates, and time-to-market, but also risks significantly reducing its product and service quality, its customer satisfaction levels, and its profit margins.
The seminar includes a strong focus on risk-management, regulatory compliance, and sample-size determination/justification.
At the start of the course, a large suite of statistical application spreadsheets is given to all students, in order to assist them in understanding and applying the course concepts and methods after returning home. Additionally, the instructor is available indefinitely by email, free of charge, to answer short questions related to the course topics, and/or to perform statistical analysis on data-sets that are provided to him in an Excel spreadsheet (in such cases, the data should be presented as a set of numbers and their QC/Design specifications, without any explanation regarding the identity of the company’s product that generated them; therefore signing an NDA would not be necessary).
- How to apply “confidence” and “risk management” to virtually all statistical techniques
- Know when to use “exact” methods rather than “approximation” methods
- Know when to use commercial software (e.g., Minitab, StatGraphics) and when to use Excel
- How to explain statistics to management
- How to interpret regulatory requirements related to statistics
- How to determine the smallest sample size needed to achieve a desired outcome.
Areas Covered :
- Basic regulatory requirements related to statistics (for medical devices and pharmaceuticals)
- Basic statistical concepts and vocabulary
- Normality, Normality Tests, and Normality Transformations
- Statistical Process Control
- Process and Product Capability assessments including: Confidence/Reliability Calculations, Tolerance Limits, Cpk/Ppk, and Reliability Plotting
- Statistical Significance tests (including testing for “Superiority”, “Non-inferiority”, and “TOST”)
- Statistical Power
- Metrology (statistical analysis of measurement uncertainty, including Gage R&R and “guard-banding”)
- QC Sampling Plans (AQL vs. LQL vs AOQL vs. other alternatives)
- Statistical justification for Process Validation sample sizes and the use of only 3 Lots
- Examples of “statistically valid rationales” regarding sample sizes.
Who will Benefit:
- R&D Manager
- QA/QC Manager
- Manufacturing Manager
- R&D Engineer
- Manufacturing Engineer
- Process Engineer
- Validation Engineer
- QC/QC Technician