James J. Filliben
William F. Guthrie
Statistical Engineering Division, CAML
The Statistical Engineering Division is pursuing a joint project with the Statistical Methods Group of SEMATECH on electronic publication of a handbook, in the style of NBS Handbook 91, on statistical methods for scientists and engineers. The handbook is intended to provide modern statistical and graphical techniques which are appropriate for the problems confronting the U.S. industry, particularly the semiconductor industry, and the NIST laboratories.
The handbook, in combination with menu-driven statistical software, will offer its readers a flexible, guided statistics tool kit for selection and operation of statistical software with minimum effort. A feature that sets this book apart from other statistical texts is its organization by problem statement rather than by statistical method. The navigation through the book is structured so that the reader can get to the appropriate material via outline, engineering question, flow- chart or statistical method. The first level outline is as follows:
1. Overview 2. Looking at Data (Exploratory Data Analysis) 3. Measurement Process Characterization 4. Process/Product Characterization 5. Building a Function-Based Process Model 6. Process Improvement 7. Process/Product Comparison 8. Process Monitoring 9. Assessing Product and System Reliability 10. Glossary of Terms and Symbols 11. Index of Engineering Questions 12. Index of Examples 13. Index of Statistical Techniques
The public domain software, Dataplot which was developed at NIST by Filliben, has been extended to a menu-driven system for this purpose.
Recently, the staff of the Statistical Engineering Division joined their counterparts at SEMATECH in presenting the background and vision for the book and demonstrating a prototype chapter and software to the SEMATECH Advisory Council. In the coming year, authors and subject matter experts for major sections will be identified; two prototype chapters will be completed; and a production editor will be identified.
Date created: 7/20/2001