James J. Filliben
Eric S. Lagergren
Statistical Engineering Division, CAML
In recent years there has been a growing awareness among engineers in industry of the virtues of statistical design and analysis tools in the on-going competitive quest for improving product and process quality. In response to this, the SED began offering the multi-day "Workshop on Improving Product and Process Quality Using Experiment Design" (originally called "Experiment Design for Scientists and Engineers"). This workshop has been offered since 1989, once or twice per year:
1989: Summer -Boulder, CO (2-day) (Filliben) NIST/NOAA
1989: Fall -Gaithersburg, MD (5-day) (Filliben) NIST
1990: Nov 26-30 -Gaithersburg, MD (5-day) (Filliben, Lagergren, Kacker) Industry
1991: Oct 28-Nov 1 -Gaithersburg, MD (5-day) (Filliben, Lagergren, Kacker) Industry
1992: Jul 13-17 -Boulder, CO (5-day) (Filliben, Lagergren, Kacker) Industry 1992: Nov 30-Dec 4 -Gaithersburg, MD (5-day) (Filliben, Lagergren, Kacker) NIST
1993: Aug 2-6 -Santa Clara, CA (5-day) (Filliben, Lagergren, Vecchia) Industry
1994: April 25-29 -Buena Vista, CA (5-day) (Filliben, Lagergren, Kacker) Industry 1994: July 4-8 -Bombay, India (5-day) (Filliben, Lagergren) Industry
1995: May/June -Gaithersburg, MD (8-day) (Filliben, Lagergren) NIST 1995: Oct 2-6 -Gaithersburg, MD (5-day) (Filliben, Lagergren) Industry
In 1995, the course was offered twice. The May/June version was 8 days and focused on NIST scientists and engineers-but with subsequent benefits to industry because of the many industrial products traceable via calibration or theory to NIST. The October version was 5 days in length and focused on industry. Attending industrial representatives included:
CERDEC Monsanto Concurrent Technologies Fluoroware, Inc. CarboMedics Eka Nobel, Inc. U.S. Naval Research Laboratory Ocean Spray Cranberries Norton Company Phoenix Medical Tech. Donaldson Co., Inc. Merck & Co. DuPont Co. McQuay International General Electric Company Ortho Diagnostic Systems Johns Hopkins University Wilcoxin Research Dow Chemical Co.
As is evident from the above list, the principles and techniques of experiment design and data analysis as presented in the course will touch a broad spectrum of industrial products and applications.
Content-wise, the workshop covers statistical designs
1) for studying a single parameter in the presence of many nuisance parameters;
2) for screening important parameters from a large set of parameters; and
3) for determining best settings of the important parameters for both mean and variability (Taguchi methods).
To reinforce underlying design and analysis concepts, students construct, run, and analyze a series of in-class, hands-on experiments. In addition, each student designs an experiment for a project they are currently working on at their own company. Student feedback has been positive.
Date created: 7/20/2001