SED navigation bar go to SED home page go to Workshops page go to NIST home page SED Home Page SED Contacts SED Projects SED Products and Publications Search SED Pages

Workshop on Improving Product and Process Quality Using Experiment Design

The importance of designed experimentation a systematic and rigorous plan for conducting an experimental investigation is gaining increased appreciation in the scientific and engineering community. The goal of a "good" design is to yield unambiguous scientific and engineering conclusions with a minimal amount of experimental effort. All experimental programs have time and cost constraints, but given these constraints, not all experiment designs are equally powerful. Underlying geometric and statistical properties of certain designs make them far superior to those commonly used in practice. A three to five-fold reduction in experimental effort (with no loss of information content) is an often-realized benefit of using such optimal designs. The widespread development and use of run-efficient designs is one of the main reasons for Japanese breakthroughs in quality engineering, and has been used as "standard operating procedure" in the Japanese engineering and scientific community for the last 20 years.

This five-day workshop covers the fundamental principles and techniques of efficient experiment design. It is based on proven design/analysis methodology which has been used and applied by NIST scientists and engineers. The methodology is a mix of "classical" methods, Taguchi methods, and as- of-yet unpublished NIST methods which have flowed out of solutions to NIST scientific problems. The workshop covers specific designs that have proven best for three very important problems:

  • ferreting out the most important parameters from a large number of potential parameters
  • converging to an optimal operating condition starting from a "best guess"
  • identifying settings of controllable parameters that reduce product and process variation (Taguchi's quality engineering approach)
The workshop will demonstrate the problem-solving role of designed experiments via applications to a wide variety of areas including maximizing yield of chemical processes, optimizing automated tooling devices, and improving semiconductor performance. By the end of the workshop, every participant will have drawn up an optimal experimental plan for his/her own project based on the principles and techniques covered by the workshop.

This workshop is for scientists and engineers in research and development or manufacturing with a goal of improving quality and productivity. Although some familiarity with the concepts of elementary statistics would be useful, it is not required. The workshop will exclude mathematical developments, and emphasize an experimentalist based evolution of concepts and techniques with reinforcement via practical applications.

The workshop text is Statistics for Experimenters by Box, Hunter and Hunter (Wiley).

Check the SED Calendar to find out when the next presentation of the workshop is scheduled.

For information on workshop content, contact james.filliben@nist.gov or via phone at (301) 975-2855.

Date created: 6/5/2001
Last updated: 6/20/2001
Please email comments on this WWW page to sedwww@nist.gov.