Fundamentals of Experimental Design

Who Should Attend

Bench chemists, assayists, bioassayists, technicians, scientists, engineers, laboratory managers, R&D managers, manufacturing and production managers, research supervisors, project managers, vice presidents, and others who need to learn, understand, and apply proven experimental design techniques for increasing the effectiveness, efficiency, and productivity of modern research and development.

The course is aimed at both beginning and experienced workers. The course assumes no previous knowledge of statistics or experimental design.

Key Topics You Will Learn About

  • How to extract information from experimental data
  • Fundamental concepts of statistical experimental design
  • Strengths and limitations of many popular experimental designs
  • How to choose appropriate experimental designs for particular applications

How You Will Benefit From This Course

  • You will master the twelve steps to successful experimentation
  • You will save resources by eliminating unnecessary experiments
  • You will improve research effectiveness, efficiency, and productivity
  • You will see how to profit from the statistical concepts of interaction and blocking
  • You will understand statistical terminology
  • You will improve your skills in communicating research strategies to co-workers
  • You will communicate more easily with statisticians
  • You will understand why traditional single-factor-at-a-time experiments are ineffective and inefficient
  • You will appreciate the "Rosetta Stone" of experimental design — the sums of squares and degrees of freedom tree
  • You will receive a brief introduction to Taguchi's philosophy and methods
  • You will develop a firm foundation for understanding advanced design techniques
  • You will be able to match appropriate experimental designs to real-world problems
  • You will learn about commercial software packages for data treatment

Day 1 Morning

  • Become familiar with symbols and terminology so you can talk with statisticians and read the statistical literature
  • Learn why statisticians use linear models as the basis of many applied experimental designs
  • Find out how to recognize non-linear models and understand when they might be useful
  • See how the mathematics of linear regression reveals unexpected information about your design
  • Avoid "software intimidation" by discovering how easily matrix least squares is achieved
  • Learn how to estimate useful responses from a fitted model

Day 1 Afternoon

  • Understand why residuals should be small — and find out the two ways to make them small
  • Know why replication is essential for good decision making
  • Discover that lack of fit is not bad — it represents an opportunity for your model to become even better
  • Explore the additivity of the seven sums of squares
  • Remove the mystery from the seven degrees of freedom with only three symbols — n, p, and f
  • Discover the Rosetta Stone of experimental design — the sums of squares and degrees of freedom tree

Day 2 Morning

  • Master the concept of factor interaction — avoid drawing wrong conclusions about what your data really say
  • Learn three reasons why statisticians use coding — and discover how to "uncode" their experimental designs
  • Find out why orthogonal designs are desirable — and why orthogonality is seldom realized in the real world
  • Be introduced to the grandparent of almost all other experimental designs: the full factorial designs
  • See how the old Yates' algorithm works, and why modern regression analysis is a more general alternative
  • Discover why two programs that analyze the same data can give answers that differ by a factor of two

Day 2 Afternoon

  • Understand why statisticians break the full factorials apart to give fractional factorial designs
  • Learn the concept of blocking and find out what you get with it — and what you don't get
  • Understand the need for screening experiments to find factors with the greatest power over the universe
  • See why saturated fractional factorial designs are perfect for screening — but often require dummy factors
  • Understand how the Plackett-Burman designs complement the saturated fractional factorial designs
  • Discover (and quickly understand) the obscure Hadamard designs
  • Realize that you now know what Taguchi designs are — and discover Taguchi's real contribution

Day 3 Morning

  • Learn how Box and Wilson constructed the central composite response surface designs
  • See how Box and Behnken constructed the Box-Behnken response surface designs
  • Work in teams on two experimental design problems that are common to most research areas
  • Discover the one key concept to understanding mixture (formulation) designs
  • Find out how statisticians use the analysis of variance (ANOVA) to extract information from data
  • Discover the amazingly simple concept behind the four "correlation coefficients" — R2, R, r2, and r
  • Understand what is behind the Fisher variance-ratio tests for regression and lack of fit, and what they mean

Day 3 Afternoon

  • Discover the importance of confidence intervals — and see how easy it is to construct them
  • Learn a simple trick for using the confidence interval calculation to construct confidence bands
  • Understand why the covariance between parameter estimates is not the same as interaction between factors
  • Appreciate the fundamental importance of the precision of the measuring process
  • Gain an intuitive understanding of how to place experiments to get better information
  • Discuss your applications and be guided toward appropriate experimental designs
  • Conclusion of course

Return Arrow