|
Abstract:
|
Statistically designed experiments are powerful tools for process improvement and optimization, yet are underutilized because new students don't feel comfortable with the methodology. Typical introductory courses present columns of positive and negative signs to show the design levels, then jump directly to main effects plots and interaction plots to summarize results. The concepts of how the design is created and the physical meaning of effects and interaction are skipped. A unique manipulative for two level factorials allows students to visualize the design space as they plan the design, insert response data into the manipulative as the experiment is run, visualize the nature and magnitude of the effects in the model, and finally calculate values for the effects. This approach clarifies the concepts of DOE, making it more likely that these techniques will be utilized.
|
