Tu-11: Should Introductory Statistics Classes Include Multifactor Statistical Design of Experiments (DoE)?


By David Fluharty (Ivy Tech Community College)


Abstract

This presentation begins with a short PowerPoint demonstrating Multifactor Statistical Design of Experiments (DoE), followed by a discussion of whether DoE should be included—as a demonstration or an activity—in Introductory Statistics Classes. DoE employs a statistically determined set of interrelated experiments to simultaneously explore the impact of multiple factors on a phenomenon of interest. The demonstration presents a simplified version of George Box’s helicopter experiment. It uses eight ‘recipe cards,’ each of which corresponds to a strictly determined paper helicopter design. By rearranging the cards, the students can calculate the main effect of 4 factors on ‘flight time.’ Students then graph the results, decide which factors matter, and develop a predictive equation. This introduction to the power of DoE accesses the cognitive, psychomotor, and affective domains of learning while providing Introductory Statistics students with insights into the process of scientific discovery, multivariate thinking, and the omnipresence of variation.


Recording

Tu-11 - Should Introductory Statistics Classes Include Multifactor Statistical Design of Experiments (DoE).pdf