Illustrating Split-Plot Designs with 3-D Models

• Activity Number: 349
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 11:15 AM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): James Alloway*+
• Companies: E.M.S.Q. Associates
• Keywords: Split-plot design ; DOE ; 3-D models ; Pedagogy

Abstract

While split-plot designs have been known for nearly 90 years, the interest in this design has recently spiked. Most statistical software used in quality improvement efforts have added or emphasized this design. Although software handles the more complex analysis, the experimental team must recognize when a split-plot design is beneficial and also be aware that the analysis must reflect any restrictions on randomization during data collection. Split-plot designs address a key objection to utilizing statistically designed experiments: randomization, which increases the time and effort to conduct experiments. If practitioners aren't comfortable with these designs, they will not use them. We must find ways to simplify the presentation of this class of designs for quality practitioners who lack extensive backgrounds in statistical methods. One approach is to utilize 3-D models to explore the structural differences between completely randomized factorial designs and factorial designs with restrictions on randomization. Once this concept is clear, it is easier to understand the software output and explain the results to others.