Mathews Malnar and Bailey, Inc. Quality engineering, applied statistical consulting, and training services for R&D, product, process, and manufacturing engineering organizations.

Design of Experiments

Course Description: This course begins with a review of graphical presentation methods, measures of location and variation, and the hypothesis tests and confidence intervals necessary to analyze and interpret designed experiments. Students will learn to design, analyze, and interpret experiments involving one or more variables using Analysis of Variance (ANOVA) and regression techniques. Specific experiment designs to be covered are: the completely randomized design, the randomized block design, factorial designs, designs for fitting simple linear and quadratic models, two-level factorial designs, fractional factorial designs, central composite designs, Box-Behnken designs, and Plackett-Burman designs. Hybrid designs which mix qualitative and quantitative variables and designs for analyzing binary response data will also be introduced. Students will use Minitab (or a comparable program) during class and DOE software skills will be reinforced with extensive classroom exercises and homework assignments. Students will analyze data sets from the textbook, data from simple experiments run during class and for homework, and data from simulations of real engineering problems. Students will be required to report their results in both oral and written form.

Prerequisites: Students should have completed the Statistical Methods for Quality Engineers course or be able to demonstrate proficiency in basic statistical methods.

Contact Hours: 32 to 48 hours. Students have better command and retention of the material if the course is delivered over an extended period of time, such as in one four-hour session each week for nine weeks.

Homework: Seven assignments requiring about 2-6 hours each (3 hours nominal) will be given.

Textbook: Mathews, Design of Experiments with MINITAB, ASQ Quality Press, 2004.

Upon completion of this course students should be able to:

• Use the normal, Student’s t, and F distributions to interpret ANOVA and regression analyses.
• Use simple residuals diagnostic graphs to evaluate the ANOVA and regression assumptions of normality, equality of variances, independence, and lack of fit.
• Analyze experiments with qualitative variables using ANOVA.
• Analyze experiments with quantitative variables using regression analysis.
• Design, build, and use screening and response surface models.
• Build and interpret statistical models including simple linear, interaction, and quadratic terms.
• Simplify models using Occam’s Razor.
• Determine experimental sample sizes to meet specified detection limits and risk levels.
• Run an experiment using the 11 step method from the planning stage through final reporting.
• Design, analyze, and interpret experiments involving qualitative and quantitative variables:
• Completely randomized and randomized block designs
• General factorial designs
• Experiments for variables with random levels, e.g. gage error studies and process capability studies
• Experiments for building simple linear and nonlinear models in one variable
• 2^k (two-level factorial) and 2^(k-p)  (fractional factorial) designs
• 3^k, central composite, and Box-Behnken designs for quadratic modeling

Return to MM&B Courses

Return to MM&B Home Page