MA 487: Design of Experiments

Prerequisites: MA 223 (Engineering Statistics) OR MA 382 (Introduction to Statistics)

Offering: In-person meetings occur on Mon. and Thurs. at 11am. Online sections meet asynchronously on Tues. and Fri.

Degree Requirements: This course counts toward

Example Syllabus: Winter 2020 Syllabus
Example Schedule: Winter 2020 Schedule

Have you had an introductory course in statistics? Are you interested in diving deeper into statistical methods for decision-making in engineering? Then this is the course for you. In MA487, we make use of data collected from controlled experiments to draw causal conclusions and identify important factors to engineering processes. Students who take MA487 have the opportunity to

  • Design and implement experiments where we collect and analyze real data.
  • Learn the R and R Markdown statistical languages.
  • Work in teams to problem-solve and make recommendations driven by statistical analysis, process costs, and practical needs.
  • Formulate a question of interest and implement a full experimental design by the end of the course.
  • The setting for this particular class is quite lively. Students should come prepared to work with classmates on a regular basis, discuss ideas during lecture, and actively participate during data collection sessions!

    Why is design of experiments important?

    Overseeing processes and improving upon currently utilized methods are two tasks many engineers face during their career. There is pressure to simultaneously increase profit margins and maintain high standards of product integrity. In order to improve upon current processes, engineers are asked to search for alternative methods or materials than those typically used. To increase profits in our rapidly changing world, innovation is key, and engineers are drivers to creating new products. Moreover, once a product is deemed fit for the consumer market, engineers oversee the processes that ensure quality.

    A natural intersection of engineering and statistics is design of experiments. The statistical tools that comprise the design of experiments subject area are integral to the informed decision-making process that engineers utilize in ensuring quality, improving product, and innovating processes. These tools allow engineers to collect samples of data that provide insights to process components that are causing improvement or deterioration in product. Analysis of data collected from a designed experiment may lead to discovery of practical process settings that are cost-efficient for the company. Furthermore, a well-designed experiment has possibility of leading the engineer to a groundbreaking discovery in their field.

    Why is statistics important?

    Individuals with strong statistical analysis skills are in high-demand. Statistician and Data Scientist positions top multiple desirable job lists online, including

    Essentially anyone working in STEM needs the ability to analyze data. This course gives you more advanced knowledge for performing your own data analyses.

    What kind of data analysis skills will I have after this class?

    You will be able to recommend appropriate designs/statistical models based upon a question of interest. By carefully selecting the model and controlling for sources which contribute to variation in processes, the collected data may be used to identify factors which effect a quantitative variable of interest as well as describe the relationships between multiple factors and the response. Topics we cover in this course include K-way ANOVA, factorial designs, block designs, fractional factorial designs, and response surface designs. Some graphs that you will be able to use/explain by the end of the course are shown below.

    alternate text

    Example 1: Interactions Plots




    alternate text

    Example 2: Latin Square Designs




    alternate text

    Example 3: Design Hasse Diagrams




    alternate text

    Example 4: Response Surface Design Results