In this paper, we present an interactive teaching approach to introduce the concept of optimal design of experiments to students. Our approach is based on the use of spreadsheets. One advantage of this approach is that no complex mathematical theory is needed nor that any design construction algorithm has to be discussed at the introductory stage. Another benefit is that the students build all necessary matrices for concrete examples starting from a sensible initial design. By modifying the initial design by trial and error, they can try to improve the properties of the parameter estimators interactively. For problems in which finding the optimal design is not evident, they can use optimization software which is readily available in the spreadsheet software.
The CAUSE Research Group is supported in part by a member initiative grant from the American Statistical Association’s Section on Statistics and Data Science Education