Design of Experiments in Civil Engineering: Are We Still in the 1920’s?

In this paper, a very useful class of statistically designed experiments, the two-level factorial design will be introduced and the pitfalls of the commonly used one-factor-at-a-time (OFAT) method of experimentation will be discussed. Applications of statistically designed experiments or DOE in civil engineering in particular will be emphasized. Statistically designed experiments have been around for the last eighty years and have been widely used with great success in many industries for the past thirty years. However, they have yet to make any inroads in civil engineering except for certain areas of environmental engineering. In fact, many experienced researchers in civil engineering have never heard of it and have continued to use the inefficient trial-and-error or OFAT methods. These include many authors of papers in journals and presenters of papers at CSCE conferences over the years. It is hope that this paper can serve as a tutorial and will provide a stimulus for experimenters in civil engineering to abandon the trial-and-error and OFAT methods of experimentation and begin to learn and use the well established statistically based design of experiments.