What is Design of Experiments
Design of Experiments is a common tool of Six Sigma and Design for Six Sigma. It can be applied in many disciplines, including product development, problem solving, clinical trials, academic research, and designing websites. Design of Experiments is about conducting strategically planned and well executed experiments. The goal of such experiments is to obtain knowledge and understanding of the phenomena under study, such as which are the most important influencing factors of a result; for example what drives customer satisfaction or the click rate on a button on a website. Knowing the most important factor, or even how much they influence the result, helps in making subsequent improvements to a product or process.
Why are the techniques from Design of Experiments useful?
In most experimental settings, the values that are obtained are subject to variation. For example, if you repeat an experiment with new raw material, results will be slightly different. Or if an experiment is spread across days and day effects are present. Design of Experiments gives techniques to carefully investigate such situations, and to come up with solutions to safeguard a trustworthy outcome and interpretation of the results. This way, the probability of making wrong decisions and their associated costs are minimized.
Another useful aspect of Design of Experiments is that it provides potentially huge shortcuts in the size of the study, saving time and money. As an example, suppose the goal is to investigate which of the, say, 10 potential influencing factors is most important. The straightforward method is to have 10 tests in a sequence, each changing one factor at a time. Design of Experiments gives the option of a (factorial) design, which would achieve the goal in only one experiment that would be only slightly larger than each of the 10 separate ones by itself!
What needs to be done in using Design of Experiments
It all starts with a goal. What is the problem, what knowledge is missing, what is already known? The consultants of CQM start with asking such questions to be able to generate a strategy for a sequence of studies. For instance, the first study might be a screening study, followed up by a more focused study. Each study needs to be planned carefully, often using schemes of the exact settings of each run or prototype for which a value should be collected. After the data is gathered, CQM can help with the statistical analysis. The analysis includes interpretation of the results and linking back to the original goal and problem.
CQM can help you get the best results!
CQM has over 35 years of consulting and training experience in using this branch of applied statistics in problems ranging from studies on consumer’s opinions to mass manufacturing settings. If you want to receive more information about CQM’s expertise please contact us.
Read more:
