The working group on Design of Experiments (DOE) is interested in various statistical aspects of the planning of experimental research. While the scientific targets of different experiments differ and depend on the application area (e.g. engineering, biology, medicine, chemistry, agriculture, etc.), one statistical methodology is required to achieve them. This unites the interests and effort of group members working in the academia, as well as in specified research areas or industry.
For example, the design of an experiment studying the transformation of a chemical compound to another may require an experimental strategy similar to that needed to study drug availability in a living organism. The reason for that is that in both cases the same type of statistical model has to be estimated using experimental data, albeit with rather different interpretation of the model parameters.
A typical challenge that we face is how much and what data to choose to collect in order to estimate a statistical model that would allow addressing the research questions of the experimental study. The design of experiments may be required to estimate models including linear, nonlinear, linear mixed, nonlinear mixed and generalized linear or nonlinear mixed effects models. Inevitably, many members of the DOE working group are also interested in data analysis and statistical modeling.
Experimental research is iterative by nature, and the specific challenges at each stage are different and often shaped up by the specific areas of applications. In all cases, it has become increasingly common to seek solutions using analytical and computer intensive methods. The Design of Experiments has become a fascinating area of research in Statistics, as well as a powerful way for statisticians to influence developments in other sciences.