Principles of screening experiments

In explorative research, the purpose of experiments is done to furnish new infonnation. An experimental design is a detailed plan of a series of experiments suitably arranged so that the observed result will give the desired information. It is therefore necessary that the experimenter, prior to any experiment, clearly states what the desired information should be. For a screening experiment, this implies that the experimenter should first specify all the potentially important variables, and then decide upon a suitable set of experimental runs so that the influence of each variable can be established. Then, the significant variables can be identified as those which have an influence above the noise level of the experimental error. 

In principle, this is a straightforward strategy. In practice, however, this is not trivial. Before any screening experiment can be laid out, an imaginative and creative process is required. During this process, a number of questions can hopefully be answered without any experiments. After this process, the problem will be more clearly stated than before. A clearly defined question is a necessary requirement for good experimental design. 

of course, impossible to give an algorithm for creativity, but some advice on how to approach the screening problem is appropriate  

Adopt an operational attitude towards the problem, i.e. to acknowledge that we can learn things by doing. After an experiment we know more than before, and this will help us to ask more detailed questions for the next experiment, see Fig. 2.1. 

---