Collecting data for research and evaluation processes are often performed via experimental design.
The notion and the methodology of modern experimental design were developed by Sir Ronald A. Fisher in 1935 in the “Design of experiments.” But the concept of experimental method (classical experimental design) was introduced much earlier by Galileo Galilei at the end of the XVIth century and later presented in the first half of the XVIIth century by René Descartes, in his book «Discours de la Méthode pour been conduire sa Maison et Chercher la vérité dans les sciences».
In classical experimental design, investigators select few factors that they think could influence the response and then choose the type of design to be used while evaluating the sample. The classical experimentation considers a specific variable while controlling and if possible eliminating the influence of other variables.
The experimental design is closely interconnected with the idea of causality in the sense that to establish causality (to show to what degree it is probable) one must use experimental design principles. Those principles are based on the idea of controlled observation. Such conditions enable to isolate causality and find out the relation of dependent and independent variables (how a causal factor affects the caused one).
The classical experimental design implies certain conditions where there are two groups (randomly assigned) taking part in the experiment. The experimental group that is exposed to/stimulated, treated, serviced by the causal (independent) factor. In the control group, this stimulus is not provided.
Testing whether the people/subjects in an experimental group are affected differently than the people/subjects in the control group is usually the core of the experiment.
When people or subjects are being studied, they are randomly assigned to two different groups (the experimental one, and the control group). If the experiment sample has matched pairs of similar individuals/subjects, they are randomly divided and assigned to different groups.
Experimental designs are the best way to meet the three basic criteria for establishing causality: the cause should precede the effect, the two factors or variables should be empirically correlated, and the empirical correlation between the two variables should not be a spurious correlation (it cannot be accounted for by other variables that cause the two variables to vary together). The empirical correlation means that if one variable changes, the other does too (positively or negatively).
The classical experimental design, though newer theories and methods later replaced it, has always been the basis for the determination of the causality between different variables. Its role in the development of the modern designs of the experiment can hardly be underestimated. This concept, in the core, is the same as it used to be for centuries, it has only undergone positive evolutional changes that enable the empirical data and the degree of probability to be measured more precisely.