“… the actual and physical conduct of an experiment must govern the statistical procedure of its interpretation.” R. A. Fisher