Operational Research

IN his admirable article on operational research in Nature of March 13, Sir Charles Goodeve makes the statement : "The application of the statistical tool is still on the increase, but there is no need or likelihood of the tool itself being developed further". Quite apart from the dangers inherent in making general prophecies about the likely advance (or lack of advance) in any field of science, this statement implies a degree of perfection in existing methods of dealing with statistical problems arising in operational research which would not, I think, be claimed by practising statisticians. In particular, the analysis of data (such as survey data) which can be classified in a number of different ways, and multi-variate analysis, both require further development.