One Man's Meat: Part 3--It's all Good Grist that Comes to our Mill

One of the features of the last decade has been the relative failure of industrial statistics to make the impact expected in the early fifties. The cause cannot be the feud that divided statisticians, the echoes of which still reverberate. It is not that the journals are filled with heavy, esoteric methodology. They are, but so are those on computing and operational research. Yet where practical computing and operational research can be and are conducted on a much lower plane than the average journal article, can be and are practised successfully by the non-specialist, statisticians have always been prone to hedge themselves about with (for example) the dangers of departing from normality, the robustness or otherwise of tests and restrictions on the size of samples (they always want more). Despite the phenomenal sales of Moroney (1951), and the lay understanding of statistical technique that must spring from them, statisticians basically dislike laymen. Statisticians are a Mecca, to which one should turn at dawn and dusk. They are pedagogues who teach deviation, but castigate the poor experimenter who deviates from the narrow path, their path, to consider some curious and interesting phenomenon.