Industrial Experimentation - edizione con copertina flessibile

INDUSTRIAL EXPERIMENTATION by K. A. BROWN LEE, M. A. FOREWORD: The present Monograph is based on an earlier MenfiO2fti4um produced by the Directorate of Ordnance Factories Explosives for the use, primarily, of those concerned with pilot plant and plant scale experiments on chemical manu facturing processes in the Royal Ordnance Factories Explosives. Much work of this type was being carried out and it had become evident that it was desirable for the results of such experiments to be subjected to critical tests of significance. A convenient account of the straightforward tests of significance, written from the point of view of the individual who has to apply them in practice without necessarily a full knowledge of their theoretical background, was not readily available, and an attempt was therefore made to prepare one. It was evident that to apply tests of significance conveniently and economically the experiments had to be planned in appropriate forms. It is considered that the methods outlined should be as much a standard tool of the industrial experi menter as a chemical balance is of the laboratory experimenter. In carrying out an industrial experiment the choice is not between using a statistical design with the application of the appropriate tests of significance or the ordinary methods the choice is between correct or incorrect methods. Even the simplest experiment requires an estimate of the significance of its results. The practice sometimes followed of consulting the statistician only after the experiment is completed and asking him quot what he can make of the results quot cannot be too strongly condemned. It is essential to have the experiment in a form suitable for analysisand in general this can only be attained by designing the experiment in consultation with the statistician, or with due regard to the statistical principles involved. The present Monograph, therefore, is intended to be a guide to both the planning and the interpretation of experiments on the industrial scale, and it is hoped that the methods described will become part of the everyday technique to those who carry out such experiments. R. C. BOWDEN, Director of Ordnance Factories Explosives Ministry of Supply. 1260tt l FEB 18 1948 3 COWTENTS amp lt r Page . .. .. ........ 3 ., . ., CONTEtffS V quot quot. quot. . quot quot ............ 4 quot quot amp gt PREFACE V - ................ 8 CHAPTER I INTRODUCTION a Experimental Error . . . . . . . . . . . . ., 9 b Classical and Industrial Experimentation . . . . . . . . 9 c Replication . . . . . . . ., . ., . . . . 10 d Experimental Design Randomised Blocks ...... 10 e The Latin Square .............. 12 f Balanced Incomplete Blocks ............ 13 g Lattice Squares ......... 14 h The Nature of quot Blocks quot . .......... 15 i Multiple Factor Experiments .......... 16 j The Three Factor Experiment .......... 17 k Higher Factorial Experiments .......... 19 CHAPTER II FUNDAMENTAL STATISTICAL CONCEPTIONS a Statistical Terminology . . . . . . . . . . 20 b Probability .............. 20 c Populations Tests of Significance .......... 20 d Significance Levels . . . . . . . . . . 21 e Computation . . . . . . . . . . . . 21 f Measures of Variability .......... 22 g The Calculation of Variance ......., 22 h The Definition of Variance . . . . . . . . 23 i Distributions . . . . . . . . t f 24 Cj Grouped Frequency Distributions .......... 25 k Log-normalDistributions .......... 26 CHAPTER III SIGNIFICANCE OF MEANS a Significance of a Single Mean ........ 28 b Confidence Limits for a Single Mean . . . ., quot 29 c Comparison of Two Means ....... 3Q i Small Samples . . . . . . quot 3 Q ii Large Samples ...... quot amp lt amp gt j d Conclusions . . .........