Volume 1

Collected papers of R. A. Fisher / edited by J.H. Bennett.

  • Ronald Fisher
Date:
1971-1974
Catalogue details

Licence: Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)

Credit: Collected papers of R. A. Fisher / edited by J.H. Bennett. Source: Wellcome Collection.

    47/616 (page 39)
    treating it in the traditionally abstract and, I think, fruitless way. Perhaps that’s why statistical science has been comparatively backward in many European countries. Perhaps we were lucky in England in having the whole mass of fallacious rubbish put out of sight until we had time to think about probability in concrete terms and in relation, above all, to the purposes for which we wanted the idea in the natural sciences. I am quite sure it is only personal contact with the business of the improvement of natural knowledge in the natural sciences that is capable to keep straight the thought of mathematically-minded people who have to grope their way through the complex entanglements of error, with which at present they are very much surrounded. I think it’s worse in this country [the U.S.A.] than in most, though I may be wrong. Certainly there is grave confusion of thought. We are quite in danger of sending highly trained and highly intelligent young men out into the world with tables of erroneous numbers under their arms, and with a dense fog in the place where their brains ought to be. In this century, of course, they will be working on guided missiles and advising the medical profession on the control of disease, and there is no limit to the extent to which they could impede every sort of national effort.’ Design and analysis of experiments When Fisher came to Rothamsted in 1919 he was brought into direct contact with workers who were concerned with the interpretation of agricul tural field experiments on crops, and with laboratory and greenhouse experiments. Ideas on the errors to which experimental results are subject were at that time confused. Knowledge of experimental error is required not only to give an idea of the general accuracy of the experimental results, but also to provide a basis for exact tests of significance. Although in the classical long-term experiments laid down by Lawes and Gilbert there was no replication other than that provided by the results of successive years, replica tion had for long been customary in many short-term field trials, and some agronomists had become aware that the differences between the replicates could be used to provide estimates of experimental error. The appropriate method of calculating such estimates when more than two treatments were under investigation was not known, however, and various alternative methods were in use, all of which were computationally laborious, and some grossly erroneous. Fisher early perceived that the analysis of variance provided a powerful technique for the separation of sources of variation in agricultural field trials. The first published application (1923) well illustrates the tentative nature of these early analyses. The results discussed were from an experi ment at Rothamsted on the effect of potash (sulphate, chloride and none) on 12 varieties of potatoes, with three replicates on each of two series, dunged and undunged respectively. The experiment was essentially of the split-plot type, the varietal plots being split for fertilizers, but by modern standards the design left much to be