On the Characteristics of Italian Statistics

Abstract

Now, nearly forty years later, after the favourable response of the then President of the Society, and the Editors of the Journal, I return once more to this subject, especially as British publications have more than once called attention to Italy's particular position regarding and to her contributions to them. An appendix to the bibliography in several editions of Yule and Kendall (19371939-1944) gives a bibliography of works on statistical methodology, pointing out that the methods developed in Italy differ considerably from those of British and American authors. Kendall and Buckland (1960) contains a number of terms used by statisticians which are distinguished by an asterisk. This Journal has published reviews by Professor Kendall of three books written by myself and my collaborators (Gini, 1939a, 1958, 1959) in each of which he draws attention to the typically characteristics of the treatment. He remarks that authors-unlike British and American-prefer applications to demographic phenomena and concentrate on the characteristics of complete collectives, only resorting to sampling as a secondary method. Kendall attributes these divergences essentially to the different traditions of the two countries. In my opinion his diagnosis is substantially correct, but the differences he notices are not the only ones, nor do I feel that he is always right about them. In any case I think they deserve to be discussed. This is the essential purpose of this article. I also wish to point out to British readers some of the advantages of the concepts and applications of statistical methods which might encourage their adoption outside Italy. As a matter of fact, the attention given to methods by British statisticians has sometimes been accompanied by a manifest tendency to discourage their adoption. When my lecture at London University was printed in this Journal in 1926, a condition was laid down that all quotations from foreign articles in which methods were applied should be excluded, leaving only quotations referring to applications made by Italians. There was serious resistance to accepting the terms used by statisticians in Kendall and Buckland (1960) and eventually not all of them were included. The reason given was that the amount of space required would be disproportionate to that accorded to authors from other countries (excepting those from Britain and America). t Both here and later I use the term Italian school of statistics (and, in contrast, British school or British and American school). This does not imply that these are schools officially set up or that all statisticians (or all British or British and American statisticians) possess identical characteristics or share the same views on all matters. I only mean that they generally follow certain scientific trends whose minor disagreements may be recognizable at home but whose typical nature is recognized abroad. This article reflects my personal views which may well differ from those of other statisticians. + Editor's note: As this article goes to press we record with regret that Professor Corrado Gini died on March 13th, 1965.