You can manage bookmarks using lists, please log in to your user account for this.
Media type:
E-Article
Title:
Revision critique de certains points de la methode representative
Contributor:
Galvani, Luigi
imprint:
W. P. Van Stockum & Fils, 1951
Published in:Revue de l'Institut International de Statistique / Review of the International Statistical Institute
Language:
French
ISSN:
0373-1138
Origination:
Footnote:
Description:
<p>Some delicate aspects of the representative method deserve even to-day to be examined in detail and, in the first place, it is pointed out that the numerous definitions given of this method differ considerably from each other in respect of their logical contents. Considering then the kinds of samples obtained according to the fundamental principles of "random selection" and "purposive selection", the author stresses their different capacity to answer the requirements of the definitions just considered. He then remarks that, while the theoretical superiority of the random selection method over the other one is indisputable, the cases are very rare in which it is possible to choose freely between the two methods. Indeed, accompanying circumstances and also the purpose of a particular enquiry often condition the actual sampling method and may necessitate the recourse to purposive sampling. In connection herewith, the theories are compared which were brought forward first by Bowley and then by Gini and Galvani, and the differences in their results are explained by the different systems of hypotheses on which those theories are based. Lastly, considering in detail the representative method as applied by means of random samples, and having distinguished the ensuing two stages by which it is characterised - deduction from the universe to the sample and induction from the sample to the universe - the author refers to the logical insufficiency of the latter stage, which was pointed out by some writers, Bowley, Gini and others. However, in spite of that imperfection most applications of the representative method - when it was possible to check the results obtained - have shown that such results were satisfactory for practical purposes. An explanation is then suggested which may clear up this puzzling contrast.</p>