1. Maurice Allais
2. Huber-Darmois
3. R.A. Fisher
4. Bourdieu
5. Mahalanobis
6. G. Calot
7. Paul Valéry
8. Edmond Malinvaud
NOTES DE LECTURE: R.A. FISHER
extrait de R.A. Fisher: (1956/1959): Statistical Methods and Scientific Inference
Tests of significance and acceptance decisions
The common tests of significance, familiarly known
as Pearson's chi-square test of goodness of fit (1900), "Student"'s t-test (1908), the z (or F)
test of analysis of variance (1924), and many others designed on the
same principles, have come in the first two quarters of the twentieth
century to play a rather central part in statistical analysis. In the
day-to-day work of experimental research in the natural sciences, they
are constantly in use to distinguish real effects of importance to a
research programme from such apparent effects as might have appeared in
consequence of errors of random sampling or of uncontrolled
variability, of any sort, in the physical or biological material under
examination. They are used to recognize, among innumerable examples
that could be given, the genuineness of a genetic linkage, the reality
of the response to manurial treatment of a cultivated crop, the
deterioration of a food product in storage, or the difference between
machines in the frequency of defective parts produced by them. The
conclusions drawn from such tests constitute the steps by which the
research worker gains a better understanding of his experimental
material, and of the problems which it presents.
......
The attempts that have been made to explain the cogency of tests of significance in scientific research, by reference to supposed frequencies of possible statements, based on them, being right or wrong, thus seem to miss the essential nature of such tests...
On the whole the ideas (a) that a test of significance must be regarded as one of a series of similar tests applied to a succession of similar bodies of data, and (b) that the purpose of the test is to discriminate or "decide" between two or more hypotheses, have greatly obscured their understanding, when taken not as contingent possibilities but as elements essential to their logic.
......
The attempts that have been made to explain the cogency of tests of significance in scientific research, by reference to supposed frequencies of possible statements, based on them, being right or wrong, thus seem to miss the essential nature of such tests...
On the whole the ideas (a) that a test of significance must be regarded as one of a series of similar tests applied to a succession of similar bodies of data, and (b) that the purpose of the test is to discriminate or "decide" between two or more hypotheses, have greatly obscured their understanding, when taken not as contingent possibilities but as elements essential to their logic.
Commentaire
Le texte ci-dessus réaffirme la position de Fisher vis-à-vis du fréquentisme radical (Neyman-Pearson), la doxa de la statistique académique. Un test de signification n'est nullement une "décision" entre les deux options (rejeter ou pire "accepter" H0); obtenue en "choisissant" a priori un seuil alpha (tel que .05 ou .01).
Le texte ci-dessus réaffirme la position de Fisher vis-à-vis du fréquentisme radical (Neyman-Pearson), la doxa de la statistique académique. Un test de signification n'est nullement une "décision" entre les deux options (rejeter ou pire "accepter" H0); obtenue en "choisissant" a priori un seuil alpha (tel que .05 ou .01).
A
propos du statut singulier de R.A. Fisher, unanimement
salué comme le plus
grand statisticien du 20ème siècle, si proche
des problèmes des chercheurs, et pourtant si
malmené par la statistique académique, j'ai
écrit dans Rouanet
& al.
1997 (Peter Lang):
Fisher's contributions to modern statistical inference have been more influential than those of any other statistician. The practice of researchers is largely Fisher-inspired. Unfortunately, fiducial inference was dismissed by mathematical statisticians. Hence the split-brain situation that is prevailing today: Fisher is the patron saint of researchers, but Neyman-Pearson is the established church of statistical inference.
Retour à la page d'accueil