Statistiques appliquées aux études de cas unique : méthodes usuelles et alternatives (notice n° 1027825)
[ vue normale ]
| 000 -LEADER | |
|---|---|
| fixed length control field | 03950cam a2200265 4500500 |
| 005 - DATE AND TIME OF LATEST TRANSACTION | |
| control field | 20250125171653.0 |
| 041 ## - LANGUAGE CODE | |
| Language code of text/sound track or separate title | fre |
| 042 ## - AUTHENTICATION CODE | |
| Authentication code | dc |
| 100 10 - MAIN ENTRY--PERSONAL NAME | |
| Personal name | Atzeni, Thierry |
| Relator term | author |
| 245 00 - TITLE STATEMENT | |
| Title | Statistiques appliquées aux études de cas unique : méthodes usuelles et alternatives |
| 260 ## - PUBLICATION, DISTRIBUTION, ETC. | |
| Date of publication, distribution, etc. | 2009.<br/> |
| 500 ## - GENERAL NOTE | |
| General note | 51 |
| 520 ## - SUMMARY, ETC. | |
| Summary, etc. | RésuméLe traitement statistique des données issues d’études neuropsychologiques de cas unique est un aspect fondamental dans le processus de validation de modèles de la cognition humaine. Pour déterminer si un individu présente des performances anormales au regard de celles présentées par un groupe témoin, ou pour détecter la présence de dissociations dans ses performances, toujours au regard de celles présentées par un groupe témoin, on utilise très largement des scores standardisés en considérant les statistiques d’échantillon comme des paramètres populationnels. Lorsque la taille de l’effectif du groupe témoin est faible (n < 15) cela conduit à augmenter la probabilité de commettre une erreur de type I. Augmentation qui sera d’autant plus importante si un faible n est associé à d’autres caractéristiques, comme par exemple une distribution asymétrique des performances du groupe témoin. Nous présentons dans cet article une recension non exhaustive de différentes solutions, relativement récentes, permettant de minimiser la probabilité de commettre une erreur de type I tout en préservant une puissance statistique acceptable. |
| 520 ## - SUMMARY, ETC. | |
| Summary, etc. | Statistics applied to single case studies: classical and alternative methodsThe statistical treatment of the data stemming from single-case studies in neuropsychology is a fundamental aspect in the validation process of human cognition models. Generally speaking the aim of single-case studies is twofold: addressing the question of whether or not a patient exhibit a (statistically significant) deficit on one hand and comparing the difference between a patient’s performances on two or more tasks — i.e. establishing dissociations of cognitive functions — on the other hand. One of the three basic approaches to answer these two questions is to compare the patient with a matched control sample. The statistical analysis widely used consists in converting the patient’s performance to a z score based on the mean and the standard deviation of the control sample. Then this z is referred to a table of areas under the normal curve. The crucial problem with this approach is that it considers the control sample as though it was a population and then treats their statistics as parameters. When the size of the control sample is reasonably large (n > 50) the parameters of the population and the estimates of these parameters converge. But when the sample size is modest (n < 15) using z score lead to an inflation of the Type one error rate and then an overestimation of the abnormality of a patient’s score. The increase of the Type one error rate is all the more large that a modest n is associated with other characteristics like a departure from normality of the scores of the control sample. The aim of this article is to set out a review of the main relatively recent different statistical methods developed to minimize the Type one error rate. We will see that a set of method proposed by Crawford and colleagues allow a good control of Type one error rate while keeping a satisfactory statistical power. |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | étude de cas unique |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | erreur de type I |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | dissociation |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | méthodes statistiques |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | type I error |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | single case studies |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | dissociation |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | statistical methods |
| 786 0# - DATA SOURCE ENTRY | |
| Note | Revue de neuropsychologie | Volume 1 | 4 | 2009-12-01 | p. 343-351 | 2101-6739 |
| 856 41 - ELECTRONIC LOCATION AND ACCESS | |
| Uniform Resource Identifier | <a href="https://shs.cairn.info/revue-de-neuropsychologie-2009-4-page-343?lang=fr&redirect-ssocas=7080">https://shs.cairn.info/revue-de-neuropsychologie-2009-4-page-343?lang=fr&redirect-ssocas=7080</a> |
Pas d'exemplaire disponible.
Réseaux sociaux