Two — Sample Rank Tests in General Models
✍ Scribed by Prof. Dr. E. Brunner; Dr. N. Neumann
- Publisher
- John Wiley and Sons
- Year
- 1986
- Tongue
- English
- Weight
- 287 KB
- Volume
- 28
- Category
- Article
- ISSN
- 0323-3847
No coin nor oath required. For personal study only.
✦ Synopsis
In the several sample case for independent observations suitable rank testa can be derived, if a location model ia aeeumed. Inmore general models it is difficult to deduce such statistics. If only two samples are considered, rank tests can be developed without probleme. For general two-sample designs rank statistics and their associated variance estimators are derived.
📜 SIMILAR VOLUMES
Starting from the discussion of a practical example a unifying concept for thederivation of meaningfully interpretable nonparametric tests for the two-sample cam is developed which may well be adapted for other designs, too. This methodology covere other well-known procedures, e.g. Fisher's exact t
The distribution-free test against ordered alternatives proposed by JONCKKEERE (1954) is based on the Kendall'e rank correlation coefficient 7. A new rank teat is proposed and illustrated with a numerical example. The proposed test is based on Spearman's and has similar functional structure as the K