|
| << vorige | volgende >> |
Tijdschrift beschrijving
Alle jaargangen van het bijbehorende tijdschrift
Alle afleveringen van het bijbehorende jaargang
Alle artikelen van de bijbehorende aflevering
Details van artikel 14 van 17 gevonden artikelen
| Titel: |
The generalized likelihood ratio test for the pearson correlation |
| Auteur: |
Wan-Ching, Liu Woodward, J. Arthur Bonett, Douglas G. |
| Verschenen in: |
Communications in statistics |
| Paginering: | Jaargang 25 (1996) nr. 2 pagina's 507-520 |
| Jaar: |
1996 |
| Inhoud: |
The exact generalized likelihood ratio (GLR) test is defined for the test of #0:P = Po against Hx:p*p0 where p is an unknown value of the Pearson correlation coefficient and Po is a known constant (-1 < p0 < 1). The exact power functions of the GLR test and the Fisher's approximate test are unknown. A simple computational procedure is developed for these tests and the exact power functions are compared with each other and with David's exact equal-tails test. Results reveal that the Fisher's test more closely approximates the GLR test than the equal-tails test and is a very close approximation to the GLR test. For positive values of Po, the GLR test is more powerful than the equal-tails test when p > po; otherwise the equal-tails test is more powerful than the GLR test. For negativevalues of p0, the result is reserved. Finally, the exact power function of the Fisher's approximate test shows that the test is conservative when a is large, but liberal when a is small. |
| Uitgever: |
Taylor & Francis |
| Bronbestand: |
Elektronische Wetenschappelijke Tijdschriften |
Details van artikel 14 van 17 gevonden artikelen
| << vorige | volgende >> |