摘要:

AbstractThe maximum range test as a closed multiple test procedure for all pairwise comparisons of several independent samples was described by ROYEN (1988) and (1989). For the number k of samples, the dimension d and some suitable degrees of freedom v between 10 and ∞ the following tables are contained in this paper:bounds for the studentized multivariate maximum range for 2≦k≦12, 2≦d≦6 and the familywise type I error rates α = 0.01 (table 1.1) and α = 0.01 (table 1.2), furthermore bounds for the studentized univariate maximum range for 2≦k≦15 and α = 0.01 (table 2). The corresponding bounds with α = 0.05 for univariate maximum ranges have been tabulated in ROYEN (1989).Some examples for the application of these bounds by maximum range t

ISSN:0323-3847    

DOI:10.1002/bimj.4710320602

出版商:WILEY‐VCH Verlag    

年代:1990

数据来源: WILEY

摘要:

AbstractDensities, cumulatives, and quantiles of the noncentral distributions are notoriously difficult to calculate. It is shown how the noncentral t distribution is represented in the canonical S‐system form, i.e., as a set of first‐order nonlinear differential equations in which the derivative of each variable equals a difference of two products of power‐law functions. By solving the S‐system numerically one rapidly obtains densities, cumulatives, quantiles, and moments of the central and noncentral t distributions, as well as power functions for the

ISSN:0323-3847    

DOI:10.1002/bimj.4710320603

出版商:WILEY‐VCH Verlag    

年代:1990

数据来源: WILEY

ISSN:0323-3847    

DOI:10.1002/bimj.4710320604

出版商:WILEY‐VCH Verlag    

年代:1990

数据来源: WILEY

摘要:

AbstractTests are proposed for comparing K treatments with a control when the samples are subject to unequal patterns of right censorhip. The tests are useful in the context of life testing experiments and clinical trials for survival analysis, where cost of experimentation is usually high, since the tests can be applied before all the units have failed. A numerical example is given to illustrate the implementation of the tests. The asymptotic distributions of the test statistics are obtained under both the null and the alternative hypotheses. Pitman efficacies of the tests are derived under the usual sequences of translation and proportional hazard alternatives. The size and the power of the tests are examined in a simulation study.

ISSN:0323-3847    

DOI:10.1002/bimj.4710320605

出版商:WILEY‐VCH Verlag    

年代:1990

数据来源: WILEY

摘要:

AbstractFallible classification scheme is known to produce inconsistent relative frequency estimates from a polychotomous variable. These estimates however can be improved upon by the use of a sub‐sample that results in a square table. In this research we highlight the method of improvemen

ISSN:0323-3847    

DOI:10.1002/bimj.4710320606

出版商:WILEY‐VCH Verlag    

年代:1990

数据来源: WILEY

ISSN:0323-3847    

DOI:10.1002/bimj.4710320607

出版商:WILEY‐VCH Verlag    

年代:1990

数据来源: WILEY

摘要:

AbstractFor a multivariate two way classification model with a single replicate per cell, likelihood ratio test for the interaction are derived assuming certain structures on the interaction. The underlying distribution is assumed to be elliptically symmetric.

ISSN:0323-3847    

DOI:10.1002/bimj.4710320608

出版商:WILEY‐VCH Verlag    

年代:1990

数据来源: WILEY

摘要:

AbstractOne way of comparing simultaneous confidence intervals is to compare the length of individual confidence intervals. Simultaneous confidence intervals are the better the shorter individual confidence intervals are. It is shown that such comparison has no sense.

ISSN:0323-3847    

DOI:10.1002/bimj.4710320609

出版商:WILEY‐VCH Verlag    

年代:1990

数据来源: WILEY

ISSN:0323-3847    

DOI:10.1002/bimj.4710320611

出版商:WILEY‐VCH Verlag    

年代:1990

数据来源: WILEY

摘要:

AbstractThe variance estimators usually applied for the generalized censored data Wilcoxon rank tests by Gehan and Peto&Prentice, are heavily biased in unbalanced problems. This paper reports the results of a Monte Carlo simulation study, where jackknifing is used to construct estimators of variance. Size, power and variance properties are compared for five variance estimators, when using different combinations of group sizes, failure and censoring patterns. The variance estimators are the permutational, the conditional permutational and the jackknife variance estimators for the statistic of Gehan and the asymptotic and the jackknife variance estimators for the statistic of Peto&Prentice. It appears that observed size, power and variance properties may be improved by using the jackknife variance estimator, when comparing to the variance estimators usually applied.

ISSN:0323-3847    

DOI:10.1002/bimj.4710320612

出版商:WILEY‐VCH Verlag    

年代:1990

数据来源: WILEY