摘要:

In this article the Tukey-Kramer procedure for multiple comparisons of pairwise differences of mean vectors in multivariate normal distributions is considered. A multivariate version of the Tukey-Kramer procedure is presented, and a generalized Tukey conjecture of the conservativeness of the simultaneous confidence intervals for all pairwise comparisons by this procedure is affirmatively proved in the case of three correlated mean vectors. Some properties of the multivariate Tukey-Kramer procedure are also presented, and simulation results for some selected parameters are given.

ISSN:0162-1459    

DOI:10.1080/01621459.1994.10476792

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

This article illustrates an analysis strategy for repeated measures data where some of the measurements are missing from the subjects on one or more occasions, with a special application to data arising from repeated measurement designs when some observations are missing after the first treatment period. Using all available data, “almost” maximum likelihood estimators for the parameters of interest are obtained under the assumption of a multivariate normal distribution of the data. Inference procedures for these parameters in small samples from popular two-treatment designs are discussed.

ISSN:0162-1459    

DOI:10.1080/01621459.1994.10476793

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Constrained and compound optimal designs represent two well-known methods for dealing with multiple objectives in optimal design as reflected by two functionals φ1and φ2on the space of information matrices. A constrained optimal design is constructed by optimizing φ2subject to a constraint on φ1, and a compound design is found by optimizing a weighted average of the functionals φ = λφ1+ (1 - λ) φ2, 0 ≤ λ ≤ 1. We show that these two approaches to handling multiple objectives are equivalent.

ISSN:0162-1459    

DOI:10.1080/01621459.1994.10476794

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Two distinct types of models are used for handling nonresponse in survey sampling theory. In a response (or quasi-randomization) model, the propensity of survey response is modeled as a random process, an additional phase of sample selection. In a parametric (or superpopulation) model, the survey data are themselves modeled. These two models can be used simultaneously in the estimation of a population mean so that one provides some protection against the potential for failure in the other. Two different estimators are discussed in this article. The first is a regression estimator that is both unbiased under the parametric model and nearly quasi-design unbiased under the response model. The second is a direct expansion estimator with imputed missing values. The imputed values are such that the estimator is both nearly quasi-design unbiased and unbiased under the combination of the parametric model and the original sampling design. The article includes a discussion of variance estimation with the goal of simultaneously estimating quasi-design mean squared error and either parametric model variance or combined (parametric model and original design) variance.

ISSN:0162-1459    

DOI:10.1080/01621459.1994.10476795

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

This article presents the Gibbs-Skovgaard algorithm for approximate frequentist inference. The method makes use of the double saddlepoint approximation of Skovgaard to the conditional cumulative distribution function of a sufficient statistic given the remaining sufficient statistics. This approximation is then used in the Gibbs sampler to generate a Markov chain. The equilibrium distribution of this chain approximates the joint distribution of the sufficient statistics associated with the parameters of interest conditional on the observed values of the sufficient statistics associated with the nuisance parameters. This Gibbs-Skovgaard algorithm is applied to the cases of logistic and Poisson regression.

ISSN:0162-1459    

DOI:10.1080/01621459.1994.10476796

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

In the Bayesian classification of satellite images, a prior distribution is used that aims to model the belief of spatial homogeneity of the underlying region. We extend this prior distribution to model certain topographical features of the area such as the position of the roads, the slopes, and the aspects. We demonstrate the effectiveness of this prior distribution in a reconstruction algorithm by means of a simulation study in which the quality of the result is assessed by a comparison of estimated and known covertypes. We apply the algorithm to real data with success.

ISSN:0162-1459    

DOI:10.1080/01621459.1994.10476797

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Pitman's Measure of Closeness: A Comparison of Statistical EstimatorsJerome P. Keating, Robert L. Mason, and Pranab K. Sen, Philadelphia: Society for Industrial and Applied Mathematics (SIAM), 1993. xv + 226 pp., $26.50.

ISSN:0162-1459    

DOI:10.1080/01621459.1994.10476798

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

ISSN:0162-1459    

DOI:10.1080/01621459.1994.10476799

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

ISSN:0162-1459    

DOI:10.1080/01621459.1994.10476801

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

ISSN:0162-1459    

DOI:10.1080/01621459.1994.10476803

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor