Table of contents, International Statistical Review, vol. 73, no. 1 (2005)

Much Ado About Nothing: the Mixed Models Controversy Revisited; Authors: Viviana B. Lencina, Julio M. Singer, Edward J. Stanek Iii

Simple Stepwise Tests of Hypotheses and Multiple Comparisons; Authors: Eugene Seneta, John T. Chen

Mixed Poisson Distributions; Authors: Dimitris Karlis, Evdokia Xekalaki

Use of Statistics in the Monetary Policy of the Czech National Bank: The Case of a Country in Transition; Authors: Ivan Matalík, Josef Arlt

Automatic Editing for Business Surveys: An Assessment of Selected Algorithms; Authors: Ton de Waal, Wieger Coutinho

Extending Power and Sample Size Approaches Developed for McNemar's Procedure to General Sign Tests; Authors: Donald R. Hoover

The $t$ Copula and Related Copulas; Authors: Stefano Demarta & Alexander J. Mcneil

Fisher and Inference for Scores; Authors: Alan H. Welsh, John Robinson

Editorial board, International Statistical Review, vol. 73, no. 1 (2005)

E. Seneta (School Math. & Statist., Univ. Sydney, Australia), Joint Editor

A.M. Manninen (City of Helsinki, Urban Facts, Finland)., Joint Editor

Ann Daniels (ISI, Voorburg, The Netherlands)., Executive Editor

R. Acx (National Bank of Belgium, Brussels), Associate Editor

T. Alanko (Statistics Finland), Associate Editor

R. Gill (Univ. Utrecht, The Netherlands),, Associate Editor

P. Guttorp (Univ. Washington, Seattle, USA), Associate Editor

W. Lutz (Internat. Inst. Appl. Systems Analysis, Laxenburg, Austria),, Associate Editor

N. Singpurwalla (George Washington Univ.,Washington, USA),, Associate Editor

I. Stamhuis (Free Univ., Amsterdam, The Netherlands),, Associate Editor

P. Verdier (INSEE, Paris, France),, Associate Editor

N. Weber (School Math. & Statist., Univ. Sydney, Australia),, Associate Editor

E. Wegman (George Mason Univ., Fairfax, USA),, Associate Editor

C. Wild (Univ. Auckland, New Zealand)., Associate Editor

T.E. Dalenius,, Former Editor

L.A.M. Vessereau,, Former Editor

W.R. Buckland,, Former Editor

O.E. Barndorff-Nielsen,, Former Editor

D.R. Brillinger,, Former Editor

B.W. Silverman,, Former Editor

D.J. Trewin,, Former Editor

V.N. Nair,, Former Editor

W.F.M. de Vries,, Former Editor

E. Arjas., Former Editor

We consider a well-known controversy that stems from the use of two mixed models for the analysis of balanced experimental data with a fixed and a random factor. It essentially originates in the different statistics developed from such models for testing that the variance parameter associated to the random factor is null. The corresponding hypotheses are interpreted as that of null random factor main effects in the presence of interaction. The controversy is further complicated by different opinions regarding the appropriateness of such hypothesis. Assuming that this is a sensible option, we show that the standard test statistics obtained under both models are really directed at different hypotheses and conclude that the problem lies in the definition of the main effects and interactions. We use expected values as in the fixed effects case to resolve the controversy showing that under the most commonly used model, the test usually associated to the inexistence of the random factor main effects addresses a different hypothesis. We discuss the choice of models, and some further problems that occur in the presence of unbalanced data.

Holm's (1979) step-down and Hochberg's (1988) step-up procedures for tests of multiple hypotheses are simple to apply and are widely used. Holm's procedure controls the familywise error rate (FWE), while Hochberg's is more powerful. This paper investigates a step-down procedure (labelled CS) of Seneta & Chen (1997) which is a sharpening of Holm's, takes into account the degree of association between test statistics, and also controls the FWE. Computation for the CS procedure may be minimized by using the procedure as an adjustment to Holm's. The computational steps are detailed, and the adjustment is then illustrated by an application to a text-book example of multiple comparisons, in which step-wise procedures are shown to perform better than the usual Tukey T-comparison. Simulation investigations in a standard comparison with a control setting show that the CS step-down procedure is more powerful than Hochberg's step-up procedure and the procedure of Simes (1986), especially in regard to error rate, and not much less powerful than an optimal, but very specific, step-up procedure of Dunnett & Tamhane (1992).

Mixed Poisson distributions have been used in a wide range of scientific fields for modeling non-homogeneous populations. This paper aims at reviewing the existing literature on Poisson mixtures by bringing together a great number of properties, while, at the same time, providing tangential information on general mixtures. A selective presentation of some of the most prominent members of the family of Poisson mixtures is made.

In the presented text the authors judge the importance of statistics in the monetary policy of the Czech National Bank (CNB) over the course of the economic transformation process, with particular consideration of changing statistical needs and the possibilities and limits of statistical data exploitation in the monetary analyses. The importance of statistics lies on the level of collection and processing of statistical information and on the level of use of statistical methods to analyse data. Since the start of the 1990s the requirements for statistics were significantly influenced by monetary policy. In the period 1990-1997, monetary targeting was the primary influential factor. Since 1998, the monetary policy is influenced by inflation targeting. Statistical priorities switched from monetary data to economy and financial market data. Much progress has been made in the use of statistical methods for analysing data. Statistics available at present cover the CNB's standard monetary-policy requirements and are on par with those in developed countries. Its further development will reflect the standard changes taking place in the more advanced countries.

Statistical offices are responsible for publishing accurate statistical information about many different aspects of society. This task is complicated considerably by the fact that data collected by statistical offices generally contain errors. These errors have to be corrected before reliable statistical information can be published. This correction process is referred to as statistical data editing. Traditionally, data editing was mainly an interactive activity with the aim to correct all data in every detail. For that reason the data editing process was both expensive and time-consuming. To improve the efficiency of the editing process it can be partly automated. One often divides the statistical data editing process into the error localisation step and the imputation step. In this article we restrict ourselves to discussing the former step, and provide an assessment, based on personal experience, of several selected algorithms for automatically solving the error localisation problem for numerical (continuous) data. Our article can be seen as an extension of the overview article by Liepins, Garfinkel & Kunnathur (1982). All algorithms we discuss are based on the (generalised) Fellegi-Holt paradigm that says that the data of a record should be made to satisfy all edits by changing the fewest possible (weighted) number of fields. The error localisation problem may have several optimal solutions for a record. In contrast to what is common in the literature, most of the algorithms we describe aim to find all optimal solutions rather than just one. As numerical data mostly occur in business surveys, the described algorithms are mainly suitable for business surveys and less so for social surveys. For four algorithms we compare the computing times on six realistic data sets as well as their complexity.

Current software and textbooks present procedures to estimate power and sample size for sign tests that only apply to settings where positive (i.e., X=1) or negative (i.e., X=-1), but not neutral (i.e., X=0) outcomes occur. However, many studies analyzed by sign tests involve the more general setting where significant amounts of neutral outcomes can occur. This paper illustrates extension of existing power/sample size approaches and software that have been developed for matched binary responses (McNemar's discordant pairs) to general sign tests with neutral outcomes occurring. An application is made to a recent study that the author collaborated on.

The t copula and its properties are described with a focus on issues related to the dependence of extreme values. The Gaussian mixture representation of a multivariate t distribution is used as a starting point to construct two new copulas, the skewed t copula and the grouped t copula, which allow more heterogeneity in the modelling of dependent observations. Extreme value considerations are used to derive two further new copulas: the t extreme value copula is the limiting copula of componentwise maxima of t distributed random vectors; the t lower tail copula is the limiting copula of bivariate observations from a t distribution that are conditioned to lie below some joint threshold that is progressively lowered. Both these copulas may be approximated for practical purposes by simpler, better-known copulas, these being the Gumbel and Clayton copulas respectively.

This paper examines the work of Fisher and Bartlett on discriminant analysis, ordinal response regression and correspondence analysis. Placing these methods with canonical correlation analysis in the context of the singular value decomposition of particular matrices, we use explicit models and vector space notation to unify these methods, understand Fisher's approach, understand Bartlett's criticisms of Fisher and relate both to modern thinking. We consider in particular the formulation of certain hypotheses and Fisher's arguments to obtain approximate distributions for tests of these hypotheses (without assuming multivariate normality) and put these in modern notation. Using perturbation techniques pioneered by G.S. Watson, we give an asymptotic justification for Fisher's test for assigned scores and thereby resolve a long standing conflict between Fisher and Bartlett.