摘要:

Four multiple range tests for specifying homogeneous subsets of treatment variances are proposed and investigated. The procedures are based on Hartley'sF-max statistic, Cochran's statistic, Bartlett and Kendall's normalizing log transformation of the sample variances, and Box's approximate test. The philosophy underlying these tests, with respect to Type-I errors, is that advocated by Newman and later by Keuls. Empirical significance levels are obtained by simulating independent random samples from normal, uniform, double exponential and Cauchy distributions. The Box test was the only procedure not seriously affected by nonnormality.

ISSN:0162-1459    

DOI:10.1080/01621459.1975.10480283

出版商:Taylor & Francis Group    

年代:1975

数据来源: Taylor

摘要:

Tukey's well-known test for additivity in a two-way classification is considered for a randomized block design under the model arising from the physical randomization process. It is shown that under certain weak assumptions the usual normal theory test is still appropriate.

ISSN:0162-1459    

DOI:10.1080/01621459.1975.10480284

出版商:Taylor & Francis Group    

年代:1975

数据来源: Taylor

摘要:

A relatively simple method for computing the exact probability and distribution of theQstatistic is proposed.

ISSN:0162-1459    

DOI:10.1080/01621459.1975.10480285

出版商:Taylor & Francis Group    

年代:1975

数据来源: Taylor

摘要:

This article is concerned with the relative merit of the procedures proposed by Lin [2], Metha and Gurland [3], and Morrison [5] for testing the equality of means of a bivariate normal distribution with correlation ρ and common variance σ2when some observations on one variable are missing. A Monte Carlo study of the powers and level of significance is conducted. Comparisons among these tests and the pairedttest are made. The preferred test is established for each combination of ρ and difference of means.

ISSN:0162-1459    

DOI:10.1080/01621459.1975.10480286

出版商:Taylor & Francis Group    

年代:1975

数据来源: Taylor

ISSN:0162-1459    

DOI:10.1080/01621459.1975.10480287

出版商:Taylor & Francis Group    

年代:1975

数据来源: Taylor

摘要:

The Likelihood Ratio Test statistic,T, is considered for the hypothesisH: θ = θ0againstA: θ ≠ θ0in the nonlinear regression modely=f(x, θ) +ewith normal errors and unknown variance. The distribution function of a random variableXsuch thatn· (T—X) converges in probability to zero is derived. UsingXto approximateT, the power of the Likelihood Ratio Test is tabulated for selected sample sizes and departures from the null hypothesis. The adequacy of the approximation ofTbyXis investigated in a Monte-Carlo study.

ISSN:0162-1459    

DOI:10.1080/01621459.1975.10480288

出版商:Taylor & Francis Group    

年代:1975

数据来源: Taylor

摘要:

In an earlier article [8] conditions were given under which maximum likelihood estimators are asymptotically normal and asymptotically efficient in certain nonstandard cases. Under these same conditions, the asymptotic distribution of the likelihood ratio is found. An easily computed asymptotic equivalent of the likelihood ratio is developed, and is applied to a problem of testing fit.

ISSN:0162-1459    

DOI:10.1080/01621459.1975.10480289

出版商:Taylor & Francis Group    

年代:1975

数据来源: Taylor

摘要:

Critical values for the one-sample and the two-sample Kolmogorov-Smirnov statistics with a particular weight function are compared and found to be greatly different. The difference is shown to be due to the marked sensitivity of the one-sample statistic to the values of the first and last order statistics of the sample distribution.

ISSN:0162-1459    

DOI:10.1080/01621459.1975.10480290

出版商:Taylor & Francis Group    

年代:1975

数据来源: Taylor

摘要:

It is shown that convergence in probability to Zipf's Law follows from a many family modification of a Bose-Einstein form of the classical occupancy model with a random number of cells.

ISSN:0162-1459    

DOI:10.1080/01621459.1975.10480291

出版商:Taylor & Francis Group    

年代:1975

数据来源: Taylor

摘要:

The application is discussed of Johnson's transformed normal density to the distribution of skewness (√b1) and kurtosis (b2) in sampling from normal mixtures (two-components, equal variances). In particular, approximations to the distribution function of √b1(andb2) are found using the Cornish-Fisher expansion withSUas kernel. Comments on the approximations in sampling from other distributions are given, and comparisons with alternative approximations (Monte-Carlo assessments, etc.) included.

ISSN:0162-1459    

DOI:10.1080/01621459.1975.10480292

出版商:Taylor & Francis Group    

年代:1975

数据来源: Taylor