Extensions of the Kruskal-Wallis procedure for a factorial design are examined under various degrees and kinds of nonnullity. It was found that the distributions of these test statistics are a function of effects other than those being tested, except under the completely null situation. Their use is discouraged. (Author/JKS)

A recent paper by Wainer and Thissen has renewed the interest in Gini's mean difference, G, by pointing out its robust characteristics. This note presents distribution-free asymptotic confidence intervals for its population value in the one sample case and for difference in the two sample situations. (Author/JKS)

Compared the analysis of variance (ANOVA) "F" and several nonparametric competitors for two-way designs for empirical alpha and power through simulation. Results suggest the ANOVA "F" suffers from conservative alpha and power for the mixed normal distribution, but is generally recommended. (Author/SLD)

An explicit statement of a statistic which is a nonparametric analog to one-way MANOVA is presented. The statistic is a multivariate extension of the nonparametric Kruskal-Wallis test (1952). In addition two post hoc procedures are developed and compared. (Author/JKS)

The author contends Glass, Peckham, and Sanders erred in discouraging the use of nonparametric counterparts of the t-test, even when it was known data were sampled from skewed distributions. He believes Wilcoxon's rank-sum test has power properties that make it preferable in most nonnormal population situations. (Author/DWH)

The basic assumption of Cooper's nonparametric test for trend (EJ 125 069) is questioned. It is contended that the proper assumption alters the distribution of the statistic and reduces its usefulness. (JKS)

A nonparametric test of dispersion with paired replicates data is described which involves jackknifing logarithmic transformations of the ratio of variance estimates for the pre- and posttreatment populations. Results from a simulation show that the test performs well under the null hypothesis and has good power properties. (Author/JKS)

The Shine-Bower error term is used to form approximate F ratios for testing various effects in the Shine-Bower single-subject ANOVA and the Shine Combined ANOVA. Results demonstrate the utility of these F ratios with respect to the probability of a type I error. (Author/JKS)

A distribution-free test for trend in one-way experimental designs is described. Critical values are derived, and a schematic diagram is given so that a computer package can be implemented. (Author/JKS)

The problem of determining which of several populations has the largest mean is considered. The procedure described by Dudewicz and Dalal is extended to the case of unequal sample sizes. (JKS)

Defining population diversity as the distribution along a continuum of homogeneity to heterogeneity with respect to one or more variables, measures are described with variance-based interpretations which can be applied to qualitative data. It is possible to use these statistics to define relationships in multivariate distributions. (Author/CTM)

Dunnett's procedure for comparing K-1 treatments with a control is discussed within the context of three nonparametric models: those of Kruskal-Wallis, Friedman, and Cochran. (Author/MH)

A statistical test for analysis of repeated measures is presented. It is shown that an appropriate chi-square statistic can be computed for related measures by taking account of the covariance among trials in computing a variance for the chi square statistic. (Author/JKS)

The one-sided problem of comparing treatments with a standard on the basis of data available in the context of a one-way analysis of variance is examined, and the methodology of S. Chakraborti and J. D. Gibbons (1991) is extended to the case of unequal sample sizes. (SLD)

Twelve variance-analytical and nonparametrical coefficients of reliability for rating scales designed for rating persons were compared to each other theoretically and empirically. Preference for two coefficients was established. The intraclass correlation coefficient appeared to be useful for the estimation of reliability as well. (Author/RL)