A general analysis strategy is proposed such that the universe of inference is increased incrementally. The strategy prevents logically incongruent findings that occasionally result when the conventional analysis strategy is employed. (Author)

Multiple comparisons involve the examination of which group or groups are actually different from other group(s) in analysis of variance results. Such comparisons usually involve one-way analysis of variance. This monograph discusses designs more complex than one-way designs. (JKS)

Logically consistent hypothesis-testing for factorial analysis of variance designs is proposed in the context of a hierarchical model. It is shown that all of the hypotheses associated with the traditional factorial model are conceptually independent and occupy the lowest levels of the hierarchy. (Author/JKS)

The authors argue that a parameterization in terms of cell means is a useful conceptual device for resolving problems inherent in the nonorthogonal analysis of variance. They demonstrate their argument by considering two widely-used methods for testing main effects. (Author/BW)

The statistical power of nine k-sample tests against ordered location alternatives under completely randomized designs are investigated. The results are intended to aid researchers in selecting appropriate statistical procedures. (Author/JKS)

The need for multiple comparison procedures for repeated measures means employing a pooled estimate of error variance to conform to the sphericity assumptions of the design in order to provide a valid test is discussed. An alternative approach which does not require this assumption is presented. (Author/JKS)

A particular strategy for investigating effects from a multivariate analysis of variance (MANOVA) is proposed. The strategy involves multiple two-group multivariate analyses. The analysis strategy is described in detail and illustrated with real data sets. (Author/JKS)

Discusses the need in designing educational studies to give more attention to recognizing the universe of generalization and to incorporate into the design a variable that represents the universe. Data analytic techniques should treat the levels of this variable as randomly selected from the universe. (Author/JEG)

The multivariate analysis of variance approach to the analysis of data from repeated measures designs that include covariates is reviewed and contrasted with univariate methods. Assumptions necessary for meaningful interpretation of the results of the multivariate analysis are discussed and contrasted to those of the univariate approaches. An…

Issues in analysis of covariance, multiple regression analysis, and the analysis of variance such as the assumption of independence and directional hypotheses are discussed. (JKS)

Repeated measures designs have traditionally been analyzed by the univariate mixed model approach, in which the repeated measures effect is tested against an error term based on the subject by treatment interaction. This paper considers an extension of this analysis to designs in which the individual repeated measures are multivariate. (Author/JKS)

This article reviews some of the less well-known hypothesis tests for variance, how they are employed, and how the results may be interpreted. Tests include testing for a single variance and the T-test for correlated variances. (CT)

Discusses two approaches to data analysis in pretest-posttest research designs in which researcher has not formed groups based on pretest or used pretest information in assigning participants to groups, thus ruling out two-way analysis of variance. Describes use of repeated-measures analysis of variance and two-group analysis of covariance in…

To examine the ability of the correction factor epsilon to counteract statistical bias in univariate analysis, an analysis of variance (adjusted by epsilon) and a multivariate analysis of variance were performed on the same data. The results indicated that univariate analysis is a fully protected design when used with epsilon. (JMB)

Humphreys' comments on double-median splits (TM 504 009) are essentially correct, but are not relevant to the original Kirby and Das article (EJ 182 444). His comments do not weaken our findings. (Author/RD)