Assignment of subjects to treatment groups on the basis of a covariate is discussed. It is contended that in making such assignments, consideration should be focused on the conditional expectations of the dependent variable given in the covariate in both treatment and control groups. Statistical development and an example are presented. (JKS)

This paper surveys available techniques and introduces an explicit statement of a new statistic to test for equality of correlated proportions in a polychotomous response design. A set of guidelines for the potential user of the techniques is provided. (Author/CTM)

The recently developed log-linear model technique for the analysis of contingency tables has many potential applications within educational research. This paper describes the two major models, log-linear and logit-linear, that are employed under this approach. (Author/JKS)

Valid statistical tests for one-subject experiments are necessary to justify statistical inferences and to ensure the acceptability of research reports to a wide range of journals and readers. The validity of randomization tests for one-subject experiments is examined. (See TM 505 800-801).(Author/JKS)

Two types of problems supposedly associated with the use of randomization tests for single-subject experiments have been discussed: the random introduction of treatments and the repeated alternation of treatments. Ways to reduce the adverse effects associated with these problems are presented. (See TM 505 799-800). (Author/JKS) (Author/JKS)

Stochastic approximation is suggested as a useful technique in areas where individuals have a goal firmly in mind, but lack sufficient knowledge to design an efficient, more traditional experiment. One potential area of application for stochastic approximation is that of formative evaluation. (CTM)

Discusses the problem of assigning subjects to one of several treatments based on their test scores. The success of each treatment is measured by different criterion. (RB)

When the experimental units are measured twice, and the response variable is dichotomous, the equality of the two proportions is usally assessed by Mc Nemar's (1947) test. In this paper, Bayesian methods are presented for testing hypotheses regarding the two success probabilities in light of complete and incomplete data. (Author/BW)

Randomization tests involve generating all possible assignments of subjects to treatments in order to create a distribution of possible test statistics. The use of a premutation group to enhance the understanding and utility of randomization tests is discussed. (Author/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)

A statistical analysis procedure is developed, based on the notion that many educational programs are dynamic interventions in natural growth processes, and is called value-added analysis. The theory of value-added analysis, and several applications are presented. (Author/JKS)

Problems associated with randomization tests in single- case experiments are discussed. This article follows a discussion of randomization tests in single case studies in the same issue of this journal. (See TM 505 799; 505 801).(Author/JKS)

This paper outlines the problems associated with the estimation of variance components in generalizability analyses using analysis of variance software, and discusses the most useful software currently available for this specialist application: the MIVQUE method of the Statistical Analysis System (SAS) procedure VARCOMP. (Author)

Glass's estimator of effect size, the sample mean difference divided by the sample standard deviation, is studied in the context of an explicit statistical model. The exact distribution of Glass's estimator is obtained and the estimator is shown to have a small sample bias. Alternatives are proposed and discussed. (Author/JKS)

The evaluation of competing analysis strategies based on estimator bias and variance is demonstrated using gains in standard scores and analysis of covariance procedures for quasi-experiments conforming to the fan-spread hypothesis. The findings do not lead to a uniform recommendation of either approach. (Author/JKS)