A calculational strategy presented in educational statistics texts for computing second and higher order semipartial correlations is critically examined and shown to be incorrect. (Author)

Estimated Type I Error rates and power are reported for the Brown-Forsythe, O'Brien, Klotz, and Siegal-Tukey procedures. The effect of aligning the data using deviations from group means or group medians is investigated. (RB)

The interrelationships among within-aggregate, between-aggregate, and total-group correlation coefficients are summarized with artificial and real-data examples. The relevance of correlation analyses at various levels of aggregation and some of the difficulties encountered in cross-level inference are also discussed. (Author/JKS)

Observed percentiles provide better estimates when they are based on grouped data rather than on raw data. This is especially true when the data set is relatively small and the population distribution is not too skewed or heavy-tailed. Applied statisticians could take advantage of this fact by implementing a program for estimating percentiles from…

Statistical issues concerning the analysis of multilevel data common in educational studies are discussed. Computer simulation results are presented to argue for a modification of two currently popular approaches to multilevel or contextual analytical procedures. (JKS)

Fisher's exact probability test is severely conservative when interpreted with reference to conventional alpha levels due to the discontinuity of the sampling distribution for 2 x 2 tables. An adjustment of the cell frequencies is proposed that results in a correction for continuity with appropriate alpha protection and increased power. (Author)

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)

Statistical techniques for summarizing results from independent correlational studies are presented. The case in which only the sample correlation coefficients are available and the case in which the original paired data are available are both considered. (Author/JKS)

Illustrates a simple nonparametric alternative that can be used to test a hypothesis that two proximity matrices on the same set of variables or objects reflect a similar pattern of high and low entries. (RC)

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)

Presents an exposition of two data reduction methods--single-link and complete-link hierarchical clustering. Emphasis is on statistical techniques for evaluating the adequacy of a completed partition hierarchy and the individual partitions within the sequence. A numerical reanalysis of data illustrates the methodology. (RC)

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)

Monte Carlo techniques were used to compare the power of Wilcoxon's rank-sum test to the power of the two independent means t test for situations in which samples were drawn from (1) uniform, (2) Laplace, (3) half-normal, (4) exponential, (5) mixed-normal, and (6) mixed-uniform distributions. (Author/JKS)

The effect of the variance stabilizing transformations on the significance level and power of the F test, applied to Binomial and Poisson variables, was examined by a Monte Carlo study. The results indicate that the significance level is not affected by the transformations. (Author/JKS)

Suggests that improving the data collected for a study will lead to the development of better models to analyze that data. Also urges researchers to have a more critical attitude toward the literal interpretation of test results. (RB)