The present study was addressed to the group AOV situation. Onte Carlo methods were employed to contrast several procedures with respect to a. ) control over Type 1 errors and b. ) power. (Editor)

The present study contrasted the robustness of the WSD and the F test to heterogeneity of variances. (Author)

Variance stabilizing transformations yield statistics whose standard errors are only influenced by the number of observations on which they are based. A solution that can be extended to inference on all of these parameters is presented and illustrated via three examples. (TJH)

Power and stability of Type I error rates are investigated for the Box-Scheffe test of homogeneity of variance with varying subsample sizes under conditions of normality and nonnormality. The test is shown to be robust to violation of the normality assumption when sampling is from a leptokurtic population. (Author/GK)

A distinction is made between statistics based on scientific theory and theory-free statistics. This distinction is discussed in the contexts of hypothesis testing, Bayesian inference, a priori planned contrasts, a new simple computational method, and alternative data interpretations. (TJH)

The logical flaw of using the methods of path analysis and structural equation modeling to make causative conclusions is demonstrated. A proper evaluation of the role of investigation versus experimentation is cited in the work of W. G. Cochran as explicated by D. R. Rubin (1983). (TJH)