Abstract
A program is described the purpose of which is to select the most powerful among four methods for conducting a priori comparisons in an analysis of variance. The approaches considered are: orthogonal contrasts, Scheffe's method, Dunn's method, and Dunnett's test. The user supplies the program with the specifications of the analysis including number of cells, sample sizes, contrast coefficients, and the desired experiment-wise Type I error risk. The program supplies the critical t ratio and the per-comparison Type I error risk for each of the relevant methods along with a specification as to which is the most powerful. Within certain limitations, the program is capable of dealing with both one-way and factorial designs. It also can consider both one-tailed and two-tailed tests.
Get full access to this article
View all access options for this article.