Ralph Bradley, University of Georgia

A Different Look at Analysis of Variance and Design of Experiments

Beginning courses on statistical methods are not favorites of unsophisticated students. Beginning students often feel engulfed by sets of formulas and at best learn how to substitute numbers into those formulas. We suggest bypassing all of this with one simple program. The user does not need to know any details of the experimental design, only the treatment applied to each experimental unit. A number of examples will be given. We now teach analysis of variance not design of experiments. Simplification of the details of statistical analyses will leave time to teach real design of experiments. Some comments on how to do this will be given and suggestions from the audience solicited.