ABSTRACT
Generally, investigators abhor negative findings. Thus, it is not uncommon for researchers to attempt any of a variety of data manipulations to find some support for the contentions underlying their study. The breadth of available statistical methods makes it possible to generate support for almost any conclusion if one is willing to disregard some of the fundamental issues concerning probability. Usually, this disregard occurs in either of two ways. One is to bend or ignore assumptions of the statistical method selected. The second is to afford an unrealistic weight to certain alpha levels selected when comparisons of data sets are made. Either can lead to faulty or misleading conclusions.
