ABSTRACT
A common feature of many experimental programs is to obtain groups of data under two or more sets of experimental conditions. The question then arises: Has the change in experimental conditions affected the data? The question may be rephrased to a more precise form: Do the means of each group differ significantly or are all groups really taken from the same population, the change in experimental conditions having had no significant effect? A variety of experimental techniques exist to answer this question. Hence, it is all too easy to select an inappropriate technique, with misleading results.
