g Suppose a company made all the salespeople who had poor sales in February go through a training program in March. In April, these salespeople had much better sales than in February. In this case, the change may simply be due to the fact that for salespeople with the same level of effectiveness, sales vary anyway, and the salespeople who were identified as poor performers were really no different from others. The improvement in April has not occurred due to the training, and it simply exhibits the fact that they had an average month. This is an example of , and this problem could be eliminated by using a proper sampling technique. cregg