ABSTRACT
The Central Limit Theorem (CLT) discussed in the previous chapter can give a false sense of security that, as far as test procedures are concerned, all is well. In reality, for all practical purposes, the mean of a random sample of X variates may be trusted to be distributed Normally. Yet there may be a loose connection between the curve property E (X ) and the location parameter within the actual model that describes X ’s distribution.
