ABSTRACT
So far we have assumed that random variables X1; X2; : : : ; Xn forming a sample are mutually independent and identically distributed. For example, in the toss of a dice, if Xt is a random variable describing the t th outcome, we may suppose that the outcome Xt does not a®ect any other outcomes Xs, s 6= t. Hence the setting of independence seems natural. If X1; : : : ; Xn show the height of n students randomly sampled in a school class, then we may also accept the setting of independence naturally.
