2. Derive the moment-generating function for the normal density.

3. Let Y n (or Y for simplicity) be b (n, p). Thus, Y / n is approximately N [p, p (1 -p) / n]. Statisticians often look for functions of statistics whose variances do not depend upon the parameter. Can you find a function, say u(Y / n), whose variance is essentially free of p?

can anyone help me with them?