One fairly simple way of analytically illustrating overdispersion is to assume some type of variability in the binomial parameter π. Let p be the random variable replacing the known binomialparameter, where p has a distribution with mean µ and variance ϕ > 0. Then, if Y is the binomial random variable, show that Var(Y ) = n2ϕ nµ − ϕnµ2 > nµ(1 − µ).