I need some help with a homework assignment. The question I'm given is: "Suppose that $X_1, X_2,..., X_n, W$ are independent random variables such that $X_i\sim Bin(1,0.4)$ and $P(W=i)=1/n$ for $i=1, 2 ,.., n$. Let

$Y=\sum\limits_{i=1}^W X_i=X_1+X_2+X_3+...+X_W$

That is, $Y$ is the sum of $W$ independent Bernoulli random variables. Calculate the mean and variance of $Y$"

Since $Y$ is a sum of Bernoulli random variables it would be a binomial random variable with mean $\mu=np$ and variance $\sigma^2=np(1-p)$, but I'm not sure how to handle this problem when $n$ itself is a random variable.

Can anyone show me how to approach this?


1 Answer 1


Using iterated expectations and law of total variance here is the approach I would use as recommended by @Glen_b in the comments above. To use these you need the means and variances of the individual random variables.

$$\mathbb{E}(X_1) =0.4$$ $$\mathbb{V}ar(X_1) =0.4\times0.6=0.24$$

and for $W$,

$$\mathbb{E}(W)=\frac{1}{n}\sum_{i=1}^ni =\frac{n(n+1)}{2n}=\frac{n+1}{2}$$ $$\mathbb{V}ar(W)=\frac{1}{n}\sum_{i=1}^ni^2 -\frac{(n+1)^2}{4} =\frac{n(n+1)(2n+1)}{2n}-\frac{(n+1)^2}{4} =\frac{n+1}{4}\times(4n+2-(n+1)) =\frac{(n+1)\times(3n+1)}{4}. $$

Now conditioning on $W$ and using total variance law,

$$ \mathbb{V}ar(Y)=\mathbb{V}ar(\mathbb{E}(Y|W))+\mathbb{E}(\mathbb{V}ar(Y|W)) =\mathbb{V}ar(W\mathbb{E}(X_1))+\mathbb{E}(W\mathbb{V}ar(X_1)) =\mathbb{V}ar(W)\mathbb{E}(X_1)^2+\mathbb{E}(W)\mathbb{V}ar(X_1). $$

Now we can use the expressions above to substitute into the last equality to get

$$\mathbb{V}ar(Y)=\mathbb{V}ar(W)\mathbb{E}(X_1)^2+\mathbb{E}(W)\mathbb{V}ar(X_1) =\frac{(n+1)\times(3n+1)}{4} \times \frac{4}{10} + \frac{n+1}{2}\times\frac{24}{100} =\frac{(n+1)\times(3n+1)}{10} + \frac{12\times(n+1)}{100} $$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.