Suppose that we have $X_1, X_2, \ldots$ is a sequence of i.i.d random variables with $E(X_i)<+\infty$ and $N$ is a random variable taking values in $\{1,2,\ldots\}$, $N$ is independent with $X_1, X_2, \ldots$.

I know that $$E\left(\sum_{i=1}^N X_i \right) = E(N)E(X_1)$$

If I divide the sum by $N$, $$M = \frac 1N \sum_{i=1}^N X_i $$ I have sample mean of $X_1, X_2,\ldots, X_N$ but $N$ is random. In this case, what is the distribution of $M$ and how to calculate $E(M)$ and $Var(M)$. Thanks.

  • $\begingroup$ The distribution of $M$ is not reachable. The technique which allows to prove the result you mention also yields $E(M)=E(X)$ and $\mathrm{var}(M)=E(1/N)\mathrm{var}(X)$. $\endgroup$
    – Did
    Jun 15, 2014 at 20:27
  • $\begingroup$ @ Did, heropup: Thank you for your comment and answer. Could you please recommend some books about conditional expectation? $\endgroup$
    – mimi90
    Jun 15, 2014 at 20:48

1 Answer 1


The law of total expectation gives $${\rm E}[M] = {\rm E}[{\rm E}[M \mid N]] = {\rm E}[(N {\rm E}[X_1])/N] = {\rm E}[X_1],$$ as you might guess. The law of total variance gives $$\begin{align*} {\rm Var}[M] &= {\rm Var}[{\rm E}[M \mid N]] + {\rm E}[{\rm Var}[M \mid N]] \\ &= {\rm Var}[{\rm E}[X_1]] + {\rm E}\bigl[\tfrac{1}{N^2} \cdot N {\rm Var}[X_1]\bigr] \\ &= 0 + {\rm Var}[X_1]{\rm E}[1/N] \\ &= {\rm Var}[X_1]{\rm E}[1/N]. \end{align*}$$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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