Let $X_1, X_2,...X_n$ be i.i.d. with $P(X_i =1)=1-P(X_i =0)=p,p \in (0,1)$. (a) Show that $\bar X_n$ is the MLE of p. (b) Find the mean $\mu_n$ and variance $\sigma^{2}_n$ of $\bar X_n$ and invoke the CLT to determine the asymptotic distribution of $\bar X_n$, properly centered and $\sqrt n$ scaled.

Thoughts: I know how to solve part (a) and also know the mean and variance of part (b). The question that bothers me is the second part of part (b). I don't know to apply CLT here with $\sqrt n$ scaled. Any explicit explanation of this part would be much appreciated.


Here is a review of some facts I suppose you already know.

$S_n = \sum_{i=1}^n X_i \sim Binom(n, p).$ Also, $E(X_i) = p\;$ and $V(X_i) = p(1-p).$ Thus, $E(S_n) = np\;$ and $SD(S_n) = \sqrt{np(1-p)}.$

If we define $Z_n = \frac{S_n - np}{\sqrt{np(1-p)}},$ then $Z_n$ is approximately standard normal. According the the CLT its distribution gets closer to standard normal as $n$ increases. I assume this is what is meant by 'centered and $\sqrt{n}$-scaled'.

Given that $\bar X = S_n/n,$ you ought to be able to re-express the fraction for $Z_n$ in the required format.

  • 1
    $\begingroup$ thanks!!! BTW, what is V? Is that the variance? If so, I think it should be $p(1-p)$. What does SD mean here? $\endgroup$ – hil316 Jan 3 '16 at 3:41
  • $\begingroup$ Right. My mistake. Fixing it. Thanks. SD=standard deviation. $\endgroup$ – BruceET Jan 3 '16 at 7:49

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.