0
$\begingroup$

Let $\bar{X}_n$ denote the sample mean of n iid random variables. Let $\bar{X}^*_n$ be the bootstrap sample mean. Does $\left|\mathbb{P}\left(n^{1/2}\bar{X}_n\leq x\right)-\mathbb{P}\left(n^{1/2}\bar{X}^*_n\leq x\right)\right|$ converge to $0$ in probability?

I am asking this question because I know from [Bickel, P.J. and D.A. Freedman (1981), Some asymptotic theory for the bootstrap] that the bootstrap can be used to approximate the distribution of the sample mean in the sense that $$\left|\mathbb{P}\left(n^{1/2}\left(\bar{X}_n-\mu\right)\leq x\right)-\mathbb{P}\left(n^{1/2}\left(\bar{X}^*_n-\bar{X}_n\right)\leq x\right)\right|\overset{\mathbb{P}}{\rightarrow}0$$ and I was wondering whether the centering is needed. Maybe to be able to use the Central Limit Theorem?

Thanks a lot!

$\endgroup$
  • $\begingroup$ What is $\mathbb{P}\left(n^{1/2}\bar{X}_n\leq x\right)$ actually? Is it asymptotically normal? And what can I say about $\mathbb{P}\left(n^{1/2}\bar{X}^*_n\leq x\right)$? $\endgroup$ – kelu Oct 14 '12 at 14:21
  • $\begingroup$ Keep in mind that the mean needs to exist for the distribution and so do higher order moments. $\endgroup$ – Michael Chernick Oct 15 '12 at 15:54
  • $\begingroup$ Yes, sorry, let's assume $\mu$ is finite, forgot that. But does that tell me anything more about my actual question? $\endgroup$ – kelu Oct 15 '12 at 20:20
  • 1
    $\begingroup$ Cross-posted at stats.stackexchange.com/questions/39297/…. $\endgroup$ – whuber Oct 29 '12 at 11:42

Your Answer

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

Browse other questions tagged or ask your own question.