Let $X_1, X_2, ..., X_n$ be a random iid sample from a population with mean $\theta$.

Now I am wondering about the intuition behind $E(X_1| \overline X ) = \overline X$, the sample mean.

If we just consider $X_1$ (or any $X_i$ for that matter) we have that $E(X_1) = \theta$ as the expected value of any random observation from a population will be the population mean.

Now given that we know $\overline X$ how come that changes what we expect to get for $X_1$? $X_1$ is still the same random observation from the population as before...but it seems that knowing the sample mean 'overrides' what we expect to get for an observation...is this correct?

  • 2
    $\begingroup$ Hint: What is $E(X_1+X_2+\cdots+X_n\,|\,\bar X)$? How do $E(X_i\,|\,\bar X)$ and $E(X_j\,|\,\bar X)$ differ? (I suspect your intuition is really concerning $E(X_i\,|\,X_j,\bar X)$.) $\endgroup$
    – whuber
    Jun 27 '14 at 18:57
  • $\begingroup$ stats.stackexchange.com/q/374989/119261 $\endgroup$ Feb 12 at 16:26

Using whuber's hint:

$\begin{align*} \left.\text{E}\left[ \bar{X} \right| \bar{X}\right] &= \bar{X}\\ \left.\text{E}\left[\dfrac{1}{n}\sum_{i=1}^n X_i \right|\bar{X}\right] &= \bar{X}\\ \dfrac{1}{n}\sum_{i=1}^n \text{E}[ X_i |\bar{X}] &=\bar{X} \\ \dfrac{1}{n}n\text{E}[ X_1 |\bar{X}] &= \bar{X}\\ \text{E}[ X_1 |\bar{X}] &= \bar{X} \end{align*}$

  • 7
    $\begingroup$ I guess the hint was meant for the OP. $\endgroup$ Jun 27 '14 at 21:04

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.