0
$\begingroup$

I don't know how to get the second line from the first line in the following: enter image description here

In the above case, $Y=(y_1, \dots , y_n)^T$ is a random sample from $N(\mu,\sigma^2)$.

My trouble is in simplifying $ E\left(\left\{\sum\limits_{i=1}^n (Y_i-\mu ) \right\}^2\right)$. What I've tried:

$$ \begin{align} E\left(\left\{\sum\limits_{i=1}^n (Y_i-\mu ) \right\}^2\right) & = E\left(\left\{\sum\limits_{i=1}^n (Y_i-\mu ) \right\}\right) E\left(\left\{\sum\limits_{i=1}^n (Y_i-\mu ) \right\}\right) , \text{ by independence} \\ & = \left(\sum\limits_{i=1}^n E \left( Y_i - \mu \right) \right)^2 \text{ by linearity of expectations} \end{align}$$

I don't see how I can get the variance from here on wards. Did I do something wrong?

$\endgroup$
  • 1
    $\begingroup$ $\sum(Y_i-\mu)$ is not independent from itself so the equality there is wrong. Use the fact that $(\sum (y_i-\mu))^2 = \sum(y_i-\mu)^2 +\sum_{i\neq j}{(y_i-\mu)(y_j-\mu)}$. Independence will work for the second sum. $\endgroup$ – radIQ Mar 5 '15 at 15:44
2
$\begingroup$

For simplicity, write $X_i=Y_i-\mu$, such that $\mathbb{E}[X_i]=0$. Then, $$\begin{align} \mathbb{E}\left[\left(\sum_{i=1}^n X_i\right)^2\right] &= \mathbb{E}\left[\left(\sum_{i=1}^n X_i\right)\left(\sum_{j=1}^n X_j\right)\right] = \mathbb{E}\left[\sum_{i=1}^n \sum_{j=1}^n X_i X_j\right] \\ &= \sum_{1\leq i,j \leq n} \mathbb{E}[X_i X_j] \tag{linearity}\\ &= \sum_{i=1}^n \mathbb{E}[X_i^2] + \sum_{1\leq i\neq j \leq n} \mathbb{E}[X_i]\mathbb{E}[X_j] \qquad\text{(independence when $i\neq j$)} \\ &= \sum_{i=1}^n \mathbb{E}[X_i^2] + 0 = \sum_{i=1}^n \mathbb{E}[X_i^2] = \sum_{i=1}^n \mathrm{Var} Y_i = n\mathrm{Var} Y_1 \end{align}$$

$\endgroup$
  • $\begingroup$ in the first line, last equality, shouldn't $X_j X_j$ be $X_i X_j$? $\endgroup$ – mauna Mar 5 '15 at 15:53
  • $\begingroup$ same case with the second line as well? $\endgroup$ – mauna Mar 5 '15 at 15:55
  • $\begingroup$ and the third. The three typos are gone. $\endgroup$ – Clement C. Mar 5 '15 at 16:01
2
$\begingroup$

Multiplying out $$ E\left[\left\{\sum\limits_{i=1}^n (Y_i-\mu ) \right\}^2\right] $$ You get terms like $$ E\left[(Y_i-\mu )^2\right] $$ as shown, but also terms like $$ 2E\left[(Y_i-\mu )(Y_j-\mu )\right] $$ for $i<j$. But the independence shows $$ E\left[(Y_i-\mu )(Y_j-\mu )\right] = E\left[(Y_i-\mu )\right] E\left[(Y_j-\mu )\right] $$ and $\mu$ was chosen so that these are all zero.

$\endgroup$

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.