Dependent Variance ?

• Apr 25th 2009, 01:01 PM
Chowder610
Dependent Variance ?
I was hoping someone can show me the proof for
V(y1-Y2) = V(y1)+V(y2)- 2 Cov(Y1,Y2)
or
V(y1+Y2) = V(y1)+V(y2)+ 2 Cov(Y1,Y2)

I understand the proof for V(y1+Y2) = V(y1) + V(Y2) for independent but im having trouble inferring for the dependent.
• Apr 26th 2009, 12:00 AM
CaptainBlack
Quote:

Originally Posted by Chowder610
I was hoping someone can show me the proof for
V(y1-Y2) = V(y1)+V(y2)- 2 Cov(Y1,Y2)
or
V(y1+Y2) = V(y1)+V(y2)+ 2 Cov(Y1,Y2)

I understand the proof for V(y1+Y2) = V(y1) + V(Y2) for independent but im having trouble inferring for the dependent.

$V(y_1+y_2)=E( [(y_1+y_2)-\overline{(y_1-y_2)}]^2)$

but

$
\overline{(y_1-y_2)}=\overline{y_1} + \overline{y_2}
$

so:

$V(y_1+y_2)=E( [ (y_1-\overline{y_1}) +(y_2-\overline{y_2})]^2)$ $=E((y_1-\overline{y_1}) ^2) + 2E((y_1-\overline{y_1}) (y_2-\overline{y_2}) )+E((y_2-\overline{y_2}) ^2)$

CB
• Apr 27th 2009, 05:33 AM
Chowder610
Thank you