Show that the variance of the sum $x + y$ of the random variable $x$ and the random variable $y$ is the sum of the variance $\sigma_x^2$ of $x$ and the variance $\sigma_y^2$ of $y$?

My attempt:

Is my first time doing statistics and there are alot of terms i am quite unfamiliar with.

Random variable: Value of a variable result from a random experiment: e.g. (the score when a die is rolled once)

Variance($\sigma^2$) : Average of squared deviations from the mean or square of standard deviation?

I know variance of sum $x + y$ means $var(x+y) = $E(x+y)^2 -$(E(x+y))^2$ But what does sum of the variance $\sigma_x^2$ of $x$ and the variance $\sigma_y^2$ of $y$ means?

Edited: My 2nd attempt: Var(x + y ) = E$(x+y)^2$ - $(E(x+y)^2)$ = E($x^2$ + 2xy +$y^2$) - [($Ex)^2$ + 2ExEy + ($Ey)^2$ ] = [E$x^2$ - (E$x)^2$ ] + [E$y^2$ - (E$y)^2$ ] + 2[Exy - ExEy] = Var x + Var y + 2cov (x,y)

Now i assume(not sure is it correct) sum of the variance $\sigma_x^2$ of $x$ and the variance $\sigma_y^2$ of $y$

= var ($\sigma_x^2$ ) + var($\sigma_y^2$) = E($\sigma_x^2$ ) -[E($\sigma_x)]^2$ + E($\sigma_y^2$ ) -[E($\sigma_y)]^2$

Since var(x) = $\sigma_x^2$,

=E(E($x^2)$ - $[E(x)]^2$) - [$E^2$ ((E($x^2)$ - $[E(x)]^2$)) + E(E($y^2)$ - $[E(y)]^2$) - [$E^2$ ((E($y^2)$

I did not further solve it because i realise is wrong as the above does not have an expression that has an xy expression. So where have i done wrong?

  • $\begingroup$ It means $E((X+Y)^2-(E(X+Y))^2$ instead. $\endgroup$ – Michael Hoppe Jan 30 at 13:49
  • $\begingroup$ @MichaelHoppe sorry, just now i made a mistake about variance of sum x + y. Anyways what you said i believe is referring to var(x+y) ? How about the case for the sum of variance σx^2 of x and the variance σy^2 of y? What is the expression? $\endgroup$ – john Jan 30 at 13:56
  • $\begingroup$ Possible duplicate of Finding the mean and variance of a linear combination of independent random variables $\endgroup$ – Did Jan 30 at 17:16

The variance of $X$ is $E\left[\big(X - E[X]\big)^2\right]$ so here we are looking at

$$E\left[\big((X+Y) - E[X+Y]\big)^2\right] \\= E\left[\big((X-E[X])+(Y-E[Y])\big)^2\right] \\ = E\left[\big(X - E[X]\big)^2\right] +E\left[\big(Y - E[Y]\big)^2\right] +2E\left[\big(X - E[X]\big)\big(Y - E[Y]\big)\right] $$

You actually want the result to be the two left terms. The right-hand term cancels out if and only if the covariance is $0$, which will happen when $X$ and $Y$ are independent or in a few other special cases


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.