Compare the variances of $X^2$, $X$ and $\sqrt{X}$.
Is there a way to effectively compare sizes using the variance equation below?
$$Var()=[(−[])^2] = [^2]−([])^2$$
Compare the variances of $X^2$, $X$ and $\sqrt{X}$.
Is there a way to effectively compare sizes using the variance equation below?
$$Var()=[(−[])^2] = [^2]−([])^2$$
Copyright © 2021 JogjaFile Inc.
How they compare will depend upon the distribution. Let's focus on the variances of $X^2$ and $X$. Suppose, for instance,
$$X=B+a,B\sim Bern(p),a\in\mathbb{R}$$
so $$X^2=(1+2a)B+a^2.$$
Then $$\text{Var}(X^2)=(1+2a)^2\text{Var}(X),$$
which is less than $\text{Var}(X)$ for $a\in (-1,0)$ but is otherwise at least as great as $\text{Var}(X).$