2
$\begingroup$

As stated in the title, I am trying to prove that if $X_n \Rightarrow X$ in distribution, then $a+bX_n \Rightarrow a+bX$ ( where $a,b\in\mathbb{R}$) in distribution using the definition as follows: $X_n$ converges in distribution to $X$ if $F_{X_n}(x) \Rightarrow F_X(x)$ for all $x$ such that $F_X(x)$ is continuous at $x$.

I have been able to prove this using characteristic functions, but now I am attempting it using only definitions.

My main idea is to use the definition of $F_X(x)=\{\omega:X(\omega)\leq x \}$ for $\omega\in\Omega$ (on $(\Omega,\mathcal{F},P)$, so then $F_{X_n}(x) \rightarrow F_X(x) \implies P\{\omega: X_n(\omega)\leq x\} \rightarrow P\{\omega : X(\omega) \leq x\} \Rightarrow P\{\omega: a+bX_n(\omega) \leq a+bx\} \rightarrow P\{\omega: a+bX(\omega) \leq a+bx\}$,

so then $a+bX_n\rightarrow a+bX$ in distribution. However, I am unsure of that last step; and feel like the continuity portion of the definition needs to be addressed.

$\endgroup$
4
$\begingroup$

It is probably worth labelling $Y_n=a+bX_n$ and $Y=a+bX$ and looking at the sign of $b$

  • If $b\gt 0$ then $F_{Y_n}(y) = F_{X_n}\left(\frac{y-a}{b}\right) \to F_{X}\left(\frac{y-a}{b}\right) = F_{Y}(y)$, with the continuity condition carrying across naturally

  • If $b\lt 0$ then the continuity condition becomes a little more important; essentially you are considering those parts of the distribution where $\mathbb P\left(X=\frac{y-a}{b}\right) =0$ and so $\mathbb P(Y = y)=0$ and so $\mathbb P(Y \le y) = \mathbb P(Y \lt y) = 1-\mathbb P(Y \ge y) = 1-\mathbb P(Y \gt y)$. For such $y$s, you have $F_{Y}(y) = 1-F_{X}\left(\frac{y-a}{b}\right)$ though not necessarily for $X_n$ and $Y_n$, and thus $F_{Y_n}(y) = 1 - F_{X_n}\left(\frac{y-a}{b}\right) + \mathbb P\left(X_n = \frac{y-a}{b}\right) \to 1 - F_{X}\left(\frac{y-a}{b}\right) + 0 = F_{Y}(y)$

  • If $b=0$, then $Y_n=Y=a$ surely, so trivially you have convergence of all types. In particular $F_{Y_n}(y)= F_Y(y)$ as this is $0$ when $y \lt a$ and $1$ when $y \ge a$

$\endgroup$

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.