Let $X_1,X_2,\ldots,X_n$ a random sample of a normal population with mean $0$ and variance $σ ^2$. Consider the random variable define as $T=\frac{\bar{X}\sqrt{n}}{S}$. Which is the distribution of $T^2$?

Standardizing, $Z=\frac{\bar{X}-0}{\sqrt{\sigma^2/n}}=\frac{\bar{X}\sqrt{n}}{\sigma}$, because $\bar{X}\sim N(0,\sigma^2/n)$. Also, $Z^2\sim \chi^2_{(1)}$.

On the other hand, we have that $$\frac{(n-1)S^2}{\sigma^2} \sim \chi^2_{(n-1)}$$

How can I get to this distributions from $T^2=\left(\frac{\bar{X}\sqrt{n}}{S}\right)^2$, in order to say that it distributes $F_{(1,n-1)}$?

Note that $\bar{X}$ and $S^2$ are estimators of $\mu$ and $\sigma^2$, respectively.

  • $\begingroup$ It is a good idea to define $\bar X$ and $S^2$. $\endgroup$ – StubbornAtom Oct 2 '18 at 19:00

For $(X_1,X_2,\ldots,X_n)$ i.i.d $N(0,\sigma^2)$ and $$\bar X=\frac{1}{n}\sum_{i=1}^nX_i\quad,\quad S^2=\frac{1}{n-1}\sum_{i=1}^n(X_i-\bar X)^2$$

, we know the exact distribution of $T$, namely

$$T=\frac{\sqrt{n}\bar X}{S}\sim t_{n-1}$$, a $t$ distribution with $n-1$ degrees of freedom.

This is because $$T=\frac{\sqrt{n}\bar X/\sigma}{\sqrt{\frac{(n-1)S^2/\sigma^2}{n-1}}}$$ is formed as the ratio of a standard normal variable and the square root of a chi-square variable, independent of the normal variable, divided by its degrees of freedom.

Now find the distribution of $T^2$.

If $f_{T}$ is the pdf of $T$, then pdf of $T^2$ would be of the form $$f_{T^2}(y)=\frac{1}{2\sqrt{y}}\left[f_T(\sqrt{y})+f_T(-\sqrt{y})\right]\mathbf1_{y>0}$$

You would find that $T^2\sim F_{(1,n-1)}$, an $F$ distribution with $n-1$ degrees of freedom.


If $U\sim \chi^2_p$ and $V\sim\chi^2_q$ and $U$ and $V$ are independent, then the ratio $$ \frac{U/p}{V/q} $$ has $F_{p,q}$ distribution. This is either the definition of the $F$ distribution, or can be proved as a theorem. Either way, your conclusion follows immediately from the additional fact that $\bar X$ and $S^2$ are independent.

  • $\begingroup$ (+1) A more direct route. $\endgroup$ – StubbornAtom Oct 2 '18 at 20:00

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.