Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

How would we go about finding the expectation and variance of the r$^{\text{th}}$ order statistic $X_{(r)}$ from a random sample $X_1, \ldots, X_n$ from a uniform distribution with density function

$$f(x)=1 \quad \quad \text{for} \ x \in (0,1)$$

and $f(x)=0$ otherwise?

Intuitively we can see the expectation is $\frac{r}{n+1}$ though I'm keen to see how that is derived from simply knowing the pdf of the distribution.

share|cite|improve this question
up vote 5 down vote accepted

Suppose you know that $$ a\Gamma(a) = \Gamma(a+1) \tag{1} $$ and that $$ \int_0^1 x^{\alpha-1} (1-x)^{\beta-1}\,dx = \frac{\Gamma(\alpha)\Gamma(\beta)}{\Gamma(\alpha+\beta)}. \tag{2} $$ From $(2)$, we get the probability density function of the Beta distribution with parameters $\alpha$ and $\beta$: $$ \int_0^1 f(x)\,dx = \int_0^1 \frac{\Gamma(\alpha+\beta)}{\Gamma(\alpha)\Gamma(\beta)} x^{\alpha-1}(1-x)^{\beta-1}\,dx=1. $$ Now we want the first and second moments $\mathbb{E}(X)$ and $\mathbb{E}(X^2)$ of a random variable $X$ with this density. $$ \begin{align} \mathbb{E}(X) & = \int_0^1 x f(x)\, dx = \frac{\Gamma(\alpha+\beta)}{\Gamma(\alpha)\Gamma(\beta)} \int_0^1 x \cdot x^{\alpha-1} (1-x)^{\beta-1} \, dx \\ \\ & = \frac{\Gamma(\alpha+\beta)}{\Gamma(\alpha)\Gamma(\beta)} \int_0^1 x^{(\alpha+1)-1} (1-x)^{\beta-1} \, dx. \tag{3} \end{align} $$ The integral identity in $(2)$ holds if $\alpha$ is any number at all; therefore it holds for $\alpha+1$: $$ \int_0^1 x^{(\alpha+1)-1} (1-x)^{\beta-1} \, dx = \frac{\Gamma(\alpha+1)\Gamma(\beta)}{\Gamma((\alpha+1)+\beta)}. $$ It follows that the product in $(3)$ is $$ \frac{\Gamma(\alpha+\beta)}{\Gamma(\alpha)\Gamma(\beta)} \cdot \frac{\Gamma(\alpha+1)\Gamma(\beta)}{\Gamma((\alpha+1)+\beta)} $$

So $\Gamma(\beta)$ cancels, and $(1)$ can be applied to change $\displaystyle\frac{\Gamma(\alpha+1)}{\Gamma(\alpha)}$ to $\alpha$, and to change $\displaystyle\frac{\Gamma(\alpha+\beta)}{\Gamma((\alpha+1)+\beta)}$ to $\displaystyle\frac{1}{\alpha+\beta}$. Therefore the expression in $(3)$ simplifies to $$ \frac{\alpha}{\alpha+\beta}. $$ A similar technique differing only in details finds $\mathbb{E}(X^2)$.

Once you have $\mathbb{E}(X)$ and $\mathbb{E}(X^2)$, you can use $\operatorname{var}(X) = \mathbb{E}(X^2) - (\mathbb{E}(X))^2$. There's some algebraic simplifying to be done after that. Remember that the variance should be symmetric in $\alpha$ and $\beta$, so if what you get is not symmetric, there's a mistake.

share|cite|improve this answer
Thanks! This is exactly what I needed. – Mathmo Jan 30 '12 at 10:49

See . The Beta distribution with parameters $r$ and $n+1-r$ has mean $\frac{r}{n+1}$ and variance ${\frac {r\, \left( n+1-r \right) }{ \left( n+1 \right) ^{2} \left( n+2 \right) }}$. Obtain these from the fact that $\int_0^1 t^a (1-t)^b\ dt = B(a+1,b+1) = \frac{a!b!}{(a+b+1)!}$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.