Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

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

Suppose $X_1,\ldots,X_n$ are iid normal random variables, $X_i \sim \mathcal{N}(\theta, \sigma^2)$. The method of moments gives estimations for $\theta$ and $\sigma^2$ by solving for them in $$\bar{X}=\theta, \qquad \frac{1}{n}\sum X_i^2 = \theta^2+\sigma^2$$ Giving $$\theta = \bar{X}, \qquad \sigma^2=\frac{1}{n}\sum(X_i-\bar{X})^2$$

Why does this give a biased estimation of $\sigma^2$? Since it seems to have $1/n$ instead of $1/(n-1)$ in front?

share|cite|improve this question

First of all you are missing some expectations throughout the post, i.e. $E[\bar{X}]=\theta$ etc. And yes, by replacing $\frac{1}{n}$ by $\frac{1}{n-1}$ you have that $$ E\left[\frac{1}{n-1}\sum (X_i - \bar{X})^2\right] = \sigma^2 $$

share|cite|improve this answer

Denote $\hat\theta(X_1, \ldots, X_n)$ and $\hat{\sigma}^2(X_1, \ldots, X_n)$ the parameter estimators: $$ \hat\theta = \frac{1}{n} \sum_{k=1}^n X_k \qquad \hat{\sigma}^2 = \frac{1}{n} \sum_{k=1}^n X_k^2 - \left( \frac{1}{n} \sum_{k=1}^n X_k \right)^2 = \frac{1}{n}\left(1-\frac{1}{n} \right) \sum_{k=1}^n X_k^2 - \frac{2}{n^2} \sum_{k=2}^n \sum_{m=1}^{k-1} X_k X_m $$ To establish the bias, we need to compute the expectation of each estimator: $$ \mathbb{E}(\hat{\theta}) = \frac{1}{n} \sum_{k=1}^n \mathbb{E}(X_k) = \frac{1}{n} \sum_{k=1}^n \mu = \mu $$ $$ \mathbb{E}(\hat{\sigma}^2) = \frac{1}{n} \left(1-\frac{1}{n}\right) \sum_{k=1}^n \mathbb{E}(X_k^2) - \frac{2}{n^2} \sum_{k=2}^n \sum_{m=1}^{k-1} \mathbb{E}\left(X_k X_m\right) \\= \frac{1}{n} \left(1-\frac{1}{n}\right) \left( \sum_{k=1}^n (\mu^2 + \sigma^2) \right) - \frac{2}{n^2} \sum_{k=2}^n \sum_{m=1}^{k-1} \mu^2 \\= \left(1-\frac{1}{n}\right) (\mu^2+\sigma^2) - \frac{2}{n^2} \frac{n(n-1)}{2} \mu^2 \\ = \left(1-\frac{1}{n}\right) \sigma^2 = \frac{n-1}{n} \sigma^2 $$ Since $\mathbb{E}(\hat{\sigma}^2) \not= \sigma^2$, the estimator is biased.

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.