Skip to content

Fisher Information and Standard Errors

Overview

The Fisher information measures the amount of information a random variable carries about a parameter.

Definition

\[ I(\theta) = E\left[\left(\frac{\partial}{\partial\theta} \log f(X;\theta)\right)^2\right] = -E\left[\frac{\partial^2}{\partial\theta^2} \log f(X;\theta)\right] \]

For \(n\) i.i.d. observations: \(I_n(\theta) = n \cdot I(\theta)\).

Standard Errors from Fisher Information

The asymptotic standard error of the MLE is:

\[ \text{SE}(\hat{\theta}_{MLE}) \approx \frac{1}{\sqrt{I_n(\hat{\theta})}} \]

Examples

Distribution Parameter Fisher Information
Bernoulli(\(p\)) \(p\) \(\frac{1}{p(1-p)}\)
Normal(\(\mu, \sigma^2\)) \(\mu\) \(\frac{1}{\sigma^2}\)
Poisson(\(\lambda\)) \(\lambda\) \(\frac{1}{\lambda}\)