Standard error

Also known as: SE, s.e.

A standard error (SE) is an estimate of the standard deviation of an estimator’s sampling distribution. For the simple OLS slope,

\[\mathrm{s.e.}(\hat\beta)=\frac{s}{\sqrt{S_{xx}}},\quad s=\sqrt{\mathrm{RSS}/(n-2)}.\]

SEs drive $t$-statistics and confidence intervals. Wrong assumptions (hetero, dependence) ⇒ wrong SEs even if $\hat\beta$ looks fine.

← Back to CS Glossary