If \(S^2\) is a random variable of possible sample variance values from a sample of size n, then

\begin{equation*} W = \frac{(n-1)S^2}{\sigma^2} \end{equation*}

is approximately \(\chi ^2(n-1).\)