# Lecture 15. B) Some Implications

• If one has access to $I\left(\theta_{0}\right)^{-1}$, or one can approximate it via some estimator say $\widehat{I\left(\theta_{0}\right)}$ and/or $\widehat{I\left(\theta_{ML}\right)}$, one can use the normal distribution for hypothesis testing as long as $n$ is large.