Considering the 'Godambe Information' G(θ)=AB−1AG(\theta) = A B^{-1} AG(θ)=AB−1A, which statement is true?
It accounts for model misspecification in the Sandwich covariance matrix estimator.
It is identical to the Fisher Information if the model is correctly specified.
It is used to calculate the power of a test.
It is independent of the sample size.