Inferential Statisticshard
0:00.0

Which condition is necessary for the Godambe Information G(θ)=AB1AG(\theta) = A B^{-1} A to be the correct asymptotic variance for an M-estimator θ^\hat{\theta} where the model is potentially misspecified?