Issues with the expected information matrix of linear mixed models provided by popular statistical packages under missingness at random dropout.
Sign Up to like & getrecommendations! Published in 2023 at "Statistics in medicine"
DOI: 10.1002/sim.9754
Abstract: Likelihood-based methods ignoring missingness at random (MAR) produce consistent estimates provided that the whole likelihood model is correct. However, the expected information matrix (EIM) depends on the missingness mechanism. It has been shown that calculating… read more here.
Keywords: information matrix; naive eim; eim; mar ... See more keywords