A note on directed adjusted profile likelihoods