We want to set up a model to look at the effect of intervention in an intervention and a control group (2 timepoints). We think it’s best to use an LME, especially since we have missing scans from some subjects on one of either timepoints.
If I understand correctly, adding ranEff ‘~1’ adds subject as a random effect to the model. What I dont understand is how this works if the baseline data from this subject is missing, could you tell me something about that? Is it correct that the followup scan is then still being used in the analysis?
What I dont understand is how this works if the baseline data from this subject is missing, could you tell me something about that?
Unlike the conventional GLM or AN(C)OVA, the solution under the LME framework is obtained through optimizing the likelihood function which does not rely on the completeness of the data. As long as those subjects missed some scans at random (e.g., the missing was so not associated with the task), it would have much impact in terms of solving the LME system.
Is it correct that the followup scan is then still being used in the analysis?
Yes, that's right.
The
National Institute of Mental Health (NIMH) is part of the National Institutes of
Health (NIH), a component of the U.S. Department of Health and Human
Services.