When fitting mixed response models, it seems challenging to end up at the right random effects. Investigate if it's possible to solve this issue by fitting models separately first, and then providing the random effects from these separate models as initial values.
When fitting mixed response models, it seems challenging to end up at the right random effects. Investigate if it's possible to solve this issue by fitting models separately first, and then providing the random effects from these separate models as initial values.