Hi all,
I'm running some four level models where, for the odd main effect, MLwIN returns an estimate of the SE as "-1.#IND000". This doesn't happen when using robust SEs. Any one know why this happens and if it can be resolved?
Thanks
Search found 2 matches
- Mon Jun 18, 2012 4:58 pm
- Forum: MLwiN user forum
- Topic: Problem computing SEs
- Replies: 0
- Views: 5562
- Mon Feb 14, 2011 3:54 pm
- Forum: MLwiN user forum
- Topic: Level incorrectly defined
- Replies: 1
- Views: 9741
Level incorrectly defined
I am trying to analyse a dataset with four levels - time within people within flats within residential blocks. I have around 550 flats within 13 different residential blocks, but MLWin is not defining the residential block level correctly. There is only a range of 1-13 but it has defined it as havin...