The difference of results between MLwiN and HLM

[lihu15402]

Dear Mr. Chris,
I hope to know how much is the difference between the results of HLM and MLwiN? Should the results given by these 2 tools be largely same?

[ChrisCharlton]

In most cases I would expect the results to be very similar for the same models. For a comparison of multilevel software packages see http://www.bristol.ac.uk/cmm/learning/mmsoftware/.

[lihu15402]

Dear Mr. Chris,
Thanks a lot! Using MLwiN, I couldn't get similar results of HLM. I can't figure out what wrong I have made. May I have your email address? If I can send my dataset, model and results to you, it is easier for you to answer my problem. My email address is lihu18@sina.com.

[ChrisCharlton]

You can find my contact details on the following page: http://www.bris.ac.uk/cmm/team/, however I don't have HLM so can only tell you if there is something obviously wrong with the way your model is set up in MLwiN. A common issue with MLwiN is that the data has to be sorted according to the model hierarchy in order to correctly estimate the model. You can check that this is the case with your model by checking the Model-Hierarchy viewer window.

[lihu15402]

Dear Mr. Chris,
I have sent worksheet files to your email box, thanks a lot!
Hu