Questions on the change of variance

Welcome to the forum for MLwiN users. Feel free to post your question about MLwiN software here. The Centre for Multilevel Modelling take no responsibility for the accuracy of these posts, we are unable to monitor them closely. Do go ahead and post your question and thank you in advance if you find the time to post any answers!

Remember to check out our extensive software FAQs which may answer your question: http://www.bristol.ac.uk/cmm/software/s ... port-faqs/
Post Reply
KK123456
Posts: 4
Joined: Tue Feb 28, 2017 11:21 am

Questions on the change of variance

Post by KK123456 »

Dear everyone:
I have a question on the drop of variance at the higher level. I am modeling 2 two-level random slope model. I found that when I just added some independent variables into the null model, and all slopes are fixed, but the variance at level-2 (school) still drops. I was a lit bit confuse about this drop. I can understand the drop of the variance at the individual level, but how does the drop at level-2 occur?

Thanks everyone!
ChrisCharlton
Posts: 1351
Joined: Mon Oct 19, 2009 10:34 am

Re: Questions on the change of variance

Post by ChrisCharlton »

Could you please provide screenshots of simple examples where this occurs to further illustration the details of this?
KK123456
Posts: 4
Joined: Tue Feb 28, 2017 11:21 am

Re: Questions on the change of variance

Post by KK123456 »

Thanks for your reply.
Here is the example:
For the null model, the random part shows that σu0^2=0.076; σe0^2=0.248
when I added several independent variables into the null model and keeps the coefficients fixed (NOT randomly varies cross schools), the random part shows that σu0^2=0.075; σe0^2=0.228
As you can see both σu0^2 and σe0^2 dropped. My question is now I have not added level-2 variables, why the σu0^2 (level-2 variance) dropped?

Thanks
ChrisCharlton
Posts: 1351
Joined: Mon Oct 19, 2009 10:34 am

Re: Questions on the change of variance

Post by ChrisCharlton »

I asked George about this and he said the following:
This is because most level 1 variables vary in their mean value across level 2 units. Thus adding level 1 variables typically explains away variation between level 2 units as well as well as between level 1 units.
Post Reply