www.cmm.bristol.ac.uk/forum

Hi John,
With only 6 faculties you would be hard pressed to estimate random effects - I guess you may get enough departments to treat as random. If you decide to fit them as fixed effects then simply putting in department effects will saturate the model at that level and you will not be able to add ...

Hi,
So there are various schools of thought here. Many people use weights that sum to 1 and I think in my MMMC paper (Browne, Goldstein and Rasbash, 2001) we did that. We have also argued (Tranmer, Steele and Browne, 2014) that it might make sense to make the squares of the weights sum to 1 as then ...

Hi JeanSebastien,
This seems more of a concept question rather than an MLwiN question. I think if you give more information about your data then it might be easier to answer and I am assuming the question is related to the earlier one you sent.
Best wishes,
Bill.

Hi JS,
You may need to give more details here as to what type of multilevel model you are fitting. Generally one would expect one row per response so it would depend on what level of data the responses were measured on - if this is dyads then you might have a multiple membership structure with ...

Hi Hakimeh,
Not sure I fully follow your research but I'll try and answer - your diagrams are a bit weird as I would have just put parent1, parent 2 etc. and then you would use whether they are addicted or not as a covariate - otherwise it looks like you only have 2 parents. I was unclear of what ...

Hi Johannes,
I am not sure that the group size differences will make much difference here - I guess you just have more information about within group variability in some groups than others. Your smallest group is still pretty big compared to many situations.
Best wishes,
Bill.

Hi Sabine,
That is really hard to say without knowing what you are hoping to test. Statistical power links to a test and will depend on what that test is and what effect size you are looking for / degree of clustering etc. There should be details in MLPowSim that will help.
Best wishes,
Bill.

Hi Liam,
If you register on the online LEMMA course there is stuff about this in unit 7 section 7.2 which is worth a read.
Best wishes,
Bill.

Hi Liam,
It is worth noting that for non-Normal multilevel models one has to be careful when interpreting higher level variance terms. The bottom level variance is always fixed by the distributional assumptions made and so when terms are added to the model that fixed variance would normally reduce ...

Hi Hakimeh,
The challenge you have is so much missing data. For any 1 place lets say that 2 people say it is beautiful and 1 picturesque out of 100 people. What you don't know is the opinions of the other 97 as you only ask them to choose 'an example' so they may all think the place in question is ...