variance of the linear predictor

[shanekav]

Hi,
Apologies in advance if this should be obvious to me. How can I calculate the variance of the linear predictor for a logistic 2 level model in MLwiN?

thanks
Shane

[ChrisCharlton]

Would you be able to clarify what you are trying to calculate, as the term could be interpreted in more than one way, as well as what you hope to use it for?

[shanekav]

Hi Chris,

I am running a number of variations of a two level logistic model and would like to make comparisons between the coefficients across the models. As we discussed before, making such comparisons is complicated in logistic multilevel models because the fixed variance at the lowest level causes a rescaling effect that leads to changes in the coefficients apart from changes coming from the model.

Joop Hox in his book Multilevel Analysis (2010, 135-6) outlines a way of calculating a scale correction factor. One part of the suggested equation is the variance of the linear predictor:

‘The variance of the linear predictor is the systematic variance in the model; the other two variances are the residual errors at the two levels’

I hope this is enough information for you.

Shane

[billb]

Hi Shane,
My immediate advice is to contact Joop directly as he may have a macro to do this in MLwiN which he has been known to use. Without reading Joop's book which I don't have time to as we speak I can't help you any further.
Apologies,
Bill.

[multilevelninja]

Hi Shane, I am looking at the same matter now. Did you manage to calculate the variance of the linear predictor? I would appreciate if you could share any details on this. Thanks!