Using weights in multivariate multilevel models

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
hannakim
Posts: 3
Joined: Sun May 14, 2017 1:00 pm

Using weights in multivariate multilevel models

Post by hannakim » Thu May 10, 2018 12:36 pm

Hi all,

I am trying to run two multivariate multilevel models, one with Propensity score weights (standardized inverse probability of treatment weights; SIPTWs) and the other without weights.
My data consists of two dependent variables, lnPRI_E (the continuous one) and PRI_E (the binary one).
I have merged (replicated) the weight variable SIPTW_1 to a long format within MLwiN using the Data Manipulation tool.

The problem here is that whenever I try to apply weights in the multivariate multilevel model, it shows an error sign indicating that "Input lists must have Gamma block components."
(When I set two continuous variables, the same message appears, but sometimes MLwiN shows results that are exactly the same as those without weights.)

Would there be anything that I'm missing or would it be impossible to run weighted multivariate multilevel models in MLwiN?
Many thanks in advance.
Attachments
MVMM_mixed_weight error.png
MVMM_mixed_weight error.png (392.53 KiB) Viewed 227 times

ChrisCharlton
Posts: 1039
Joined: Mon Oct 19, 2009 10:34 am

Re: Using weights in multivariate multilevel models

Post by ChrisCharlton » Mon May 14, 2018 9:32 am

I checked this with Harvey Goldstein, who is responsible for the weighting algorithm, and he thinks that you cannot use weights in this way with a bivariate model where one variable is binary. As we do not recommend using weights for discrete models in MLwiN (see http://www.bristol.ac.uk/cmm/software/s ... ights.html) this combination has not been tested.

Post Reply