Multiple membership in a 3-level repeated measures setup

[drddtorres]

Okay, so I have read through chapter 19 of the user's guide and through the MLwiN Practical Module 13. None of this is very clear for how to setup a longitudinal model. I worked very hard today getting my data into MLwiN and making the -ivmlwin- macro run (mostly), but I'm wondering if my data need to be in wide format to run the multiple membership model. Anyone? Anyone? Right now the data are in long format with one row per year for each person represented in the data.

Thanks,
Diego

[ChrisCharlton]

Looking at chapter 19 it looks like the data does need to be in wide format, however it appears that multiple membership in IGLS also involves setting constraints. These would interfere with the constraints set by the ivmlwin macro, so unfortunately I don't think that you can use both of these at the same time. If you weren't using the ivmlwin procedure then we would normally recommend that multiple membership models are estimated in MCMC which is why the documentation for multiple membership with IGLS in the user guide isn't very extensive.

[drddtorres]

C'est la vie, no? Oh, well.
Diego