www.cmm.bristol.ac.uk/forum

I am working on a longitudinal dataset with students nested in classrooms nested in schools.

For the longitudinal dataset, I have 5 measurements that will be modeled with a nesting in students (compare chapter 13 of the MCMC manual; http://www.bristol.ac.uk/cmm/software/m ... nuals.html)

During the assessments, the students were allowed to move to another classroom within the school. That requires multiple membership models, with students given a weight for their presence in a classroom (e.g., all waves in the same classroom versus five waves in five different classrooms; compare chapter 16 of the MCMC manual).

The challenge is to combine these modelling strategies into on model. However, I don't know how to specify such a repeated measures multiple membership model. Given that such longitudinal analyses are not common, I wonder if anyone knows about a solution?

Hi GijsHuitsing,
I am unclear whether you have multiple membership or simply cross-classification? If your structure is that each measurement is nested within 1 student and 1 class but the classes change for measures then this is simply a cross-classification and you simply tick the cross-classified button when setting up the model.
The repeated measures nature you describe simply results in a 3-level model and it is possible to have 1 of these levels as multiple membership if each measure is associated with more than 1 class. You basically need your class ids and weights set up as per chapter 16 but simply add an additional level for student and have a column with the student ids. The classification window that sets up multiple membership then asks you for each level.
Hope this helps,
Bill.

Thanks for you reply Bill.

I struggled with it, and found out that it is possible to combine multiple membership models with repeated measures.
The data were structured as followed:

ID School Class Weight WAVE Outcome
10 10000 1001 0.3333 1 5.26
10 10000 1002 0.3333 2 4.89
10 10000 1003 0.3333 3 7.56
etc.

In this way, students are allowed to move classroom between waves.

Hi GijsHuitsing,
But the issue here is there is a 1-1 mapping of outcomes to classes so probably no need to use multiple membership unless you are going to included some sort of cumulative effects for earlier class-rooms.
One uses a multiple membership model if say you only had response at wave 3 but knew the 3 classes attended and then there is a 3-1 mapping of classes to outcomes.
Best wishes,
Bill.