www.cmm.bristol.ac.uk/forum

The model of my research is cross-classified multiple membership (CCMM). The model I study is 3level, 1-level is time, 2-level is individual, and 3-level is school (group). That is, at 3-level, individuals belong to several groups. I have now constructed this with the following formula.

In order to construct this expression with R2MLwiN code, I wrote the following again
1) Y~1+time+(1+time|group1)+(1+time|id)+(1|group2)+(1|time)
2) Y~1+time+(1+time|group1)+(1+time|id)+(time|group2)+(1|time)

But I would like to ask you three questions about the three parts
1. Should I be (1 | group2)? Or (time | group2)?
2. The model I study is 3level, 1-level is time, 2-level is individual, and 3-level is school (group). In other words, the 1-level error term can be regarded as the random effect of time, but also is used as the time variable. In R2MLwiN, can be the error term put into (1 | time)?
3. As I said earlier, my research is cross-classified multiple membership. The cross-classified is judged to have two options: XC = TRUE in estoptions and multiple membership in drop.data = FLASE and mm.
How do you configure cross-classified multiple membership options?

For an introduction to fitting these types of models I would recommend reading through the MCMC manual chapter 16 replication materials at http://www.bristol.ac.uk/cmm/software/r2mlwin/examples/, along with the corresponding text in chapter 16 of the MCMC guide (http://www.bristol.ac.uk/cmm/media/soft ... mc-web.pdf).