

Multilevel mixture loglinear model 

Message/Author 


Hi, In a multilevel model, "%WITHIN%" AND "%BETWEEN%" sections appear before "MODEL CONSTRAINTS". If a new parameter is specified in "MODEL CONSTRAINT", how can I make it a between level effect? In the example below, (L1_0,L1_11,L2_11) are the actual parameters to be estimated. For some modeling reasons, they can't appear directly in the MODEL section, but have to be reparameterized as a function of some other parameters in the model. If I want some parameters, say L1_11 and L2_11, to vary between clusters, how do I do that? MODEL %WITHIN% %c#1% [x1$1] (t1_1); [x2$1] (t1_2); ... %BETWEEN% MODEL CONSTRAINT: NEW(L1_0 L1_11 L2_11); t1_1=L1_0+L1_11; t1_2=L1_0+L2_11; 


You would need to use between parameters to define new between parameters. 


Dear Mplus team, I am wondering if we can run a twolevel loglinear model in Mplus. If yes could you please direct me to an example from the user's guide? Thank you in advance. 


Have a look at the singlelevel loglinear analysis of ex 7.15 which is based on mixture variables that are observed. Then combine that with the twolevel mixture analysis of ex 10.6. 

Back to top 

