Skip to content

4 binary DVs, subjects nested within schools

1 message · Yvonnick Noel

#
Paul,
What you need is a Rasch model (and by no means a multinomial model).

As to a mean of fitting this model with glmer, See:

http://www.ualberta.ca/~baayen/publications/baayenDavidsonBates.pdf

Also, comparing slopes between groups is OK, but comparing slopes on a 
multivariate response vector is less trivial. Item response models are 
suited for that.

Best,

Yvonnick Noel
University of Brittany at Rennes
Department of Psychology
Rennes, France