Displaying 1 result from an estimated 1 matches for "mps_home".
Did you mean:
mpi_home
2005 Apr 27
0
Fitting a kind of Proportional Odds Modell using nlme, polr, lrm or ordgee
...j1}+....+p_{ijk}= Q_{ijk}
and Q_{ijm}=1, Q_{ij1}=1.
i = 1,...,r (r=5) #number of studies
j = 1,....,n_{i} #subject j from study i
k = 1,...,m (m=5)#Category
My problem is how to fit this model in R although I have a SAS code from
Whitehead which is availabe at
http://www.rdg.ac.uk/mps/mps_home/misc/publications.htm . There it is
fit by using Proc NLMIXED. The result for beta is there:
The problem is that alpha_{ik} represents the k_th intercept for the
i_th study. A "normal" Proportional-Odds model
has only alpha_k which represents the k-th intercept without any
associato...