Hi, all:
I am trying to estimate a conjunctive model with two dimensions, theta1 and
theta2. The overall response probabiliy P(i, j) is given by the product of
response probabilities from two subprocesses, EP(i,j) and MP(i,j), which is a
function of theta1 and theta2, respectively. The model is specified as the
following.
model
{
for (i in 1 : N) {
for (j in 1 : T) {
logit(ep[i,j]) <- theta1[i] -beta1;
logit(mp[i,j]) <- theta2[i] -t1*dgr[j] -t2*surf[j] -beta2;
p[i, j] <- ep[i,j]*mp[i,j];
resp[i,j] ~~ dbern(p[i,j]);
}
theta1[i] ~~ dnorm(0, 1);
theta2[i] ~~ dnorm(0, 1);
}
# Priors
beta1 ~~ dnorm(0,1.0E-4); beta2 ~~ dnorm(0,1.0E-4);
t1 ~~ dnorm(0,1.0E-4); t2 ~~ dnorm(0,1.0E-4);
}
Although the model can run, the resulting parameters suggest that sth. is
wrong. It seems to me that some indetermincy in the final parameter scale.
sometimes estimates for one dimension looks fine, but those for the other
dimension looks wild in terms of scale,which will give a near 1.0 probability
for one of the subprocess. sometimes the reverse will happen.
Does anyone have some experiencs or suggestion to fix the problem?
Thanks in advance.
Xiangdong Yang, PhD,
Center for Educational Testing and Evaluation
University of Kansas
Phone: (785)864-3337
-------------------------------------------------------------------
This list is for discussion of modelling issues and the BUGS software.
For help with crashes and error messages, first mail [log in to unmask]
To mail the BUGS list, mail to [log in to unmask]
Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html
Please do not mail attachments to the list.
To leave the BUGS list, send LEAVE BUGS to [log in to unmask]
If this fails, mail [log in to unmask], NOT the whole list
|