I have to disagree that the two models produce different results.
Once the *+ error was corrected, my simulations of gamma and beta ranged
over the same range of values. At least with the small data provided, I
would argue that these are identical models that produce identical
results.
Finn Krogstad
U. Washington, Forest Engineering
http://students.washington.edu/fkrogsta
On Thu, 18 Jan 2001, Elena Moltchanova wrote:
> Thanks to Finn Krogstad for pointing out a mistake in my recent e-mail.
> Unfortunately that was in e-mail only and doesn't change the fact that the two
> programmes work differently.
>
> Some other remarks:
>
> I do have a data on ones[]<-1 in the second model. I have recorded it in the
> data part.
> x is indeed a property of a person so that it's more logical to write it as a
> 1:J array. However the data Y itself cannot be recorded as a n*J array because
> a number of rats per person varies.
>
> Thanks for your comments.
>
> Best regards,
> Elena Moltchanova
>
>
>
>
>
> Finn Krogstad wrote:
>
> > I like your problem, and would be interested in finding out what responses
> > you get. I don't have any useful observations at this point, but some
> > useless observations are
> >
> > 1. you didn't include the ones[i]<-1 line in model 1
> > 2. x[] is more properly a propery of person the vaccine was applied to, so
> > your array of x=c() is of length J not I and
> > logit(p.v[i])<-gamma+beta*x[person[i]]*epsilon[person[i]]
> > which suggests an odd relation between beta and epsilon
> >
> > OOPS I just discovered your problem.
> > your logit(p.v[i]) lines are not the same in the two models
> >
> > Changing *epsilon to +epsilon gives results within one MC error of each
> > other.
> >
> > Finn Krogstad
> > U. Washington, Forest Engineering
> > http://students.washington.edu/fkrogsta
>
> -------------------------------------------------------------------
> To mail the BUGS list, mail to [log in to unmask]
> You can search old messages at www.jiscmail.ac.uk/lists/bugs.html
> 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
>
-------------------------------------------------------------------
To mail the BUGS list, mail to [log in to unmask]
You can search old messages at www.jiscmail.ac.uk/lists/bugs.html
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
|