Some good advice from Finn Krogstad;
Please see previous discussion at
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0203&L=bugs&P=R239&I=-3
before sending Finn your replies.
Ken Rice
Bugs list manager
---------- Forwarded message ----------
Date: Mon, 27 Jan 2003 10:51:45 -0800
From: Finn Krogstad <[log in to unmask]>
To: [log in to unmask]
Subject: RE: Bayesian estimators using remission duration of 6-MP
Dear Bugsies,
Much as I hate to restart the old argument in which one side says,
"Bayesian MCMC is dangerous. Users should start with simple models and
gradually build up to their model of interest," and the other side says,
"You're a bunch of elitist pigs!" But should we really be using the
discussion list to give people just enough code to get themselves in
trouble?
It would be one thing if they say, "I have been playing with the 'Mice:
Weibull regression in censored survival analysis' example, and it
differs from my problem in the following ways... How do I make these
revisions?" I suspect most of us have been at presentation in which a
non-Bayesian has presented the results of "the Bayesian" analysis,
without acknowledging or even understanding what they are bringing to
the analysis.
I am all in favor of helping newbies into BUGS, and I regularly assist
with questions posted on this list. But I am concerned about giving
code to users, which even if it exactly represents their model, an
inexperience user can still fail to identify non-convergence (how many
of us can say we haven't done that?) or not even know what they are
supposed to be looking for. I know all (or at least some) of the
mistakes I have made over the years, and I know how easy it is to make
them. The only real protection is to have made enough mistakes on the
easy stuff, so that you know what mistakes look like on the tough stuff.
I think that the BUGS guys have done a great job moderating this list
over the years, but I really think that questions of form "can you send
me the code for this sort of model?" should not get onto the list,
because we are such a kind hearted bunch that one of us might make the
mistake of giving code to someone who is too inexperience to write, run,
or understand the results of the code.
Finn Krogstad
University of Washington
College of Forest Resources
Seattle, WA 98195
PS So as to skip past the flaming process, I will stipulate that I am
an elitist pig, so we can skip that step an move on to actually deciding
what to do.
> -----Original Message-----
> From: (The BUGS software mailing list)
> [mailto:[log in to unmask]] On Behalf Of Lotfi khribi
> Sent: Monday, January 27, 2003 2:52 AM
> To: [log in to unmask]
> Subject: Bayesian estimators using remission duration of 6-MP
>
> Dear BUGS users,
> I would like now if any BUGS user have a program for
> example 6.4 page 181 (Remission Duration from a Clinical
> Trial for Acute Leukemia it's a clinical trial of a drug
> 6-mercaptopurine (6-MP)in 42 children with acute
> leukemia) and the way that we have the table 6.5 page 182 in
> the book "Survival Analysis Techniques for Censored and
> Truncated Data". and if It's possible that you send me this
> program (Bayesian estimation of the survival function by the
> Gibbs sampler), cause I need this to make a comparison with
> other methods. Thank you and have a nice day.
> e-mail: [log in to unmask]
-------------------------------------------------------------------
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
|