R help - Jun 2008 - dgamma in WinBUGS and JAGS (rjags)

Hello,

In WinBUGS 1.4 manual
(http://www.mrc-bsu.cam.ac.uk/bugs/winbugs/manual14.pdf), the gamma
density is presented as dgamma(r,mu) where r and mu are the shape and
rate parameters, respectively. In JAGS (rjags) manual version 1.0.2,
May 9, 2008 (http://www-fis.iarc.fr/~martyn/software/jags/jags_user_manual.pdf),
on page 26 the gamma density is presented as dgamma(mu,r) instead of
dgamma(r,mu). Since JAGS uses the BUGS language, I was wondering if
this is a typo or it is in fact different from WinBUGS.


Thanks,
Reza

Seyed Reza Jafarzadeh <srjafarzadeh <at> gmail.com> writes:
> 
> Hello,
> 
> In WinBUGS 1.4 manual
> (http://www.mrc-bsu.cam.ac.uk/bugs/winbugs/manual14.pdf), the gamma
> density is presented as dgamma(r,mu) where r and mu are the shape and
> rate parameters, respectively. In JAGS (rjags) manual version 1.0.2,
> May 9, 2008
(http://www-fis.iarc.fr/~martyn/software/jags/jags_user_manual.pdf),
> on page 26 the gamma density is presented as dgamma(mu,r) instead of
> dgamma(r,mu). Since JAGS uses the BUGS language, I was wondering if
> this is a typo or it is in fact different from WinBUGS.
> 
  I don't know, but I would suggest (1) trying a couple of little
simulations to try it out for yourself -- for example with
the following BUGS code

mu <- 1
r <- 1
x ~ dgamma(mu,r)

 and see what the distribution looks like. 

and (2) e-mailing Martyn Plummer, the author (since it seems
likely that this is a documentation bug, he probably would
like to hear from you)

  Ben Bolker