[R-sig-ME] Parameters and unobserved random variables - was Re: lmer: ML and REML estimation

Robert A LaBudde ral at lcfltd.com
Mon Mar 30 03:26:06 CEST 2009

At 04:25 PM 3/29/2009, Murray Jorgensen wrote:
>Rolf, I didn't ask any question about the EM algorithm, or about MML 
>for that matter, I was just indicating the context in which my query 
>first arose. To avoid any tangents to my tangent and to suppress any 
>fuzziness which Rolf and I both deplore I will repeat the question:
>_Are their any Bayesians who attempt to make a distinction between 
>parameters and unobserved random variables and if so, how and why?_
>(Possibly the wrong group to address such a question to, though I 
>could attempt to defend my choice of group if required!)

I believe the correct question to ask should be "Is there any 
distinction between parameters and latent variables in the Bayesian 

Offhand, I don't see any substantiative distinction from a 
methodology point of view. But, under some forseeable circumstances 
in some experiment, latent variables might become manifest, but 
parameters never will. Variables might be latent in one experiment, 
where they are not measured, and manifest in others, where they are.

(Reminds me of Erwin Schroedinger's cat.)

Robert A. LaBudde, PhD, PAS, Dpl. ACAFS  e-mail: ral at lcfltd.com
Least Cost Formulations, Ltd.            URL: http://lcfltd.com/
824 Timberlake Drive                     Tel: 757-467-0954
Virginia Beach, VA 23464-3239            Fax: 757-467-2947

"Vere scire est per causas scire"

More information about the R-sig-mixed-models mailing list