[R-sig-ME] Rasch with lme4

[Doran, Harold]

There is another issue that I don't think has been raised yet. Assume
the fixed effects model is longitudinal in the sense that it contains
multiple scores per student and has students and schools as factors. It
turns out that the school effects are then identified only on the basis
of those students that move between schools.

Below is a link to a paper that discusses this in the context of
organizational studies. It is not the thesis of the paper, but it is
noted and discussed. I have a paper, which is more really notes on a
page jotted down, with all of the mathemtcail reasons why this would
occur. 

"Using Stata for a memory-saving fixed-effects estimation of the
three-way error-components model"
http://ideas.repec.org/p/boc/dsug08/07.html

> -----Original Message-----
> From: r-sig-mixed-models-bounces at r-project.org 
> [mailto:r-sig-mixed-models-bounces at r-project.org] On Behalf 
> Of Andy Fugard
> Sent: Tuesday, June 09, 2009 1:21 PM
> To: Chuck Cleland
> Cc: R-SIG-Mixed-Models at R-project.org Models
> Subject: Re: [R-sig-ME] Rasch with lme4
> 
> Chuck Cleland wrote:
> > 
> > Allison, P.D. (2005).  Fixed Effects Regression Methods for 
> > Longitudinal Data Using SAS. Cary, NC: SAS Institute.
> > 
> >   covers some of this territory.  As someone else pointed out, a 
> > limitation is that including "Subject" as a factor 
> precludes inclusion 
> > of specific subject explanatory variables (e.g., gender).
> 
> Many thanks - shall track that down.  Though I haven't yet 
> been able to replicate the limitation.
> 
> A
> 
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list 
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>