[R] 'R' library for Rasch modelling?

[Jonathan Baron]

On 07/12/02 11:00, Bill Bell wrote:
>
>I'm new to 'R', so forgive me if I've used the wrong term.
>
>I'd like to model some Likert-style items using Rasch methods.
>
>I've been unable to find mention of these for 'R'. Would someone happen to
>know of any?

I once fit some data using a Rasch-type model with nls in the nls
library, using commands roughly like:

  logit <- function(x) {exp(x)/(1+exp(x))}
  icc.tmp<-nls(score~C*logit((mscore-A)/B)+D)
               start=list(A=astart[i],B=.1,C=.5,D=.5),
               control=nls.control(tol=.025))

These were not Likert-type items but test items that were scored
as correct or incorrect.  The model would be more complicated
with several different cutoffs, and I don't want to take the time
to figure out how to do it.  It is similar to Thurstone category
scaling, I assume.

I note that the data I had were very messy, and the model fit
poorly and often failed to converge at all.  If I did this again
I would either make sure I had better data or else I would use
optim() (in the base package) instead of nls, or both.

Note also that various Rasch models allow you to specify
various parameters, such as an upper bound, lower bound,
midpoint, and slope.

-- 
Jonathan Baron, Professor of Psychology, University of Pennsylvania
Home page:            http://www.sas.upenn.edu/~baron
R page:               http://finzi.psych.upenn.edu/
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._