[R-sig-ME] Master's Partial Credit with lmer [was:RE: lmer]

[Bill Auty]

I approached this problem by creating a dichotomous item for each 
threshold of the rated item. For a 4-level rating (0,1,2,3,4), I created 
3 dichotomous items. A score of 0 translates to 0,0,0; 1=1,0,0; 2=1,1,0 
and 3=1,1,1. Then I combined my created items with the originally 
dichotomous items and ran models using lmer as described in the paper.

I don't know if the mathematics works out exactly, but I think this is 
logically equivalent to the Partial Credit Model. If someone see a 
theoretical reason not to do this, I'd be interested to know.

Clearly, this tactic weights the data from the rated item more than a 
dichotomous items (3x in my example), but that was appropriate for my 
purposes.


Iasonas Lamprianou wrote:
> Thank you Doran for your response. If anyone else is aware of any other R package that can run multilevel Rasch/IRT models, please respond.
> 
> Jason
> 
> Dr. Iasonas Lamprianou
> Department of Education
> The University of Manchester
> Oxford Road, Manchester M13 9PL, UK
> Tel. 0044  161 275 3485
> iasonas.lamprianou at manchester.ac.uk
> 
> 
> --- On Sat, 6/12/08, Doran, Harold <HDoran at air.org> wrote:
> 
>> From: Doran, Harold <HDoran at air.org>
>> Subject: Master's Partial Credit with lmer [was:RE: [R-sig-ME] lmer]
>> To: lamprianou at yahoo.com, r-sig-mixed-models at r-project.org
>> Date: Saturday, 6 December, 2008, 3:34 PM
>> The answer is no, the PCM cannot be run using lmer. Also, it
>> is best not to ask a new question by replying to a different
>> thread.
>>
>> Harold
>>
>>
>> -----Original Message-----
>> From: r-sig-mixed-models-bounces at r-project.org on behalf of
>> Iasonas Lamprianou
>> Sent: Fri 12/5/2008 5:10 PM
>> To: r-sig-mixed-models at r-project.org
>> Subject: Re: [R-sig-ME] lmer
>>  
>> Dear friends, 
>> does anyone know how (if) I can run a multilevel Partial
>> Credit Rasch model using lmer? I am aware of the
>> "Estimating the Multilevel Rasch Model: With the lme4
>> Package" but I think that this only refers to the
>> dichotomous Rasch case. Or, alternatively, redirect me to
>> any other free package that can handle multilevel Rasch/IRT
>> models. 
>> Thanks
>>
>> Dr. Iasonas Lamprianou
>> Department of Education
>> The University of Manchester
>> Oxford Road, Manchester M13 9PL, UK
>> Tel. 0044  161 275 3485
>> iasonas.lamprianou at manchester.ac.uk
>>
>>
>> --- On Thu, 4/12/08,
>> r-sig-mixed-models-request at r-project.org
>> <r-sig-mixed-models-request at r-project.org> wrote:
>>
>>> From: r-sig-mixed-models-request at r-project.org
>> <r-sig-mixed-models-request at r-project.org>
>>> Subject: R-sig-mixed-models Digest, Vol 24, Issue 4
>>> To: r-sig-mixed-models at r-project.org
>>> Date: Thursday, 4 December, 2008, 11:00 AM
>>> Send R-sig-mixed-models mailing list submissions to
>>> 	r-sig-mixed-models at r-project.org
>>>
>>> To subscribe or unsubscribe via the World Wide Web,
>> visit
>> 	https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>>> or, via email, send a message with subject or body
>>> 'help' to
>>> 	r-sig-mixed-models-request at r-project.org
>>>
>>> You can reach the person managing the list at
>>> 	r-sig-mixed-models-owner at r-project.org
>>>
>>> When replying, please edit your Subject line so it is
>> more
>>> specific
>>> than "Re: Contents of R-sig-mixed-models
>>> digest..."
>>>
>>>
>>> Today's Topics:
>>>
>>>    1. Logisting regression for same-different speaker
>>>       classification (Leonardo LANCIA)
>>>
>>>
>>>
>> ----------------------------------------------------------------------
>>> Message: 1
>>> Date: Wed, 3 Dec 2008 12:03:02 +0100 (CET)
>>> From: Leonardo LANCIA
>>> <Leonardo.Lancia at univ-provence.fr>
>>> Subject: [R-sig-ME] Logisting regression for
>> same-different
>>> speaker
>>> 	classification
>>> To: r-sig-mixed-models at r-project.org
>>> Message-ID:
>>>
>> <9555723.429.1228302182606.JavaMail.root at frontal1>
>>> Content-Type: text/plain; charset=iso-8859-1
>>>
>>> Dear List,
>>>
>>> I would like to use a mixed logistic regression model
>> as a
>>> classifier which decides if two speech signals
>> representing
>>> two istances of the same phoneme (uttered in a
>> specified
>>> phentic context) are produced by the same speaker or
>> not. To
>>> do that I should use a huge number of predictors (more
>> or
>>> less 50 acoustic features). More over, for each
>> acoustic
>>> feature I should specify a random interaction with the
>>> following factors : the phonetic label attached to the
>>> acoustic signals, and a phonetic label correspoding to
>> the
>>> context from which the acoustic signals are extracted.
>>> I am not interested in hypotesis testing but I would
>> like
>>> to have an estimation of the contributoin to this task
>> of
>>> each of the predictors and an estimate of the
>> correction
>>> coefficients associated to the random effects.
>>> Do you think that a mixed logistic regression would do
>> the
>>> job or should I move to Support vector machines
>> algorithms?
>>> Leonardo Lancia
>>>
>>>
>>>
>>> ------------------------------
>>>
>>> _______________________________________________
>>> R-sig-mixed-models mailing list
>>> R-sig-mixed-models at r-project.org
>>>
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>>>
>>> End of R-sig-mixed-models Digest, Vol 24, Issue 4
>>> *************************************************
>>
>>
>>
>> _______________________________________________
>> R-sig-mixed-models at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
> 
> 
> 
> 
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

-- 
Bill Auty
Education Measurement Consulting
1126 NW 29th St Corvallis, OR 97330
(541) 207-3632 (office) (541) 760-2148 (cell)