[rkward-devel] item response theory plugin

[Stefan Rödiger]

Am Donnerstag 14 August 2008 22:04:46 schrieb meik michalke:
> hi,
>
> i've been planning this for a long time, now i have finally started working
> on a plugin for data analysis unter the item response theory approach :-)
>
> i've implemented parameter estimation for dichotomous data last night, to
> fit the rasch, two parameter logistic and birnbaum three parameter model.
> the plugin relies on the ltm package:
>  o http://wiki.r-project.org/rwiki/doku.php?id=packages:cran:ltm
> almost all options of the rasch(), ltm() and tpm() functions are available,
> only the "parscale" control of tpm() and the two factor model of ltm() are
> missing, as of now.
>
> my future plans are to develop additional dialogs for goodness-of-fit tests
> and ltm specific plot options (e.g. test information curves). but of course
> any suggestions are wellcome :-)
>

Hello meik,

thanks for the plugin.

> here's the plugin:
>  o http://reaktanz.de/stuff/R/irt_plugin-0.01.tar.bz2
> and some screenshots:
>  o http://reaktanz.de/stuff/R/scrsht3_01_data_model.png
>  o http://reaktanz.de/stuff/R/scrsht3_02_options.png
>  o http://reaktanz.de/stuff/R/scrsht3_03_advanced_control.png
> there's also a quick&dirty wizard, but i'm not sure if it's really
> useful...
>
I think it will take some time until your plugin is reviewed, at least we have to wait for a final 
comment by Thomas.

> since the plugin needs the ltm package, i've included a simple library(ltm)
> call in the preprocess function. is there a better way to make sure it's
> loaded, except doing it manually?
>

<?
function preprocess () { ?>
require (ltm)
...

is commonly used in RKWard plug-ins.

>
> viele gruesse :: m.eik

viele gruesse
Stefan