[Rasch] Rasch Difficulty Error

Stephanou, Andrew Stephanou at acer.edu.au
Fri Jul 2 20:56:23 EST 2010

I believe you know that the error on an ability estimate is greatest for low and high scores on a test, that is for people for whom the test is either too easy or too difficult.
The same can't be said for the most difficult or easiest item in a test.  If most of the information of the sample of people providing the data is available for the most difficult item in a test, then the error on the location of this item is not going to be the greatest.  The error on the location of an item depends on the position of the distribution of people providing the data relative to the location of the item.  Imagine a distribution of people centred on the most difficult item in a test, that is a test that is too easy for this group.  The error on this item is not going to be as large as the error on the other items for which there is less information.
I hope this helps.
Andrew Stephanou


From: rasch-bounces at acer.edu.au on behalf of Joyce
Sent: Fri 2/07/2010 6:48 PM
To: Stephanou, Andrew
Cc: rasch
Subject: [Rasch] Rasch Difficulty Error


I would need some help with interpreting a Rasch result that I have obtained. 

I have an understanding that the item difficulty errors of the is reflected in the person-item map, ie, if a question is given a item difficulty error of 0.77, it should be on the top of the person-item map. However, my rasch analysis has produced a different result. The most difficult question in the person-item is not having the highest difficulty error. 

Is there an explanation for this or am I interpreting this wrongly? 



-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20100702/7ac3c201/attachment.html 

More information about the Rasch mailing list