[Rasch] Raw score mean and classical item difficulty

Roger Graves rgraves at uvic.ca
Fri Dec 9 05:53:09 EST 2011


22 is 4.8% higher than 21.
.61 is 5.1% higher than .58.
Close enough for rounding error?


On 8-Dec-11, at 2:55 AM, Anthony James wrote:

> Dear all,
> There are two tests each having 25 items. Each test is taken by 50  
> persons.
> The mean of classical item difficulties for the two tests (computed  
> within their group) is almost the same (.61 &. 58).
> The raw score mean of the two groups are different by one point     
> (21 & 22 ).
> Can these data be real?
> Is it possible for two groups to have almost the same mean  
> proportion of correctly answering the items (item difficulty) but  
> be different in means by one point out of a maximum of 25?
>
> Cheers
> Anthony
>
> _______________________________________________
> Rasch mailing list
> Rasch at acer.edu.au
> Unsubscribe: https://mailinglist.acer.edu.au/mailman/options/rasch/ 
> rgraves%40uvic.ca




More information about the Rasch mailing list