[Rasch] Facet with one examinee

Parisa Daftari Fard pdaftaryfard at yahoo.com
Mon Aug 8 03:26:49 EST 2011


Dear Rasch Member,
 
Facet is usually used to search for the possible bias in the scores generated through multiple judgments by different raters or judges. In such situations, there are several examinees, some raters, and some scores. Of course some other nominal variables might be examined as intervening variables. 
 
However, recently bias is discussed when approaching latent trait and competence. For example, whether native speakers approach specific speech act similarly or might be biased towards the appropriate response because of their cultural or social background. In such situations, We have Raters (40 for instance) but only one paper; I mean different raters judge a single situation on likert scale from 1 to 5 to see how much the response to such situation is appropriate. This means that Raters are examinees. Now I wonder how would you think that we need to approach the data 
1.  I have a three facet model one Rater (many raters), Question types (2types), Items (1-3) [considering the fact that the raters are the examinees]
1, 1, 2, 2, 3 
1, 2, 2, 2, 3 
2, 1, 2, 2, 3 
2, 2, 2, 2, 3
.
.
. 
20, 1, 2, 2, 3 
20, 2, 2, 2, 3 
OR
2. I consider question type as dummy variable and have the data as follows 
1, 2, 2, 3, 2, 2, 3 
2, 2, 2, 3, 2, 2, 3 
3, 2, 2, 3, 2, 2, 3 
.
.
.
20, 2, 2, 3, 2, 2, 3 
 
As the situation is new to Facet, I should highly appreciate your help and comments beforehand.
 
Best,
Parisa
 
Parisa Daftarifard
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20110807/74d67e02/attachment.html 


More information about the Rasch mailing list