[Rasch] Unidimensionnality

squik squik psquik at gmail.com
Sun May 14 21:56:15 EST 2006


Dear Rasch members,

I hope that you could help me and thank you in advance.
I work (as a student) in a lab of psychology on a test dealing with stress.
I recently discovered the Rasch model and decided to use it to this purpose,
with Bigsteps. We decompose the stress in two components. We would like to
get three scores, one for each of the component and one for the global level
of stress.

What I would like to know is the following points :
Do I have to test the unidimensionality separately for each component and
then for all items or is it satisfactory to test it only on all items ?

Do I have to run big steps three times to get a measure for each item,
namely twice for the items of the two components and then once for all items
or is it satisfactory to run Bigsteps only once on all items ?

Secondly, before I arrived in this lab, items were calibrated in one run to
give one and only one score. Now, because I would like to get at least two
scores (one for each of the two components), I would like to know whether it
is possible to re-use a measure previously computed on all items in one and
only one run or whether I have to re-compute separately (through a one step
item banking procedure for instance) the measure for items belonging to each
component.

Sincerely and best wind to all of you.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20060514/3f003779/attachment.html 


More information about the Rasch mailing list