[Rasch] FA/MRM Vs. LLTM in test validation

Purya Baghaei puryabaghaei at gmail.com
Thu Oct 23 21:14:02 EST 2014

Dear colleagues,

I have a question about the cognitive processes which are studied in test
validation using factor analysis (FA) and multidimensional Rasch model
(MRM) on one side and those investigated by the linear logistic test model
(LLTM) on the other.

For understanding cognitive processes underlying test performance we can
use FA or MRM. The other less commonly used methods is LLTM. I think there
is one fundamental difference between the two techniques: In FA/MRM
cognitive processes which are assigned to factors are separate dimensions
of ability on which the standings of individuals can vary. That is, when FA
or MRM fit, the test is not unidimensional and dimensions of processes are
separate dimensions of ability for which we have person estimates.

In LLTM, however, unidimensionality is a prerequisite. That is, the test
must be unidimensional according the Rasch model before we can estimate
LLTM. That is, processes confirmed by LLTM are not necessarily separate
dimensions of ability. How does this affect the nature of processes
confirmed by LLTM Vs those ascertained by FA/MRM?



Purya Baghaei, Ph.D
English Department,
Islamic Azad University,
Ostad Yusofi St.
91871-Mashhad, Iran.
Phone: +98 511 6635064/5
Fax: +98 511 6634763
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20141023/b499cb18/attachment.html 

More information about the Rasch mailing list