[Rasch] Likert Scales

Mike Linacre (RMT) rmt at rasch.org
Sun Oct 23 11:24:04 EST 2005


Dear Martin:

Thank you for your questions.

For investigations like yours, take a look at papers such as 
www.rasch.org/rmt/rmt101k.htm

Investigation of the functioning of the rating scale(s) should be part of 
the validation of every instrument that uses polytomies.

OK?

Mike L.


At 09:17 AM 10/21/2005 -0700, you wrote:

>Hi,
>I am Martin Schaaf, Marriage, Family & Child Counselor.  I'm attracted to 
>this list because I've discovered that Rasch analysis helps me decide what 
>to believe, what tools to use, et cetera.
>
>I have some provocative questions:
>
>When Likart response scales were developed, they were _assumed_ to be 
>scales, as best as I can tell.  That is, "Always" was assumed to be a 
>stronger response than "Almost Always," et cetera. (At the end of my 
>message is a list of Likart response items culled from the Net.
>
>QUESTION 1:  Could Rasch analysis be used to study the Likart response 
>items for scalability?  (My suspicion is, yes, but how?)
>
>QUESTION 2:  Would the scale items change with change of context?  (My 
>suspicion is, Yes.)
>
>QUESTION 3:  Should part of test development include scaling the Likart 
>response items for each test, perhaps even each item?  (Again, my 
>suspicion is, Yes.)

Mike Linacre
Editor, Rasch Measurement Transactions
rmt at rasch.org www.rasch.org/rmt Latest RMT: 19:1 Summer






More information about the Rasch mailing list