[Rasch] Validity in Rasch

Mike Linacre rmt at rasch.org
Tue Mar 22 12:14:04 EST 2011


Thank you for your post, Juanito.

All these can be seen in Winsteps Table 3.2

Please remember that these are guidelines. They do not always apply. They 
do not apply to all datasets. Some may be counter-productive in some 
datasets. Please think carefully about the meaning of your data and your 
rating scale as you apply them.

Think of these as similar to guidelines when driving a car. "Keep to your 
side of the road."  "Drive at the speed of the traffic flow." "Obey the 
speed limit." But we can all think of situations where we disregard these 
guidelines, or where these guidelines are contradictory.

Mike L.

At 3/21/2011, you wrote:
>To: Mike Linacre <rmt at rasch.org>
>
>Dear Dr Linacre,
>
>I got the following guidelines from your paper 
>(<http://www.winsteps.com/a/linacre-optimizing-category.pdf>http://www.winsteps.com/a/linacre-optimizing-category.pdf) 
>.  Thank you for that paper.  At the right of each guideline (in italics 
>font) are my comments/questions.
>
>1. At least 10 observations of each category. - It can be tested by 
>inspection from the frequency distribution.
>2. Regular observation distribution. - It can be tested by inspection from 
>the frequency distribution.
>3. Average measures advance monotonically with category. Where can I find 
>this from the Winsteps outputs?
>4. OUTFIT mean-squares less than 2.0. - I can easily locate this.
>5. Step calibrations advance - I cant identify/locate this from the 
>Winsteps outputs.  Please guide me.
>7. Step difficulties advance by at least 1.4 logits. - I cant 
>identify/locate this from the Winsteps outputs.  Please guide me.
>
>A million thanks in advance.
>J

Mike Linacre
Editor, Rasch Measurement Transactions
rmt at rasch.org www.rasch.org/rmt/ Latest RMT:  24:4 Spring 2011
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20110321/f05256f2/attachment.html 


More information about the Rasch mailing list