[Rasch] Constructing Variables

Lang, William Steve WSLang at tempest.coedu.usf.edu
Wed Dec 12 05:52:31 EST 2007

Hmmmm....not a bad question, but I think of "constructing variables" with Rasch as a different (or higher) standard than the "hypothetical construct" that most of us were taught in sophomore psychology.  The notions of conjoint measurement, interval level measures, and person fit are sometimes missing in less thoughtful instrument development models.  Maybe Rasch folks (Rashies?) don't always describe a philosophy, but a philosophy is decidedly underlying the development and analysis of the instruments and scales.  Simply hitting the button for Winsteps would not suffice.

For example, if you read this week's issue of Educational Researcher with 7 articles on validity, you will see hints at their confusion about variable construction, but none of the articles mention "mapping variables" like Mark Wilson (Constructing Measures: An item Response Modeling Approach) or produce Kidmaps (Applying the Rasch Model, Bond & Fox).  Both the beginnings and endings of creating Rasch models are "different" than seemingly described by the authors in ER.  

To me, a discussion of validity without such concepts or methods is so limiting that there mostly seems to be lot of reorganizing of old ideas in ER without consideration of what I would consider Rasch basics (sorry about the soapbox comment).  If one wants confidence and understanding of the measures they use and the information produced, then Rasch models provide a method that is powerful, understandable, and practical.  I believe the Constructing Variable theme implies that different process.  I don't know what will happen in 2008, but there are certainly folks who produce measures with Rasch that are "job-related" (Assessment Teacher Competency: Five Steps to Valid Measurement using the CAATS Model, Wilkerson & Lang, 2007, Corwin Press).

Steve Lang
University of South Florida St. Petersburg

-----Original Message-----
From: rasch-bounces at acer.edu.au on behalf of Snider-Lotz, Tom
Sent: Tue 12/11/2007 10:01 AM
To: Rasch Mailing List (rasch at acer.edu.au)
Subject: FW: [Rasch] Constructing Variables
Hello all --

I'm a bit surprised at the lack of response to my question (below) about constructing variables.  I'd thought this would be an easy one!  In place of a lengthy answer, I'd be grateful for a reference that I could follow up.

Thanks, all.

 -- Tom

From: rasch-bounces at acer.edu.au [mailto:rasch-bounces at acer.edu.au] On Behalf Of Snider-Lotz, Tom
Sent: Friday, December 07, 2007 11:55 AM
To: Rasch Mailing List (rasch at acer.edu.au)
Subject: [Rasch] Constructing Variables

The IOMW conference next March looks very interesting, but I must admit that I don't understand the theme, "Constructing Variables."  I know this is a common phrase in the Rasch world, and has been the theme for several issues of the Journal of Applied Measurement -- but still I don't know quite what it means.  What is the difference between a constructed variable and one that's not constructed?  What's involved in constructing a variable?  If I run test results through Winsteps, are the item measures constructed variables?

I work in employment selection testing, and when I think of variables I think of test scores, supervisory ratings of performance, number of words typed per minute, number of units sold, and so forth.  Would these be included in the domain of variables to be discussed at this conference, and are any of them "constructed"?

Thanks for any clarification you might offer.

  -- Tom Snider

-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailinglist.acer.edu.au/pipermail/rasch/attachments/20071211/5e9b0612/attachment.html 

More information about the Rasch mailing list