[Rasch] Measure from one item response

Thomas Salzberger thomas.salzberger at gmail.com
Tue Sep 25 20:07:51 EST 2012

To reiterate Richard's point:
I am currently working on single item measures chosen from a multiple
item measure with demonstrated fit.
(Not that I am advocating the widespread use of single item measures,
but given that some people use them it is interesting to investigate
how a Rasch-based selection works and how such a single item measure
operates and performs.)
A single item measure (7 point rating scale) has a minimum SE(person)
of about 1.3 in one example (other examples similar). And that's only
the case in a rather limited area, of course.
But even a four item scale does not come close to a SE of 0.5 (what
might be a decent degree of uncertainty).
If you are not really interested in individual person measures but
group means, it is possibly okay.


2012/9/24  <rsmith at jampress.org>:
> Of course the person SEM for this estimate would be very large (Imagine the SEM for a 4 item MC test.).
> Richard M. Smith, Editor
> Journal of Applied Measurement
> P.O. Box 1283
> Maple Grove, MN  55311, USA
> website:  www.jampress.org
> phone: 763-268-2282
> fax: 763-268-2782

Thomas.Salzberger at gmail.com
Thomas.Salzberger at wu.ac.at

More information about the Rasch mailing list