[Rasch] Misfitting persons

Stuart Luppescu lupp at uchicago.edu
Mon Dec 4 01:52:43 AEDT 2017

On Sun, 2017-12-03 at 21:02 +1100, Nicholas Reynolds wrote:
> ·         The scale has four reverse worded items, all of which are
> misfitting

This is a very common phenomenon. There are two principal causes:
1) People do not carefully read the items and do not notice the
negative valence, and
2) The reversed scale does not have the same threshold values as the
non-reversed scale. That is, for example, "agree reversed" does not
have the same meaning as "disagree".

I have found that putting the reversed items in their own response
group somewhat mitigates problem 2). It does nothing for 1), which is
why I try to avoid negatively worded items whenever possible.

Stuart Luppescu
Chief Psychometrician (ret.)
UChicago Consortium on School Research

More information about the Rasch mailing list