Date: Wed, 7 Sep 2005 19:37:22 +0100
Reply-To: starborn@statisticsdoc.com
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Keith Starborn <starborn@statisticsdoc.com>
Subject: Re: Test - Retest
In-Reply-To: <B14120EE5C432443B21102F7925DAD020266B31F@COKE.uwec.edu>
Content-Type: text/plain; charset=ISO-8859-1; format="flowed"
Keith Starborn
www.statisticsdoc.com
Dear Jessica,
A few suggestions - before conducting a test-retest analysis of the
stability of
the scores over time, it would be good to first establish at least the
following conditions:
1. At each time point, the scales have high internal consistency,
2. At each time point, the scales have a similar dimensional structure (the
suggestion to use SEM to look at the invariance of factor structure is an
excellent one),
3. There is reason to believe, theoretically, that the scores should be
relatively stable over time. What is being measured? Some constructs are
changeable for good reason. Who is being surveyed? Some subjects may be
exposed to conditions that are intended to differentially change scores on the
construct.
To the extent that these conditions are met, an analysis of the correlation
between scores over time is more likely to be meaningful.
HTH,
KS
Quoting "Franson, Jessica Leigh" <fransonj@UWEC.EDU>:
> I'm working with someone who recently conducted some survey research.
> He surveyed the same group of people twice over a certain period of time
> and wants to test the validity of the results by comparing both data
> sets. What's the best way to go about this?
>
> Thanks,
> Jessica Franson
>
>
> Jessica Franson
> Note: New email is fransonj@uwec.edu <mailto:fransonj@uwec.edu>
> Instructional Technologist/Statistical Software Consultant
> Network for Excellence in Teaching
> Learning &Technology Services
> University of Wisconsin-Eau Claire
> (p) 715.836.4734
> (f) 715.836.6001
> http://www.uwec.edu/net <http://www.uwec.edu/net>
>
>
>
--
For personalized and experienced statistical consulting,
visit www.statisticsdoc.com
