Date: Sun, 14 Jul 1996 10:52:44 -0400
Reply-To: "David A. Rowe" <darowe@COE.UGA.EDU>
Sender: "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From: "David A. Rowe" <darowe@COE.UGA.EDU>
Subject: Re: intra-class correlation
In-Reply-To: <31E64675.5E0@psychologie.uni-kassel.de>
Martin,
It depends upon which intraclass correlation (ICC) you want. The
ICC is actually a family of correlation coefficients, one of which is the
Cronbach's alpha provided in the Reliabilities program of SPSS. This
particular one is an ICC based upon a 2-way ANOVA model, which does not
include mean differences between trials (items, whatever). You probably
need to know which kind of ICC you wish to calculate before your question
can be answered. I hesitate to give you the procedures for any particular
one, when it may not be appropriate, and I don't really have the time to
provide all of the procedures. Maybe you could start by reading
articles by Berk, Algina, Bartko, and Feldt before making your decision.
Of course, if you wish to dsecribe in a little more detail your specific
research situatiion, it might be somewhat easier to provide a specific
answer. If you need the full references for these articles, let me know,
and I'll send them (don't have them close at hand right now).
=============================================================================
| David A. Rowe | office: (706) 542-4786 | "A student from whom |
| University of Georgia | home: (706) 543-3277 | nothing is ever asked |
| Dept. Exercise Science | fax: (706) 542-3148 | which he cannot do, |
| Ramsey Center | email: | never achieves all that |
| 300 River Road | darowe@coe.uga.edu | he can" |
| Athens GA 30602-3654 | | - John Stuart Mill |
=============================================================================
On Fri, 12 Jul 1996, Martin Haenze wrote:
> Does anybody know how to compute intraclass correlations
> by SPSS?
>
| 
