Date: Fri, 7 Mar 1997 10:43:28 -0800
Reply-To: Dawn Brand <brandd@OHSU.EDU>
Sender: "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From: Dawn Brand <brandd@OHSU.EDU>
Subject: sample size & interrater reliability
I have 2 questions.
First, I have 3 raters that abstract patient information (both categorical &
interval type data). As a group they abstract on about 1500 patients a year.
On a yearly basis, I need to select a sample from this pool of 1500 patients so
that each rater can reabstract patients that one of the other raters abstracted
previously. This is so that I can do an interrater reliability analysis on this
data. What I would like to know is what is the best way for me to determine the
appropriate sample I need to select from this pool?
My second question has to do with using Kappa on the categorical data. Since I
have 3 raters and Kappa can (using spssx 6.1.4) can only compare 2 raters at a
time, would it be appropriate to calculate a Kappa for each pair of raters and
then use the average of the 3 Kappas as the final Kappa for the particular
variable of interest?