Richard,
It appears that you have RATER and EQUIPMENT crossed and TIME nested within
specific combinations of RATER and EQUIPMENT, so you do not need to break
this problem into separate analyses for RATER and for EQUIPMENT effects. If
you analyze both RATER and EQUIPMENT in the same design, you can examine
some interesting interaction effects, such as the possibility that
measurements by one piece of equipment are associated with higher
inter-rater agreement as well as higher within-rater temporal stability.
One can imagine a scenario in which one piece of equipment yields
information that is not only rated more accurately but is also rated with
higher levels of agreement between raters and more temporal stability.
Best Regards,
Stephen Brand, Ph.D.
www.StatisticsDoc.com
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of
Shoemaker, Richard G Dr CIV DOD CAPMED FBCH
Sent: Wednesday, January 04, 2012 3:01 PM
To:
[hidden email]
Subject: Set up variables for analysis
I am confused about the best way to set up this data for
validity/reliability analysis: subjects are 52 teeth set individually in an
artificial mandible (26 have a fracture and 26 are intact controls), an
image of each tooth is made with two different types of equipment (call them
T1 and T2), a software program produces the images on a screen and three
different raters look at all 104 images in random order and rate them as: 1)
definite fracture, 2) indefinite, 3) definite intact tooth. Then at 8 weeks
the raters are provided the same images, again assorted randomly and asked
to rate them. Can this be set up and analyzed in a mixed or factorial repeat
measures design? Or must this be broken down to two separate problems: rater
reliability and equipment validity?
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Locked
