Dear colleagues,
I've been looking at the computation of the intraclass correlation
coefficient in randomised studies comparing two administrations (paper vs
electronic) of a quality of life questionaire in two periods of n subjects.
I was hoping that I might be able to ask a couple of questions?
(1) In publications, I've not seen one yet that accounts for the period
effect in the estimation of the ICC. Does anyone know of one which does and
is there a
textbook or methods article that recommends this as the approach? I would
think this is the best approach as it increases precision?
(2) Does anyone know of a site with SAS code to do this analysis?
Many thanks for any advice.
kind regards
Malcolm Morrissey
statistician
_________________________________________________________________
Windows Live™ Messenger has arrived. Click here to download it for free!
http://imagine-msn.com/messenger/launch80/?locale=en-gb
|