Hello
I would be grateful to receive some relevant advice on calculating a confidence
interval for an average Kappa statistic, where this average has been obtained
over multiple simple Cohen Kappa statistic each of which are related. (The
relatedness arises from the fact that the individual Kappas are evaluated for
specific categories of segments of a human blood vessel. Each individual Kappa
has therefore been obtained for a specific category of segment but the patient
sample is the same in each case.) The intention is to obtain an agreement
statistic over the patient sample for the whole vessel and therefore I am
attempting to find an appropriate way to pool the Kappa values rather than treat
the data as one grand sample from the outset and ignore the relatedness. I am
not convinced that a simple average is best here either but I was unable to find
a suitable alternative in the literature. Help in this area would also be
valued.
Many thanks
Best wishes
Margaret
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|