News Focus
News Focus
icon url

hirogen

07/15/13 6:46 AM

#164039 RE: DewDiligence #164028

XOMA-

There are two things going on here: i) the VH scale itself (the matter you addressed above); and ii) what measurements are reported in the two abstracts you posted. The 87.7% number in the first abstract you posted is measuring something quite different from the 0.53 (and 0.75) concordance numbers in the second abstract.



Yes I understand. What I was trying to say was the 87.7% number would probably be different if they had baselined the test using the SUN/NEI scale. Using the coarser scale would likely lead to greater contribution of between/within graders to variance and consequently the k value from the first abstract (.91) would not show such a dramatic improvement relative to the 2nd abstract .53/.75 numbers.