Good morning I would apreciate any suggestion on the following: For a sample of 1400 compositions, each rated on a scale 0-15 by two raters randomly chosen from a pool of 120, would you say that ICC, Pearson's, or Cohen's k (after categorising the grades) would be an apporopriate method for checking interrater reliability? The current layout is: rows-->subjects, column1-->raterA, column2-->raterB

Many thanks Vassilis