Open nicebread opened 7 months ago
sum(scores across all papers) / sum(max across all papers)
mean ( sum(scores per paper) / sum(max per paper) )
Option 1 implies a stronger weighting of papers with a higher max score Option 2 says: "Every paper is equally weighted, regardless of the max score".
I think option 2 more reflects the general philosophy of RESQUE.
Does max per paper vary/depend on entries of the candidates (e.g., max is lower for non-empirical papers)? I also think option 2 (regardless of max score) is the better solution.
sum(scores across all papers) / sum(max across all papers)
mean ( sum(scores per paper) / sum(max per paper) )
Option 1 implies a stronger weighting of papers with a higher max score Option 2 says: "Every paper is equally weighted, regardless of the max score".
I think option 2 more reflects the general philosophy of RESQUE.