On the interplay between normalisation, bias, and performance of paper impact metrics

Marcel Dunaiski, Jaco Geldenhuys, Willem Visser