We investigate the impact of choosing between four data sources on the relative performance measurement of scholars’ research output. Our results show that rankings can differ considerably and that the observed correlations are often even lower than those reported in previous studies. We show...