Relative Performance Measurement of Researchers: The Impact of Data Source Selection

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 that data source selection generally has a higher effect than does measure selection and that performance measures based on Google Scholar might provide information that complements the more traditional data sources. We find greater differences for accounting, which suggests that decision makers must pay more careful atten-tion when using rankings in this field.