The h_? index was recently proposed to measure the degree of scientific leadership. While the concept is useful and interesting, namely as a complement to the traditional performance analysis, the metric suffers from important shortcomings. We argue that scientific leadership should be evaluated: (i) taking into account information of the moment the paper is produced/published; (ii) in the specific context of the paper, meaning that only previous work relevant for the paper should be taken into account. Based on these two principles, we introduce an alternative approach, using self-citations as a source of information, which eliminates the shortcomings inherent to the h_? index. The new measures proposed in this study can be used to complement the traditional performance assessment, namely through the application of the h-index.

--

• Ciências da Computação e da Informação - Ciências Naturais
• Ciências da Comunicação - Ciências Sociais
• Outras Ciências Sociais - Ciências Sociais