It can often be difficult to gauge researcher productivity and impact, but these measures of effectiveness are important for academic institutions and funding sources to consider in allocating limited scientific resources and funding. Much as in the lab, where it is important for the results to be repeatable, developing an algorithm or an impartial process to appraise individual faculty research performance over multiple disciplines can deliver valuable insights for long-term strategic planning. Unfortunately, the development of such evaluation practices remains at an embryonic stage.
Several methods have been proposed to assess productivity and impact, but none can be used in isolation. Beyond assigning a number to an investigator—such as the h-index, the number of a researcher’s publications that have received at least that same number of citations, or a collaboration index, which takes into account a researcher’s relative contributions to his or her publications—there are additional sources of data that should be considered. At our institution, Memorial Sloan Kettering Cancer Center (MSKCC) in New York City, there is an emphasis on letters of recommendation received from external expert peers, funding longevity, excellence in teaching and mentoring, and the depth of a faculty member’s CV. For clinicians, additional assessments of patient load and satisfaction are also taken into consideration by our internal committees evaluating promotions and tenure. Other noted evaluation factors include the number of reviews and editorials an individual has been invited to author; frequency of appearance as first, middle, or senior author in collaborations; the number of different journals in which the researcher has published; media coverage of his or her work; and the number of published but never-cited articles.
Here we propose a new bibliometric method to assess the body of a researcher’s published work, based on relevant information collected from the Scopus database and Journal Citation Reports (JCR). This method does not require intricate programming, and it yields a ...
