We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.DL

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Computer Science > Digital Libraries

Title: Non-cumulative measures of researcher citation impact

Abstract: The most commonly used publication metrics for individual researchers are the the total number of publications, the total number of citations, and Hirsch's $h$-index. Each of these is cumulative, and hence increases throughout a researcher's career, making it less suitable for evaluation of junior researchers or assessing recent impact. Most other author-level measures in the literature share this cumulative property. By contrast, we aim to study non-cumulative measures that answer the question "in terms of citation impact, what have you done lately?"
We single out six measures from the rather sparse literature, including Hirsch's $m$-index, a time-scaled version of the $h$-index. We introduce new measures based on the idea of "citation acceleration". After presenting several axioms for non-cumulative measures, we conclude that one of our new measures has much better theoretical justification. We present a small-scale study of its performance on real data and conclude that it shows substantial promise for future use.
Comments: 11 pages; journal version is open access and has link to data
Subjects: Digital Libraries (cs.DL)
Journal reference: Quantitative Science Studies (2020) 1 (3): 1309-1320
DOI: 10.1162/qss_a_00074
Cite as: arXiv:2105.09354 [cs.DL]
  (or arXiv:2105.09354v1 [cs.DL] for this version)

Submission history

From: Mark C. Wilson [view email]
[v1] Wed, 19 May 2021 18:45:34 GMT (58kb,D)

Link back to: arXiv, form interface, contact.