sign in sign up
<<
Home , Impact factor

EDITORIAL EDITORIAL Impact, not impact factor

Inder M. Verma Editor-in-Chief, PNAS

When the English philosopher Herbert Spen- Reflecting that limitation, a 2005 editorial cer introduced the phrase “survival of the in reported that 89% of the journal’s fittest” in 1864, he could not have imagined impact factor of 32.2 at the time could be that it would summarize the plight of young attributed to 25% of the papers published scientists years later (1). As competition for during the relevant period (3). Moreover, as coveted faculty appointments and research Garfield himself observed, only 0.5% of the 38 funding continues to intensify, today’sre- million items cited from 1900 to 2005 were cited more than 200 times; half of the articles searchers face relentless pressure to publish received no , and a quarter were not in scientific journals with high impact factors. substantive reports of original research (2). But only a few decades ago, when I began Furthermore, the common practice of my scientific career as a virologist in the journals based on vanishingly small Inder M. Verma. 1970s, the common outlets in my field were in impact factor—sometimes down to three journals that specialized in . Work decimal points—lends the metric a semblance that straddled disciplines was often pub- of precision and discriminatory power. which an applicant’sarticleswerepublished lished in journals catering to a broad reader- Admittedly, it is unrealistic to think that reaches a threshold; failuretoreachthethresh- ship. Most researchers read and published the choice of journal in which to publish is old can influence career advancement. Some articles with little regard to the purported unimportant; exciting and important papers institutions even offer substantial monetary impact of the journals themselves. Faculty are often published in journals with high incentives to publish in journals with high appointments, promotions, and the award impact factors. But the above facts must give impact factors (there are anecdotal reports of the scientific community pause regarding the of research grants were often largely based sliding scales!). usefulness of the impact factor as an indicator on perceived future impact of the work, To be sure, the scientific community is of the quality of individual articles published not on whether the research was published overburdened with responsibilities, including in so-called “high-impact” journals. The scientific community writing and reviewing grants and articles, Unfortunately,thetidehassinceturned. must not rely exclusively teaching, and reading hundreds of applica- , founder of the Institute for tions for assistant professorships, tenure, Scientific Information, which later became on the impact factors and promotions. So it is easy to equate high- impact work with journals with high impact part of Thomson Reuters, introduced the of journals. factors. But not all papers with high impact concept of journal impact factors. Originally are or can be published in such journals. conceived to help guide librarians’ decisions in high-impact journals. How, for example, Anditisequallyimportanttobearinmind regarding journal subscriptions, this metric can a researcher serving on an appointment/ that what matters in the end is the impact of has long since been used to rank- scien- promotion/review committee be certain of the category to which an applicant’s research ar- a given body of work on the tific journals. According to Garfield, “Ajour- ticles belong: the majority of articles with low of a scientific field. At PNAS, we ask authors nal’s impact factor is based on 2 : the citations or the minority with high citations? to write a 120-word statement of significance numerator, which is the number of citations When it comes to judging the quality and of the work to indicate its impact in the in the current year to items published in the significance of a body of work, there is no field. Other institutions and funding agen- previous 2 years, and the denominator, which substitute for qualitative . And it cies are beginning to ask candidates to state is the number of substantive articles and re- the significance of their important papers, ” bears repeating that the impact factor is not views published in the same 2 years (2). That an article-level metric, nor was it intended as which should aid in evaluations. calculation has inherent limitations even a yardstick for comparing researchers’ scholarly As arbiters of the importance and merit of ’ when used to measure a journal s scientific contributions. However, at many institutions publications, the scientific community must impact. Determining impact by tallying up performance assessments hinge greatly on this not rely exclusively on the impact factors of citations garnered over a two-year period number, which currently wields outsize influ- journals, whose acceptance criteria can be can inflate a journal’s overall impact if only ence on the advancement of scientific careers. based on an array of considerations, includ- a small number of articles, especially in pro- So much so that job applications at some uni- ing and subject areas. I am gratified lific but rapidly changing areas of research, versities are not even processed until appli- that the scientific community, concerned account for the vast majority of citations. cants have published at least one paper in a about the undue influence of impact factors, (Consider the recent explosion of interest in high-impact journal with first authorship. At has begun to seriously address the issue. areas such as stem and genome other institutions tenure is granted when the The San Francisco Declaration of Research editing, for example.) combined impact factor of the journals in Assessment—a set of guidelines forged under

www.pnas.org/cgi/doi/10.1073/pnas.1509912112 PNAS | June 30, 2015 | vol. 112 | no. 26 | 7875–7876 Downloaded by guest on September 24, 2021 the auspices of the American of Cell in Washington, DC, a symposium orga- journal impact factors and curtail the com- Biology in 2012 and targeted toward re- nized by Randy Schekman, editor-in-chief munity’s overreliance on them. At PNAS, searchers, publishers, and funding agencies, of eLife, raised several of the aforemen- which is celebrating its 100th anniversary, among other stakeholders—is an encourag- tioned concerns. Sustained engagement in the focus has always been on impact, not ing step toward a reasonable evaluation of such efforts can help prevent the abuse of impact factors. scientific output (4). At conferences, conver- sations about the improper use of impact factors have begun to raise the scientific 1 Spencer H (1864) The Principles of Biology, p 444. Available 3 Anonymous (2005) Not-so-deep impact. Nature 435(7045): community’s awareness of the issue. At at https://archive.org/stream/principlesbiolo05spengoog#page/n460/ 1003–1004. mode/1up. Accessed June 1, 2015. 4 Anonymous (2012) San Francisco Declaration on Research Assessment. the 152nd annual meeting of the National 2 Garfield E (2006) The history and meaning of the journal impact ASCB Annual Meeting. Available at https://archive.org/stream/ Academy of Sciences held in April 2015 factor. JAMA 295(1):90–93. principlesbiolo05spengoog#page/n460/mode/1up. Accessed June 1, 2015.

7876 | www.pnas.org/cgi/doi/10.1073/pnas.1509912112 Verma Downloaded by guest on September 24, 2021