sign in sign up
<<
Home , Impact factor, Scientometrics (journal)

Journal of Taibah University Medical Sciences (2013) 8(2), 69–71

Taibah University Journal of Taibah University Medical Sciences

www.sciencedirect.com

Editorial Article

Bandwagon of for journal

Salman Yousuf Guraya, FRCS

Department of Surgery, College of Medicine Taibah University, Almadinah Almunawwarah, Kingdom of Saudi Arabia

Received 1 April 2013; revised 5 April 2013; accepted 10 April 2013

Introduction manipulate IFs. Looking into these crucial issues, I have de- scribed the merits and demerits of IF with details of misappli- Bibliographic literature has developed a diverse range of tools cation and manipulation techniques. in evaluating research mainly based on various ways of count- 1 What is journal impact factor and how it is calculated? ing . In to prioritize the choice of quality infor- mation sources, researchers and scientists are in need of reliable decision aids. The ‘‘impact factor’’ (IF) is the most The field of scholarly publications has witnessed enormous 7 commonly used aid for deciding which journals growth in the recent past. To quantify the readership, journal’s should receive a scholarly submission or attention from re- IF is used as a universally agreed metric. First introduced by search readership. It is a journal-focused indicator that quan- Garfield, IF is defined as the number of citations within a given tifies the readership a journal attracts. It does not necessarily year to items published by a journal in the preceding two years indicate quality, but high impact factors indicate a probability divided by the number of citable items published by the journal 8 of high quality. As an arithmetic mean of data originating in these two years. It is the average number of citations a paper from all authors of a journal with a high variance, it is inappli- of a journal attracts in the two years following its publication. cable to evaluate individual scientists.2 Even recent , Thomson Reuters’ Web of Scienceâ (WoS) has been such as the h-index3 and the g-index,4 that attempt to measure the standard automated tool to identify article counts both the scientific productivity and apparent impact of a re- and conduct . and subscrip- searcher depend on the researcher’s citation record over time. tion-based are other citation databases available since 9 The IF has extensively penetrated academia and academic 2004. Thomson Reuters’ (JCR) is , which has provoked significant modifications in an annual subscription-based online citation data report for publishing strategies by the academic publishers and editors5 any journal included in its Web of Scienceâ database, which in- and in authors’ publishing behavior.6 Creating a higher num- cludes the (SCI) and the Social Science ber of mutually referenced papers from the same body of evi- Citation Index (SSCI). Thomson Reuter’s WoS currently covers dence, timing publications to have maximum exposure for almost 12,000 active journals and over 3,000 proceedings vol- 10 accruing citations, and increasing the number of citation- umes. This has raised from 8,684 titles in 2000, but it is still only attracting review papers has become a common practice to a third of the scientific serials listed in Ulrichswebä which is incomplete itself. The JCR draws on citation reports of more than 8000 journals from more than 3300 publishers in over 66 Corresponding author: Professor of Surgery, College of Medicine countries.11 Citations are compiled annually, and each unique Taibah University, Almadinah Almunawwarah, Kingdom of Saudi article-to-article link is counted as a citation. Arabia. E-mail: [email protected], [email protected]. under responsibility of Taibah University. IF has been shown to have the following pitfalls;

(a) Time window of impact factor A relatively short time window of two years is used to calculate IF although Production and hosting by there are many disciplines in which citations to their

1658-3612 Ó 2013 Taibah University. Production and hosting by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jtumed.2013.04.001 70 S.Y. Guraya

papers do not reach a peak during this short time. The on basic science resulting in three to five times more cita- IF of a journal is determined by highly cited papers tions of basic medicine than its clinical counterpart. because the distribution of citations of papers is always Consequently, basic science journals have a higher IF skewed. Olson, by using a five-year window, concluded than clinical science journals which does not reflect the that majority of articles in his study were cited for the real essence of scientific citedness of the research. first time three years after publication.12 Another study Reviews are more likely to be cited than original reported citations of 31% and 69% for two-year and research papers. Journals publishing a substantial num- five-year windows, respectively.13 Thus the mean IF of ber of review articles consequently attract more citations journal using five-year IFs are logically substantially and thus are likely to achieve a higher IF.23 greater than the corresponding means for two-year IFs. (g) Retracted articles Invalid articles such as retracted ones (b) Self citations and active manipulations of impact factor may still continue to be cited by other researchers as a Citations are viewed as the ‘currency’ of modern science valid work. Steen24 examined all retractions from MED- and their analysis has become increasingly important for LINE from 1966 through 1997 and found that many journal editors, authors and readers.14 Authors are often papers still cited retracted papers as valid research long tempted to liberally and inappropriately cite their own after the retraction notice. Consequently, retracted and previous publications in an attempt to raise their scien- invalid articles pose significant in the calculation tific among the researcher community. On the of IF of the journals. same note, in 1997 Leukemia had been accused of forc- ing the authors to cite more articles from Leukemia.15 Alternative journal scientometrics (c) Coverage and English- preference by the SCI data The SCI covers less than one fourth of peer reviewed jour- Based on the outlined issues in calculating the IF of the nals worldwide, and exhibits a preference for English lan- journals, a growing body of researchers has suggested vari- 16 guage journals. Most of the publications included in ous remedies; Asai found that more accurate statistics could WoS databases until recently had been from English-lan- be calculated if the period count is based on months rather guage journals. This resulted in severe gaps of WoS dat- than a year.25 Accordingly, he proposed an Adjusted Impact abases in comparison with other databases for citations Factor to count a weighted sum of citations per month over 17 of papers in non-English-language journals. The accu- a time period of four years. Gla¨nzel and Schoepflin reported racy of how citations are collected at ISI significantly influ- that three-year citation window was a good compromise be- ences the final IF and statistics. A fact-finding tween fast growing disciplines and slowly theories.26 by suggested a significant undercount of Hirst introduced the Disciplinary Impact Factor (DIF) to ‘‘citable’’ items in Nature Genetics in1996andanerrone- overcome the subject bias.27 It is based on the average num- ous inclusion of ‘‘citable’’ items other than those defined ber of times a journal was cited in a sub-field rather than 18 by ISI itself for Nature in 2000. the entire SCI database. For the assessment of individual (d) Impact factor is an arithmetic measure of the journal, authors, author-focused metrics may be employed which can’t predict quality of articles The majority of Nature are calculated on the basis of citations of only the author published articles from year 2002 to 2003 received under to be evaluated. For almost every letter of the alphabet, a 20 citations in 2004; 2.7% of the papers received over citation based index has been proposed. Of those a-, b-, c-, 19 100 citations with a record holder of 522 citations. In d-, e-, f-, g-, h-, j-, k-, L-, m-, n-, p-, q-, r-, t-, u-, v-, w-, x-, 2009, a single paper attracting 5,624 citations pushed y-, and z-indices, some of them admittedly very new, only the IF of A up from under 3 to the h-index has gained a widespread use.28 49.93, with all other papers of the journal having attracted three or less citations.20 Such variation renders H index and its applications attempts to use IF for the evaluation of single papers or authors absurd. (e) Impact factor is an incomplete journal-focused metric In 2005 Hirsch proposed that a scientist has index h if h of his Currently, researchers and publishers are using the or her Np papers have at least h citations each and the other growing databases WoS, SciVerseR Scopus, Google (NpÀh) papers have 6h citations each. It is probably the most Scholar, which are not comprehensive in contents and simple, author-focused index, defined as the number of papers computed data. The extent of incompleteness can vary of an author with citation number Ph. To get a higher h index, 29 mainly depending on, among others, discipline, location an individual needs at least 2h + 1 extra citations. For exam- and language of the scientist.21 This reflects a big flaw in ple, to increase the index from 4 to 5, at least 9 citations are calculating the citations for IF, necessitating the need needed. The higher the h index the more citations are needed for more comprehensive scientometrics for evaluating to increase it. It means that the difference between higher h in- the journal’s scientific strength. dex values (25 and 26, for example) is much greater than be- (f) Subject areas and categories of articles Articles in rapidly tween lower values (4 and 5, for example). The h index can growing disciplines and subjects are cited more often be employed to measure the research output of scientific insti- 30 31 than more traditional research fields, in particular theo- tutions and countries. Its only disadvantage is for younger retical and mathematical areas.22 This diversity leads to scientists with lower publication numbers, but it is at least the wide variance of IFs across subject categories and based on the author’s publications. Since it can easily be adversely affects the underrepresented fields of research. manipulated by unethical self-citations, Schreiber has rightly A given research field is often additionally cited by suggested to exclude self-citations from its calculations and 32 related fields. Hence, clinical medicine draws heavily use ‘‘the honest h index (hh)’’. Bandwagon of impact factor for journal scientometrics 71

Conclusion 14. Smith DR. Impact factors, scientometrics and the history of citation-based research. Scientometrics 2012; 92(2): 419–427. 15. Smith R. Journal accused of manipulating impact factor. BMJ The IF cannot assess the quality of individual articles, due to 1997; 314(7079): 461. the qualitative variety of citations derived and computed from 16. Moed HF, Burger W, Frankfort J, Van Raan AF. The use of a journal. For the evaluation of individual researchers, jour- bibliometric data for the of university research nal-focused metrics are inapplicable. Author-focused metrics, performance. Res Policy 1985; 14(3): 131–149. such as the h index, are to be used. 17. Sangwal K. Citation and impact factor distributions of scientific Due to major flaws and inaccuracies of IF, the researchers journals published in individual countries. J Informetrics 2013; and the publishers have to seek a more reliable and accurate 7(2): 487–504. measure of journal scientometrics. Currently, despite signifi- 18. Adam D. Citation analysis: the counting house. Nature 2002; 415(6873): 726–729. cant criticism against IF, it represents the most popular quan- 19. Campbell P. Escape from the impact factor. Sci Environ titative metrics of the journal. Politics 2008; 8(1): 5–7. 20. Dimitrov JD, Kaveri SV, Bayry J. Metrics: journal’s impact factor References skewed by a single paper. Nature 2010; 466(7303): 179. 21. Van Raan AF, Van Leeuwen TN, Visser MS. Severe language 1. Garfield E. The history and meaning of the journal impact factor. effect in university rankings: particularly Germany and France are JAMA 2006; 295(1): 90–93. wronged in citation-based rankings. Scientometrics 2011; 88(2): 2. Krell F-T. The journal impact factor as a performance indicator. 495–498. Eur Sci Editing 2012; 38: 3–5. 22. Seglen PO. Why the impact factor of journals should not be used 3. Hirsch JE. An index to quantify an individual’s scientific research for evaluating research. BMJ 1997; 314(7079): 498. output. Proc Natl Acad Sci U S A 2005; 102(46): 16569. 23. Dong P, Loh M, Mondry A. The’’ impact factor’’ revisited. 4. Egghe L. Theory and practise of the g-index. Scientometrics 2006; Biomed Digit Libr 2005; 2(7): 1–8. 69(1): 131–152. 24. Steen RG. Retractions in the scientific literature: is the incidence of 5. Brown H. How impact factors changed medical publishing––and research fraud increasing? J Med Ethics 2011; 37(4): 249–253. science. BMJ 2007; 334(7593): 561. 25. Asai I. Adjusted age distribution and its application to impact 6. Lawrence PA. The mismeasurement of science. Curr Biol 2007; factor and immediacy index. J Am Soc Info Sci 1981; 32(3): 17(15): R583–R585. 172–174. 7. Guraya SY. Journal of Taibah University Medical Sciences; 26. Gla¨nzel W, Schoepflin U. A bibliometric study on ageing and Journey to Success. J Taibah Univ Med Sci 2011; 6(2): 57–60. reception of scientific literature. J Info Sci 1995; 21(1): 8. Garfield E. Citation indexes for science. A new dimension in 37–53. documentation through association of ideas. Int J Epidemiol 2006; 27. Hirst G. Discipline impact factors: a method for determining core 35(5): 1123–1127. journal lists. J Am Soc Info Sci 1978; 29(4): 171–172. 9. MENTIONED I. The history and meaning of the journal impact 28. Hirsch J. An index to quantify an individual’s scientific research factor 2006. output that takes into account the effect of multiple coauthorship. 10. Garfield E. The meaning of the impact factor. Revista internac- Scientometrics 2010; 85(3): 741–754. ional de psicologı´a clı´nica y de la salud. Int J Clin Health Psychol 29. Guns R, Rousseau R. Simulating growth of the h-index. J Am Soc 2003; 3(2): 363–369. Info Sci Technol 2009; 60(2): 410–417. 11. Vanclay JK. Impact factor: outdated artefact or stepping-stone to 30. Bornmann L, Marx W. The h index as a research performance journal certification? Scientometrics 2012; 92(2): 211–238. indicator. Eur Sci Editing 2011; 37(3): 77–80. 12. Olson JE. Top-25-business-school professors rate journals in 31. Csajbo´k E, Berhidi A, Vasas L, Schubert A. Hirsch-index for operations management and related fields. Interfaces 2005; 35(4): countries based on essential science indicators data. Scientometrics 323–338. 2007; 73(1): 91–117. 13. Stonebraker JS, Gil E, Kirkwood CW, Handfield RB. Impact 32. Schreiber M. Self-citation corrections for the Hirsch index. EPL factor as a metric to assess journals where OM research is (Europhysics Letters) 2007; 78(3): 30002. published. J Operations Manag 2012; 30(1): 24–43.