Bibliometrics
Bibliometrics (Greek: biblion "book" and métron "measure") is the teaching of the measurement (metric) of scientific publications .
Mathematical and statistical methods are used to measure publications such as books, articles and magazines. Bibliometrics, for example, creates ranking lists that show which scientists or universities publish how many articles in a year. It is a sub-discipline of scientometry (measurement of the sciences) that measures general information about the scientific community, such as the number of university graduates or the countries of origin of Nobel Prize winners.
Bibliometrics does not make any statements about the quality of scientific publications, but deals exclusively with their quantities. In doing so, she tries to determine, for example, the empirical regularities in the growth in the number of publications of books, the distribution of topics across specialist journals and the number of citations of a journal article ( citation analysis ). Bibliometric analyzes are usually produced in written tabular form by their own departments within libraries or universities. The data for such analyzes are mainly collected from specialist databases and citation databases and made available for a fee.
Similar sub- disciplines of scientometry are patentometry (measurement of publications that are patents ) and webometry (measurement of data from the World Wide Web ). Since 2016, graphic elements in publications have been systematically examined using new methods, which the developers call " Viziometrics ".
Bibliometric indicators
Citation analysis
The most important measured variables are the citation rates. They are created by means of citation analyzes, in which it is determined how often a scientific publication has been cited by other publications. In addition, it also shows how often the entirety of publications by a specific scientist or scientists employed at a specific research institution (e.g. at a university) have been cited. Based on citation analyzes, rankings can be drawn up in which the most cited publications, scientists and groups of scientists (e.g. scientists at the University of Vienna compared to those at the University of Graz ) are determined.
- Journal Impact Factor
The various journals are also compared on the basis of the citation frequency. The Journal Impact Factor is a calculated number that provides information about how often the articles in a particular scientific journal are cited. It can be calculated in different ways and on the basis of different raw data. The best known is the impact factor identified in the Journal Citation Reports . In 2012, the medical journal CA - A Cancer Journal for Clinicians was the front runner with 154 points , and the journal Science took 20th place with 31 points.
- h index
The h-index quantifies the scientific influence of an author based on the number of his publications and the number of times they have been cited. All publications by a person are numbered in descending order in the order they are cited. The Hirsch factor is then the last value in the list for which the number of citations is greater than or equal to the number assigned in the sequence.
Alternative metrics
As an alternative to traditional citation analysis, legacy metrics measure impact on the web.
Citation databases
To determine citation frequencies, citation databases operated especially for this purpose are queried. These are filled with data that come from the evaluation of a different large part of the overall scientific publication. The two most influential citation databases are the Web of Science (from Thomson Reuters ) and Scopus (from Elsevier ), to which access can be bought with high license fees. Freely accessible citation databases are for example Eigenfactor , CiteSeer , Research Papers in Economics or Citebase .
application
One area of application of bibliometrics is the quantitative evaluation of scientists and academic institutions based on their publications.
The numbers and regularities ascertained using bibliometrics are of practical importance , particularly for libraries (see also German library statistics ).
In addition to the measurability of scientific outputs, bibliometrics enables topic areas to be observed (see citation analysis): Which topics are currently being discussed scientifically, in which area is a lot published?
Institutions and journals
The most important bibliometric specialist journal is Scientometrics , founded in 1978 , which annually awards the most important award, the "Derek J. de Solla Price Medal", first awarded in 1984. An international conference has been held every two years since 1987 and is organized by the International Society for Scientometrics and Informetrics (ISSI), which was founded in 1993 as part of the conference .
history
Scientific papers, studies and investigations that have a bibliometric character can, depending on the definition, be identified as early as the 12th century in the form of Jewish indices. The term bibliometrics itself was introduced in 1969 by Alan Pritchard . As early as 1934, however, the term bibliométrie was used by Paul Otlet , the founder of documentation . The name librametrics , proposed by SR Ranganathan in 1948, failed to gain acceptance. One of the first important results of bibliometrics was the establishment of a power law in the frequency of the use of individual words, as it was discovered by Estoup in 1916 and formulated by George Kingsley Zipf in 1935 and later developed by George Udny Yule , Benoît Mandelbrot and others. The phenomenon was already 1913 of Felix Auerbach found in rank and size of cities and is known as Zipf's Law known.
In 1926 Alfred J. Lotka discovered a similar relationship between the number of publications by a person and the number of people with an equally high publication output ( Lotka's law ).
P. Gross and E. Gross were the first to use citations as bibliometric data sources in 1927. They counted and analyzed the citations from each article in a chemical journal and came up with a list of journals that they considered essential to chemical education. The analysis of citations has become a standard tool in the evaluation of scientific journals and researchers.
This was mainly due to the 1963 has Eugene Garfield built Science Citation Index contributed from which the also in 1963 by Garfield presented Impact Factor is calculated. Further relationships such as the bibliographic coupling and co- citation presented by Fano in 1956 , which are also the subject of bibliometric research, can be read from the citation graph .
Another subject of bibliometrics is the Matthew effect postulated by Robert K. Merton in 1968 , which has found its way into many bibliometric models in various forms.
Other important results are Bradford's law, discovered by Samuel C. Bradford in 1934, and the refutation of the Ortega hypothesis by Cole & Cole (1967). Alongside Zipf's Human behavior and the principle of least effort (1949), Little science, big science by Derek de Solla Price , which was published in 1963 and thus established modern scientometry , is the most influential work in the field.
Bibliometrics is usually used as a branch of Scientometry (1966) for the quantitative investigation of science and scientific processes. At the same time, it is a specialty of informetrics (1979), the quantitative investigation of information and thus a sub-area of information science . Since published information units are primarily examined in bibliometrics, it can also be assigned to library science . With the increase in online publications on the Internet , there is increasing overlap with webometry .
criticism
This method of evaluating the personal performance of scientists based on the citation of their articles is not without controversy. A report by the International Mathematical Union warns against the widespread practice of drawing inadmissible conclusions from bibliometric data. In the case of small magazines, the impact factor fluctuates strongly from year to year. Nor can the quality of an article necessarily be inferred from the quality of a magazine. "On the contrary: if a journal A based on a good hundred articles per year has an average citation rate of 0.4 and a journal B in the same subject with a good fifty articles has an impact factor of 0.8, then the probability of error in the judgment lies , a randomly selected essay in A is better than one in B, if it is more than sixty percent.
But if the assessment of individual articles is questionable due to their place of publication, then, as the mathematical report states, a judgment is made about individual scientists and comparisons between them are not possible. This also applies to other, somewhat more refined indicators such as the so-called and currently popular Hirsch index .
The “Leiden Manifesto for research metrics” shows the disadvantages of the bibliometric indicators currently in use and it contains 10 principles that should be observed in research evaluation. This includes z. B. to consider the special features of the different subjects: Scientific publications that are not in English or appear in books are not taken into account in the impact factor. Top-ranked journals in mathematics have an impact factor of 3, while those in cell biology are 30.
An example for reading
For someone new to bibliometrics, it may be a welcome opportunity to look at a relevant example. The bibliometric analysis of a linguistic journal that only publishes articles in the field of quantitative linguistics can be reached with one click . Yanni Lin and Haitao Liu presented a study of this journal.
literature
- Frank Havemann: Introduction to Bibliometrics . Society for Science Studies, 2009 ( PDF ).
- R. Ball, D. Tunger: Bibliometric analyzes - data, facts and methods - basic knowledge of bibliometrics for scientists, science managers , research institutions and universities . Self-published by Forschungszentrum Jülich, 2005, ISBN 3-89336-383-1 ( access to the full text ).
- JR Cole, S. Cole: The Ortega hypothesis . In: Science . No. 178 , 1972, p. 368-375 .
- DJ de Solla Price: General Theory of Bibliometric and other Cumulative Advantage Processes . In: Journal of the American Society for Information Science . tape 27 , no. 5-6 , 1976, pp. 292-306 .
- HD White, KW McCain: Bibliometrics . In: Review of Information Science and Technology . tape 24 , 1989, pp. 119-186 .
- Alan Pritchard: Statistical Bibliography or Bibliometrics? In: Journal of Documentation . tape 25 , no. 4 , 1969, p. 348-349 .
Magazines
- Scientometrics. 1978-
- Bibliometrics - Practice and Research. 2012-
See also
Web links
- Wolfgang Glänzel: A concise introduction to bibliometrics & its history , University of Leuven, ECOOM
- Introduction and script
- Bibliometric service of the University Library Vienna
- Society for Bibliometrics: Portal with articles on the subject of bibliometrics
- List of links to bibliometrics (conferences, journals, people ...) ( Memento of February 23, 2008 in the Internet Archive ) (Archived website)
- European Summer School for Scientometrics (esss)
- Research blog - Bibliometrics category
- Timeline of bibliometrics
- Terminosaurus Rex: Homepage of Information Science
- Competence center bibliometrics
- Bibliometrics2012 - 1st International Bibliometrics Conference and Trade Fair
Individual evidence
- ↑ Viziometrics: Analyzing Visual Information in the Scientific Literature (accessed June 8, 2016)
- ↑ About Viziometrics - [1] ( Accessed June 8, 2016)
- ^ Ball, Rafael .: Bibliometrics: Simple - understandable - comprehensible. De Gruyter, Berlin 2013, ISBN 3-11-029375-7 , p. 53 .
- ↑ Milos Jovanovic: A little early history of bibliometrics . In: Information. Wissenschaft & Praxis , Volume 63, Issue 2, pp. 71–80.
- ↑ http://www.mathunion.org/fileadmin/IMU/Report/CitationStatistics.pdf
- ↑ Jürgen Kaube, Die Bibliometric Verblendung , FAZ, July 24, 2008, p. 36
- ↑ https://www.nature.com/polopoly_fs/1.17351!/menu/main/topColumns/topLeftColumn/pdf/520429a.pdf
- ^ Yanni Lin, Haitao Liu: A Bibliometric Analysis of Glottometrics , in: Glottometrics 39, 2017, p. 1–37 (PDF full text ).