The Impact Factor was developed by Eugene Garfield, the founder of the Institute for Scientific Information (presently a part of Thomson Reuters), in 1955 following ideas inspired by Vannevar Bush's famous article "As We May Think" in 1945 available at http://adammikeal.com/courses/chi/files/jan26.bush.pdf . The original article "Citation indexes for science: a new dimension in documentation through association of ideas" by Garfield was published in Science 1955;122:108-111 and is available online at http://www.garfield.library.upenn.edu/papers/newsci.htm.
The Journal Impact Factor can be used to evaluate or compare a journal's relative importance to other journals in the same field. It shows an average citation rate per article for the past two years. Thomson Reuters publishes annually Journal Impact Factors for journals included in Web of Science data base in its Journal Citation Reports.
The IF for a journal (j) is defined as the number of times articles published in the two years prior to the JCR edition year have been cited in the JCR edition year (C) divided by the number of articles published in those two years (P).
JIFj = C ⁄ P
A diagram of Impact Factor calculation.
It should be noted that only research papers (articles) and review articles (citable articles) are considered in the calculation of the denominator. Letters, editorials etc are not taken into account. However, the numerator reflects citations to all document types including self-citations. Normally the non-citable articles are seldom cited, and thus do not considerably affect the IF of a journal. JCR also provides journal IFs without journal self-cites.
The two-year period used in calculating the IF is a compromise. If it were based on one year's articles only, it would give even greater weight to rapidly changing fields. If it were based on longer periods of citations and sources, the measure would be less current.
The calculation of IF for Cell (2007).
Image source: Journal Citation Reports (Thomson Reuters)<http://www.isiknowledge.com/> 2.4.2009
The Impact Factors of journals fluctuate from year to year, and part of this variation is random. The relatively small number of papers contributing to the Impact Factor means that there is a large random component in the variation. As there are many more journals with a low Impact Factor than journals with a high one, rankings for the low impact ones are less stable than for the high impact ones. For high impact journals, noise and fluctuations have only a small influence on the impact, and do not lead to any change in ranking. However, journals are often compared for a fixed year, without taking into account the higher variation for small journals.