The impact of self-citation

Abstract

Citation rates are used to calculate journal impact factors and for measuring the personal impact of individual scientists (Aksnes, 2003; Fowler and Aksnes, 2007), which can affect chances of academic appointments, advance in career, pay increases and grant funding (Hyland, 2003; Adler, 2009). However, the current method used to calculate citation rates has its limitations. The most widely used index published by the Institute of Scientific Information is the 2-year impact factor (Della Sala and Grafman, 2007, 2009) which calculates citation rate by summing the total citations during a given year to articles published in the previous two years and dividing this by the number of ‘source items’. What constitutes a ‘source item’ is unclear, but the number of source items can obviously affect citation rate (see Della Sala and Brooks, 2008), and can be manipulated to affect journal impact factor (see Brumback, 2009). Journal impact factors can also be highly influenced by the publication of a few highly cited papers (Buchtel and Della Sala, 2006) and it has been suggested that around 20% of all articles tend to account for 80% of all citations. In addition, it has been argued that the arbitrarily set time window of two-years disadvantages slowmoving disciplines, such as neuropsychology (Della Sala and Crawford, 2006, 2007; Della Sala and Grafman, 2009). Another factor that may affect citation rate is self-citation. Authors may choose to cite their own work in order to make their earlier work visible, to help reinforce an argument, to create an image of scientific authority (Hyland, 2003) or simply for self-aggrandizement (Hyland, 2003; Fowler and Aksnes, 2007). Self-citation can inflate the perception of an article’s or a scientist’s scientific impact, particularly when an article has many authors, increasing the possible number of self-citations (Schreiber, 2007; Della Sala and Brooks, 2008), and thus there have been calls to remove self-citations from citation rates (Schreiber, 2007). However, Fowler and Aksnes (2007)