Biased Citation Practice and Taxonomic Parochialism

Scientific endeavour has one goal: to unravel mechanisms underlying the real world with utmost objectivity. Rigorous logic and experimentation are as essential for this purpose as is the consideration of extant knowledge. Bias is the demon of science at any level, and our incessant effort is to escape its temptation. It is always worth to scrutinize our success (Gluud 2006; Lortie et al. 2007). The publication process involving peer review and editorial advice and decision is one way to keep biases down. The adoption of published evidence as basis to build on future research is another mechanism towards this end. The use of previous results as research basis depends on its relevance for a study, which is a function of its specific importance and general quality. Previous research that we use as a basis is referred to in our papers and hence included in their reference lists. A glance at the reference list of a paper hence allows to appraise its context. Any form of bias in the selection of articles to which a study refers, intended or accidental, will hamper this judgement. Nevertheless, there is ample evidence that such unwanted bias exists (Nicolaisen 2007; Bornmann & Daniel 2008). The probability of a paper to be considered and hence cited in subsequent articles is influenced, among others, by social networks (Case & Higgins 2000), the number of authors (Leimu & Koricheva 2005), the authors’ institutional affiliation (Crane 1965; Podsakoff et al. 2008), the author’s and the journal’s reputation (Merton 1968; Callaham et al. 2002; Judge et al. 2007), whether results are ‘positive’ or ‘negative’ (i.e., supportive of the hypotheses tested or not; Kjaergard & Gluud 2002; Leimu & Koricheva 2005; Nieminen et al. 2007), regional affiliations (Wardle 1995; Paris et al. 1998; Grange 1999; Wong & Kokko 2005), whether the authors are from native English speaking countries (Tregenza 2002; Leimu & Koricheva 2005), the alphabetical position of authors’ surnames (Tregenza 1997), and the citing authors’ motivation to persuade the reader (Merton 1968; Gilbert 1977; MacRoberts & MacRoberts 1996); although there are opinions and studies qualifying these observations (Garfield & Welljamsdorof 1992; White 2004; Radicchi et al. 2008). An important reason for bias is the limitation of journal space, which is why often not all sources on which a study is based can be cited (MacRoberts & MacRoberts 1989). Provided this extensive evidence of biased citation practice, any use of bibliometric data for the evaluation of research seems highly problematic (Paris Correspondence Michael Taborsky, Behavioural Ecology, Institute of Ecology and Evolution, University of Bern, Wohlenstrasse 50a, CH-3032 Hinterkappelen, Bern, Switzerland. E-mail: [email protected]