Weak evidence for determinants of citation frequency
in ecological articles
Andre´ Andrian Padial
•
Joa˜o Carlos Nabout
•
Tadeu Siqueira
•
Luis Mauricio Bini
•
Jose´ Alexandre Felizola Diniz-Filho
Received: 28 April 2009 / Published online: 7 May 2010
 Akade´miai Kiado´, Budapest, Hungary 2010
Abstract Citation frequency has been considered a biased surrogate of publication merit.
However, previous studies on this subject were based on small sample sizes and were
entirely based on null-hypothesis significance testing. Here we evaluated the relative
effects of different predictors on citation frequency of ecological articles using an infor-
mation theory framework designed to evaluate multiple competing hypotheses. Supposed
predictors of citation frequency (e.g., number of authors, length of articles) accounted for a
low fraction of the total variation. We argue that biases concerning citation are minor in
ecology and further studies that attempt to quantify the scientific relevance of an article,
aiming to make further relationships with citation, are needed to advance our under-
standing of why an article is cited.
Keywords Citation frequency  Ecological articles  Merit  Biases
Introduction
Citation frequency can be considered a formal, albeit controversial, way to quantify the
overall impact of an article or the relevance of a scientific journal (Garfield 1972). To
assess the academic performance of individual researchers, institutes, universities and even
countries, for several purposes such as funding and selection of applicants for a position,
decision makers can use citation frequency as a surrogate for publication quality (Aksnes
A. A. Padial (&)  J. C. Nabout  T. Siqueira  L. M. Bini  J. A. F. Diniz-Filho
Programa de Po´s-graduac¸a˜o em Ecologia e Evoluc¸a˜o, Departamento de Ecologia, Universidade
Federal de Goia´s, Rodovia Goiaˆnia-Nero´polis km 5, Campus II, Setor Itatiaia, PO Box 131, Goiaˆnia,
GO CEP 74001-970, Brazil
e-mail: aapadial@gmail.com
A. A. Padial
Campus Nova Andradina, Instituto Federal de Educac¸a˜o, Cieˆncia e Tecnologia de Mato Grosso do Sul,
Nova Andradina, MS, Brazil
T. Siqueira
Universidade Federal de Sa˜o Carlos, Sa˜o Carlos, SP, Brazil
123
Scientometrics (2010) 85:1–12
DOI 10.1007/s11192-010-0231-7

2003a; Leimu and Koricheva 2005a). However, a small number of scientific journals
publish most of the ‘‘relevant’’ articles (a pattern known as Bradford’s law) and, thus,
receive more citations (Nicolaisen and Hjørland 2007). Indeed, approximately 20% of
articles account for 80% of all citations (known as the 20/80 phenomenon), and the
majority of publications are never or rarely cited (Garfield 2006). This may falsely lead us
to believe that the majority of articles are of poor quality or contribute little to the
advancement of science. Citation can be biased and not a good surrogate of the value of a
scientific article. In fact, a number of studies have suggested that this last possibility is
plausible (Table 1).
In general, studies investigating biases in citation of articles focus on (van Dalen and
Henkens 2001): (i) journal characteristics, e.g., journals with higher impact factor attract
more citations, (ii) article characteristics (e.g., lengthy articles receive more citations), and
(iii) author characteristics, e.g., famous authors of certain countries (in general, native
English speakers) are more cited. Leimu & Koricheva (Leimu and Koricheva 2005a, b) and
Borsuk et al. (Borsuk et al. 2009) discussed numerous other explanatory variables related
to citation frequency of articles in ecology (see Table 1). If citation frequency is biased by
other factors unrelated to scientific relevance, thus this metric cannot be regarded as an
objective and impartial tool for academic evaluation (Leimu and Koricheva 2005a).
Table 1 Some recent studies presenting variables that presumably explain citation
Subject Determinants of citation
frequency
Number
of
papers
analyzed
Statistical
strategy used
Reference
Chemistry Number of authors 1,733 Exploratory
inference
Kademani et al. (2007)
Crime and
Psychology
Author nationality; Number of
author citation; Type of
publication
428 Negative
Binomial
Regression
Walters (2006)
Ecology Journal Impact Factor; Study
outcome (positive or negative
result); Article length;
Number of authors; Author
nationality; Author affiliation
university
23–5883 Correlation
Analysis;
ANOVA;
ANCOVA
Leimu and Koricheva
(2005a, b), Borsuk et al.
(2009)
Mathematics Article availability 2,765 Exploratory
inference
Davis and Fromerth (2007)
Medical
Sciences
Abstract length; Author
affiliation university; Author
nationality; Number of
authors; Number of cited
References; Number of
pages; Research approach
(e.g. experimental or
empirical studies); Research
topic; Sample size of the
research; Study outcome
(positive or negative result);
Type of publication
34–1274 Exploratory
inference;
Correlation
Analysis;
ANOVA
Celayir et al. (2008), Lokker
et al. (2008), Gehanno et al.
(2007), Nieminen et al.
(2007), Bhandari et al.
(2007), Kostoff (2007),
Pasterkamp et al. (2007)
Physical
Sciences
Article availability 74,521 Exploratory
inference
(Moed 2007)
2 A. A. Padial et al.
123