Is there a clubbing effect underlying Chinese research citation Increases?

Abstract

There is increasing evidence that citations to Chinese research publications are rising sharply. A series of reasons have been highlighted in previous studies. This research explores another possibility—whether there is a “clubbing” effect in China's surge in research citations, in which a higher rate of internal citing takes place among influential Chinese researchers. Focusing on the most highly cited research articles in nanotechnology, we find that a larger proportion of Chinese nanotechnology research citations are localized within individual, institutional, and national networks within China. Both descriptive and statistical tests suggest that highly cited Chinese papers are more likely than similar U.S. papers to receive internal and localized citations. Tentative explanations and policy implications are discussed.

Full text

Is There a Clubbing Effect Underlying Chinese Research
Citation Increases?
Li Tang
School of Public Economics and Administration, Shanghai University of Finance and Economics, Shanghai
200433, China. E-mail: tang006@gmail.com
Philip Shapira
Manchester Institute of Innovation Research, Manchester Business School, University of Manchester,
Manchester M13 9PL, United Kingdom, and School of Public Policy, Georgia Institute of Technology, 685
Cherry Street NW, Atlanta, GA 30332-0345. E-mail: pshapira@mbs.ac.uk
Jan Youtie
Enterprise Innovation Institute and School of Public Policy, Georgia Institute of Technology, 75 Fifth Street
NW, Atlanta, GA 30308. E-mail: jan.youtie@innovate.gatech.edu
There is increasing evidence that citations to Chinese
research publications are rising sharply. A series of
reasons have been highlighted in previous studies. This
research explores another possibility—whether there is
a “clubbing” effect in China’s surge in research cita-
tions, in which a higher rate of internal citing takes place
among influential Chinese researchers. Focusing on the
most highly cited research articles in nanotechnology,
we find that a larger proportion of Chinese nanotechnol-
ogy research citations are localized within individual,
institutional, and national networks within China. Both
descriptive and statistical tests suggest that highly cited
Chinese papers are more likely than similar U.S. papers
to receive internal and localized citations. Tentative
explanations and policy implications are discussed.
Introduction
China’s rise in science in general and particularly in some
cutting-edge fields has been widely documented (Adams,
King, & Ma, 2009; Zhou & Leydesdorff, 2007). Evidenced
by the number of publications indexed in the Web of Science
(WoS), China’s scientific output has demonstrated rapid
growth. In 2005, China was ranked fifth by a number of
scholarly publications across all disciplines, after the United
States, the United Kingdom, Germany, and Japan (Organisa-
tion for Economic Co-operation and Development [OECD],
2007). By 2010, China had moved to second place in terms of
publication output after the United States (Moiwo & Tao,
2013; Zhang, Patton, & Kenney, 2013). A2020 target to rank
among the world’s top-five countries by aggregated scientific
paper citations was set by Chinese government planners in
2006—a target that was achieved in 2012 (Bound, Saunders,
Wilsdon, & Adams, 2013). While normalized citation impact
measures for China’s research papers as a whole are still
below those of the United States and other leading science
nations, Chinese research publications are increasing in
quality and visibility, with several fields in China now at or
close to the respective world citation impact averages.
Chinese citation averages are relatively high in mathematics
and agricultural sciences, and have grown strongly over the
last decade in such areas as computer science, space science,
biology and chemistry, and materials science (Bound et al.,
2013; Moiwo & Tao, 2013). In the nanotechnology domain,
China as of 2010 published the most WoS-indexed papers
(Kostoff et al., 2012; Tang & Shapira, 2011a; Zhou &
Leydesdorff, 2007). Kostoff (2012) found that the number of
China’s highly cited papers, or “heavy hitters,” in nanotech-
nology continues to grow. When benchmarked with other
leading countries, the citation gap of Chinese nanotechnol-
ogy papers has narrowed over the years (Hu & Rousseau,
2013; Youtie, Shapira, & Porter, 2008).
What are the drivers behind China’s rise in citations?
Several accounts have been explored in previous studies.
One explanation is that the quality of Chinese research has
been increasing over time, with better papers being
Received October 28, 2013; revised March 23, 2014; accepted March 24,
2014
© 2015 The Authors. Journal of the Association for Information Science
and Technology published by Wiley Periodicals, Inc. on behalf of
ASIS&T •Published online 5 March 2015 in Wiley Online Library
(wileyonlinelibrary.com). DOI: 10.1002/asi.23302
JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGY, 66(9):1923–1932, 2015
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction
in any medium, provided the original work is properly cited.

published in journals with higher impact factors, which in
turn leads to them attracting more citations (Appelbaum,
Parker, & Cao, 2011; Guan & Ma, 2007). Suttmeier, Cao,
and Simon (2006) posited that the fierce competition among
the numerous Chinese scientists has pushed forward better
research. Jin, Rousseau, Suttmeier, and Cao (2007) high-
lighted the role of diasporic linkages in China’s internation-
ally collaborative research. In a study conducted by Tang
(2013), Chinese knowledge moderators, who collaborate
intensively with both Chinese and U.S. researchers, have
facilitated China’s rise in research quality. Other scholars
have argued that China’s research is now more visible and
understandable to the global community, evidenced by the
growing numbers of Chinese publications written in English
and greater indexing of Chinese journals in the WoS (Huang,
Notten, & Rasters, 2011; Lin & Zhang, 2007; Ren &
Rousseau, 2002). Some studies have further suggested that
more Chinese papers are published in heavily cited research
disciplines such as physics, bioscience, and interdisciplinary
research (e.g., Shapira & Wang, 2010). Still others have
posited that rising citations are due to China’s rapidly
expanding collaboration network, especially at the interna-
tional level (Adams et al., 2009; Costas, van Leeuwen, &
Bordons, 2010).
In this article, we examine yet another possibility:
Chinese researchers are more likely to be cited by internal
and localized networks within their own country than are
their counterparts in other nations. Using nanotechnology as
a test bed, we explore whether there is a “clubbing” effect in
China’s growth of research citations. The article is organized
as follows. The next section briefly reviews extant research
related to self-citations and clubbing phenomena in forward
citations. A description of data sources and methods follow.
We then examine a set of highly cited Chinese and U.S.
nanotechnology publications to explore the factors underly-
ing citations and apply regression modeling to test for
country differences in “clubbing” effects. The article con-
cludes with a discussion of tentative explanations, limita-
tions, and policy implications.
Background
“Clubbing” Effects on Forward Citations
The concept of “clubbing” has been used in research that
has examined situations where elite authors extensively cite
each other (Opsahl, Colizza, Panzarasa, & Ramasco, 2008),
or more broadly to understand the development of dominant
communities of researchers in science (Colizza, Flammini,
Serrano, & Vespignani, 2008). Stimulated by this notion,
this study looks at the clubbing of forward citations of
Chinese highly cited research and compares it against U.S.
counterparts. We want to ascertain if there exists a national
difference in clubbing effects between the two countries. To
refine and operationalize the phenomenon of clubbing asso-
ciated with citation impact, we first review several relevant
issues, including discussion about the potential role of self-
citation. Self-citations are not necessarily the same as the
internal citations represented in the clubbing effect from the
perspective of the individual author, but the concept is analo-
gous at the country level.
Citation and Self-Citations
Citation is a generally accepted proxy of research quality
or, to be more accurate, of visibility or scientific impact
(Garfield, 1979; Redner, 1998). One noise factor in the cita-
tion indicator is self-citation; that is, when an author cites his
own research (Moed, 2002; Noyons, Moed, & Luwel, 1999).
The potential confounding effects of self-citation have been
widely discussed in the bibliometric community. Some
scholars have argued that self-citation significantly affects
results and have called for the removal of self-citations in
impact measurement. Van Raan (1998) demonstrated that
the fraction of self-citations is higher in collaborative papers
and especially in international coauthored research. Hall,
Jaffe, and Trajtenberg (2001) noted that by interacting with
the phenomenon of citation truncation, self-citation spill-
overs have an upward bias on coefficients in more recent
cohorts.
On the contrary, others have argued that at the macro
level, or when a sufficiently large number of publications are
examined, there is no need to exclude self-citations (Glänzel
& Meyer, 2003; Glänzel & Thijs, 2004a; Pichappan &
Sarasvady, 2002). According to a survey conducted by
Bonzi and Snyder (1991), there are no significant differ-
ences in the motivations for self-citations and citation to
others. Persson, Glänzel, and Danell (2004) argued that the
self-citation rate of collaborative papers has a diminishing
effect on citation inflation. Based on an analysis of 1991 to
1999 self-citations, they found that not only did the absolute
number of self-citations increase to a lesser extent than did
non-self-citations but also the relative shares of self-citation
decreased slightly over time. Rehn and Kronman (2008)
further explained that this decline may be because scholars’
citing behavior in scientific communication is mostly merit-
based. They speculated that authors in the United States cite
themselves more than the average amount, albeit without
empirical evidence. Combining both perspectives, Gedik
(2012), in his recent study, posited that whether to remove
self-citations is debatable because inclusion overestimates
localization while exclusion of self-citations underestimates
localization.
Relatively little research has been devoted to the analo-
gous clubbing effect of internal citations among a group of
elite researchers, with few exceptions on own-group prefer-
ence on languages and journals. Yitzhaki and collaborators
(Bookstein & Yitzhaki, 1999; Egghe, Rousseau, & Yitzhaki,
1999) developed an indicator of relative own-language pref-
erence to depict the citing behavior of own-group preference
on the language aspect. Ren and Rousseau (2002) analyzed
the internationality of citations received by Chinese journals
in the fields of physics and chemistry. Focusing on 18
Chinese journals in the Institute of Science Index (ISI) in
1924 JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGY—September 2015
DOI: 10.1002/asi

1998, they observed that many citations of Chinese journals
originated from other Chinese journals. Later, based on
inequality theory and a weighted Lorenz curve, Egghe and
Rousseau (2004) proposed a framework to measure own-
group preference. Despite these works, it is not clear to what
extent the clubbing of citations influences the total citations
received at micro (individual), meso (institutional), and
macro (national) levels. It also remains unknown whether
the extent of citation clubbing varies by country.
Methods
Data
We combine text mining and regression analysis to
examine the clubbing phenomenon among highly cited U.S.
and Chinese nanotechnology papers, or “heavy hitters.”1The
selected domain is nanotechnology, which was chosen
because of China’s rapid emergence in this multidisciplinary
domain in both research quantity and impact. We focus on
heavy hitters because they account for a large proportion of
citation output in highly skewed publication and citation
data (Allison & Stewart, 1974; Kostoff, 2012; Lotka, 1926;
Phelan, 1999).
Our database consists of two parts. One contains papers
of heavy hitters while the other contains all articles which
have cited these heavy hitters. We use the Georgia Tech
global nanotechnology publication data set (2011) to iden-
tify heavy hitters’ nanotechnology articles. For more details
on how this large-scale data set was created, including its
multistage complex Boolean search strategy, please refer to
Porter, Youtie, Shapira, and Schoeneck (2008) and Youtie
et al. (2008). We confine our cross-country comparison to
the two most actively publishing nations: the United States
and China.2The top-20 most cited nanotechnology articles
for the single years of 2000, 2004, and 2008 were selected
for the United States and China, respectively, which resulted
in 120 seed articles (i.e., 20 ×3×2=120). The initial year
reflects the period just before the introduction of China’s
National Development Plan for Nanoscience and Nanotech-
nology (comparable to the U.S. National Nanotechnology
Initiative, which was introduced several months earlier). The
middle period represents a phase of moderate growth in
China’s nanotechnology publications. The latter period
covers a phase of rapid growth of nanotechnology publica-
tions. The full bibliographical data of these articles’ forward
citations (N=62,338) were accessed in September 2011 and
linked with each of the 120 seed articles.3
For the heavy hitters, only original research articles are
included. We did not differentiate fields when selecting
highly cited articles so that we can observe whether differ-
ences between citation patterns of highly cited research by
the United States and China also vary by field (although all
highly cited seed articles fall in the nanotechnology
domain). We stop at the 2008 cohorts to capture recent
research developments in nanotechnology, yet allow time (at
least 2 years) for each article to garner citations.
For the citing data set, we follow common practice and
have included three types of documents: normal articles,
review articles, and letters. Other types of papers (e.g.,
meeting abstracts, proceeding papers, corrections, and bio-
graphical items, etc.) have been removed. Several rounds of
data cleaning and standardization were conducted, with the
aid of VantagePoint software,4which included automatic
name matching developed by Tang and Walsh (2010) and
manual validation.
Analysis of Results
Basic Descriptives
Unsurprisingly, all 120 heavy hitter nanotechnology
papers are written in English. Articles written in English are
more likely to be cited by researchers than are other papers
of the same level of quality (Hu, Carley, & Tang, 2012;
Liang, Rousseau, & Zhong, 2013). English is now the key
language of communication in the global scientific commu-
nity, although we also need to keep in mind limitations in the
coverage of non-English language journals in the WoS.
U.S. authors take a leading role in China’s most cited
nanotechnology research. Among China’s top-60 heavy
hitting nanotechnology papers, 24 are coauthored with
scholars outside China, and 13 of these have U.S. collabo-
rators. In sharp contrast, none of the U.S. heavy hitting
nanotechnology papers involve a Chinese coauthor.
Tables 1 compares the collaboration profiles of the U.S.
and Chinese cohorts. Echoing Aksnes (2003), we find that
heavy hitting nanotechnology papers are typically multi-
authored, multi-affiliated, and often internationally collab-
orative. The median number of authors per paper is 6 for the
United States and 5 for China, but Chinese scholars are
slightly more likely to collaborate across research institutes
and national boundaries (Table 1). Citations to our 120 seed
articles range from 43 to 2,362 (M=416, SD =43). Journals
in which these citing articles appear include high-profile
sources such as Nature (2011 journal impact factor =36.28)
and field-specific journals such as Solid State Communica-
tions (2011 journal impact factor =1.65).5Table 2 lists the
top-five subject categories and journals in which Chinese
and U.S. heavy hitters published. Regardless of the small
1Following common practice, a whole counting method is adopted in
the allocation of country origin based on the reported affiliations in the
bylines. So a Sino–U.S. collaborated paper is allocated as both a U.S.
publication and a China publication.
2In this article, China refers to mainland China, including Hong Kong
and Macau.
3Duplicate citations have been removed in the two citing files. That is,
if Article P1 cited more than one Chinese heavy hitter paper, Article P1 only
appears once in the citing database for China.
4VantagePoint is a text-mining tool for bibliometric and patent data-
bases; see https://www.thevantagepoint.com
5Journal Citation Reports, 2011, Thomson Reuters, Institute for Scien-
tific Information Web of Knowledge; http://wokinfo.com/products_tools/
analytical/jcr/ (retrieved March 22, 2014).
JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGY—September 2015 1925
DOI: 10.1002/asi

size of our heavy hitter set, which is sized as such because it
comprises the most highly cited publications, U.S. and
China cohorts are quite comparable in terms of publication
journals and research foci (Table 2).
We next examine the citation distribution of the heavy
hitting nanotechnology papers (Table 3). The United States
easily leads China in all absolute counts of citations. Against
the expected decreasing trend of citations due to citation
truncation, the total number of citations to Chinese 2004
cohorts surpassed that of the 2000 cohort. The number of
U.S. articles citing China heavy hitters varies by year, the
proportion of U.S. papers citing Chinese heavy hitters
increases over time: from 16% in 2000 to 17% in 2004 and
to 22% in 2008. We do not find the same pattern for U.S.
heavy hitters, which may suggest an increased research
impact of Chinese-authored publications over time. Addi-
tionally, Chinese publications accumulate a substantially
greater share of citations from China than do U.S.-authored
papers. As demonstrated in the table, 52% of papers citing
Chinese heavy hitter articles are from China while only 36%
of papers citing U.S. heavy hitters are U.S.-authored.
The clubbing effect of Chinese forward citations is not
confined to the national level alone; it also is evident at the
institutional and individual levels within China. Figure 1
presents box plots which display the proportions of within-
institution, individual, and country-level citations for China
and the United States. As illustrated in Figure 1, almost all
box plots for China (except the clubbing share at the insti-
tutional level in 2008) are longer than comparable U.S.
groups, indicating that overall, Chinese publications are
more heterogeneous than are the U.S. counterparts in terms
of the clubbing effect. In addition, without exception, all box
plots associated with Chinese-authored publications are
noticeably higher than are the equivalents for the United
States, suggesting that Chinese researchers tend to internally
cite their own work to a greater extent at all three levels.
Statistical Regression
The descriptive statistics suggest that a larger proportion
of Chinese nanotechnology research citations are localized
within individual, institutional, and national boundaries.
We next test whether these distinctions hold when control-
ling for confounding factors such as research domain,
extent of collaboration, and year. Our null hypothesis is
straightforward:
H0: China’s research has a stronger citation clubbing effect
than does the United States.
The alternative hypothesis is:
H1: There is no statistical country-level difference in the
citation clubbing effect between China and the United States.
The regression model is as follows:
Y X where Y represents a clubbing effect
ii i
=+
ε
i,
Y
Club Person
Club Inst
Club Country
i=
⎡
⎣
⎢
⎢
⎢
⎤
⎦
⎥
⎥
⎥
_
_
_
Xiis a vector of characteristics impacting Yi, and εiis an
error term. Table 4 presents variables and measures in this
section.
Measurement
Dependent variable. We measure the clubbing effect from
three dimensions of within-level citation (micro, meso, and
macro levels). For each heavy hitter, we calculate
• Share of citations by primary authors (i.e., first author and
corresponding author),
• Share of citations by papers with coauthors from the same
institution, and
• Share of citations by papers with coauthors in the same
country.
These three outcome variables are proportions bounded
by 0 and 1 inclusive. For individual self-citations, we
confine our analysis to primary authors due to well-known
author ambiguity issues (Galvez & Moya-Anegon, 2007;
Onodera et al., 2011)
Independent variable. Our explanatory variable is China’s
heavy hitters (i.e., a highly cited Chinese nanotechnology
articles). This is a dummy variable, and we code it as 1 if
the heavy hitter involves at least one scholar from China;
otherwise, it is coded as 0.
Control variables.
Collaboration scope. Both forward citation and internal-
country/institution/individual citation distributions are func-
tions of a variety of factors. Glänzel and Thijs (2004b) found
that the number of coauthors has less effect on self-citations
than it does on external citations. Costas et al. (2010) found
TABLE 1. Collaboration scopes of heavy hitters: China versus the United
States.
No. of Authors No. of Affiliations
Min. MMdnMax. Min. M Mdn Max.
China 2 6.8 5 64 1 2.3 2 33
United States 2 11 6 192 1 2.2 1 12
No. of Countries ΣCitations
Min. M Mdn Max. Min. MMdnMax.
China 1 1.6 1 11 43 217.4 193.5 576
United States 1 1.3 1 5 90 614.4 493.5 2362
Note. Heavy hitters comprise a distributed set of the most highly cited
papers in nanotechnology published by Chinese and U.S. authors in 2000,
2004, and 2008 (N=120) (see text for further details).
1926 JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGY—September 2015
DOI: 10.1002/asi

that the number of research centers is positively correlated
with individual self-citations. Drawing on this line of work,
we have estimated three collaboration-scope variables:
AFFILIATIONS (number of organizations with which the
authors are affiliated), AUTHORS (number of coauthors),
and COUNTRIES (number of countries in which the orga-
nizations are situated).
6The country coverage of our data is 99%. Table 3 only lists statistics for
those citing records reporting country affiliations.
7The third country citation refers to citations from countries other than
the United States or China. As we adopted the whole counting method for
country allocation, the third country citing paper can be international col-
laborated articles between the United States and China, but it also must
have involved another country.
TABLE 2. Research foci of heavy hitters: China versus the United States.
Rank
Top-five subject categories Top-five publication sources
China United States China United States
1 Chemistry, Multidisciplinary Multidisciplinary Sciences Journal of the American Chemical Society Science
2 Materials Science, Multidisciplinary Chemistry, Multidisciplinary Applied Physics Letters Nature
3 Physics, Applied Materials Science,
Multidisciplinary
Angewandte Chemie International Edition Journal of the American
Chemical Society
4 Chemistry, Physical Nanoscience & Nanotechnology Nature Nano Letters
5 Multidisciplinary Sciences Physics, Applied Science Nature Nanotechnology
Note. Heavy hitters comprise a distributed set of the most highly cited papers in nanotechnology published by Chinese and U.S. authors in 2000, 2004,
and 2008 (N=120) (see text for further details).
TABLE 3. Citations distribution of heavy hitters.
Pub_Year ΣCitations6
Domestic
citations
International
citations
Citations origins (share)
United States China Third Country7
United States 2000 18,205 6,333 11,872 6,333 (35%) 4,140 (23%) 9,771 (54%)
2004 10,196 3,498 6,698 3,498 (34%) 2,789 (27%) 5,216 (51%)
2008 8,465 3,407 5,058 3,407 (40%) 1,671 (20%) 4,703 (55%)
Total 36,866 13,238 23,628 13,238 (36%) 8,600 (23%) 19,690 (53%)
China 2000 4,520 2,136 2,384 743 (16%) 2,136 (47%) 2,102 (47%)
2004 5,259 2,988 2,271 902 (17%) 2,988 (57%) 2,014 (38%)
2008 3,266 1,627 1,639 727 (22%) 1,627 (50%) 1,416 (43%)
Total 13,045 6,751 6,294 2,372 (18%) 6,751 (52%) 5,532 (42%)
FIG. 1. Box plots of clubbing effects of forward citations to heavy hitter nanotechnology papers (0 =United States 1 =China). [Color figure can be viewed
in the online issue, which is available at wileyonlinelibrary.com.]
JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGY—September 2015 1927
DOI: 10.1002/asi

International collaboration. In parallel with its invest-
ment in domestic-research activities, China has increasingly
integrated itself into the international-collaboration arena.
Previous studies have suggested that there is a positive
correlation between international collaboration, particularly
Sino–U.S., and cumulative citations (Tang & Shapira,
2011b, 2012). To remove the confounding factor of
collaboration beyond national borders, we add two dummy
variables US_CNCOL (Sino-U.S. collaboration) and OTH-
COLLAB (third-country collaboration) in the regression
model.
Research domain. The extent of citation clubbing also
varies substantially by research field (Snyder & Bonzi,
1998). We adopt the mega field approach developed by
Porter and colleagues (see Garner, Porter, Borrego, Tran, &
Teutonico, 2013; Porter & Rafols, 2009; Porter & Youtie,
2009) and apply it to the subject categories in our data set,
which returns five metadisciplines: Biology and Medicine,
Physical Science and Technology, Environmental Science
and Technology, Psychology and Neurology, Computer
Science, and Engineering (The original list has six metadis-
ciplines, but we excluded the Social Sciences category
because it is outside this study’s scope). We then generate a
set of five dummy variables and add them to the regression
model to control for research domain effect.
Year cohorts. Intuitively, research is often noticed and
cited by people working in the same or closely related
groups before coming to the attention of other groups
(Gupta, Campanha, & Pesce, 2005). Taking the temporal
effect of within-group citations into consideration, with
years representing closely related groups, we add two
dummy variables for publication years in the regression
model. The reference year is 2000. Summary descriptive and
correlation statistics (Tables 5 and 6) indicate no significant
issues for the regression. Unsurprisingly, AFFILIATIONS
and COUNTRIES are highly correlated, but they appear in
different models, so this level of correlation does not affect
the regressions.
Regression Results
Density plots show that dependent variables follow a
conventional generalized linear model (GLM) distribution.
Considering the nature and distribution of the three outcome
variables, we adopt the fractional logit regression approach
proposed by Papke and Wooldridge (1996). STATA Version
12 was used for modeling.
As illustrated in Table 7, it is clear that there are
significant country differences in clubbing effect
(HCP_CTRY) at all three levels. Chinese researchers are
more likely than are their U.S. heavy hitter counterparts to
internally cite work at the country, institution, and indi-
vidual levels. For robustness testing, we also ran multivari-
ate regressions, as these three outcome variables represent
a loosely defined clubbing effect. As shown in Table 7, the
results are similar to those obtained in GLM regression.
TABLE 4. Variable description.
Construct Variable Description
Clubbing effect CLUB_PERSON Percentage of self-citations of primary authors (including first author and correspondent author)
CLUB_INST Percentage of institutional citations; that is, no. of citing paper affiliated with (co)authors’ affiliations
/total citations
CLUB_CTRY Percentage of home country citations; that is, proportion of citations from author’s or coauthor’s
affiliated country
Home country of heavy
hitters
HCP_CTRY 1 if home country is China; 0 if the United States
Sino–U.S. collaboration US_CNCOL International collaborated article involved countries of the United States and China =1; otherwise =0
Third country
collaboration
OTHCOLLAB At least one author associated with an affiliation outside China and the United States =1;
otherwise =0
Scope of research
collaboration
AUTHORS No. of coauthors
AFFILIATIONS No. of affiliations associated with coauthorship
COUNTRIES No. of coauthors’ countries of affiliation
Research discipline SUBJECT F1–F5: Metadiscipline based on Thompson ISI subject category
Publication year PUB-YEAR 2000, 2004, 2008
TABLE 5. Summary of descriptive statistics (N=120).
Construct Variable MSDMin Max
Clubbing effect CLUB_PERSON .05 .05 .00 .35
CLUB_INST .11 .09 .01 .49
CLUB_CTRY .46 .17 .10 .98
Home country of heavy
hitters
HCP_CTRY .50 .5 0 1
Sino–U.S. collaboration US_CNCOL .11 .31 0 1
Third country
collaboration
OTHCOLLAB .21 .41 0 1
Scope of research
collaboration
AUTHORS 8.93 19.59 2 192
AFFILIATIONS 2.25 3.3 1 33
COUNTRIES 1.47 1.14 1 11
1928 JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGY—September 2015
DOI: 10.1002/asi

All pvalues associated with the overall model are less than
.0001, indicating that the multivariate analysis of variance
model is statistically significant regardless of what type of
multivariate criteria are used. The independent variables
explain 14, 46, and 42% of the variance in the outcome
variables CLUB_PERSON,CLUB_INST,and CLUB_C-
TRY, respectively. We did not observe consistent patterns
of effects from other variables such as research disciplines,
publication years, international collaborations, or collabo-
ration scopes.
Discussion and Conclusion
Major Findings
This study contributes to further understanding of China’s
rising research paper citations. As discussed earlier, there are
several perspectives on the causes of China’s rapid growth in
citations. Some deem this citation growth to be a valid
indicator of China’s rise in science from a research-impact
perspective. Behind this type of argument lies an assumption
that citation is a merit-based indicator. Others have argued
TABLE 6. Correlation matrix.
Variable 1 2 3 4 56789
1CLUB_PERSON 1.00
2CLUB_INST .58 1.00
3CLUB_CTRY .34 .31 1.00
4HCP_CTRY .24 .50 .46 1.00
5US_CNCOL .10 .19 −.20 .35 1.00
6OTHCOLLAB .16 .30 −.11 .06 .02 1.00
7AUTHORS −.13 .06 .04 −.11 .03 .29 1.00
8AFFILIATIONS −.07 .39 −.16 .02 .28 .43 .49 1.00
9COUNTRIES .03 .40 −.25 .13 .45 .62 .36 .88 1.00
TABLE 7. Regression results.
Panel 1 GLM Panel 2 MANOVA
CLUB_PERSON CLUB_INST CLUB_CTRY CLUB_PERSON CLUB_INST CLUB_CTRY
HCP_CTRY .473* .910*** .736*** .022 .0849*** .196***
(2.54) (5.04) (6.27) (1.960) (5.480) (6.410)
AUTHORS −.0288** −.000 −.001 .001
(−2.59) (−1.18) (−1.91) (1.600)
AFFILIATIONS .0432*** −.003 .0155** .0214*
(5.07) (−.74) (3.200) (2.240)
COUNTRIES −.304*** .003 −.020 −.126***
(−3.69) (.230) (−1.31) (−4.20)
OTHCOLLAB .524*.371 .209 .028 .0489* .064
(2.11) (1.95) (.98) (1.780) (2.260) (1.490)
pub_year2 .288 .179 .114 .013 .011 .021
(1.47) (1.08) (.95) (1.190) (.710) (.690)
pub_year3 .225 .239 .269* .010 .018 .032
(.93) (1.60) (2.08) (.850) (1.100) (.990)
Discp2 .0395 .0115 −.0422 .002 −.002 −.002
(.15) (.05) (−.26) (.160) (−.11) (−.06)
Discp3 .0794 −.0487 −.106 .004 −.011 −.022
(.32) (−.25) (−.81) (.270) (−.58) (−.57)
Discp4 −.310 .231 −.0853 (.015) .024 −.012
(−1.27) (1.06) (−.51) (−.96) (1.090) (−.28)
Discp5 −.577 −.703* .0326 −.022 −.058 .021
(−1.29) (−2.01) (.08) (−.84) (−1.61) (.300)
_cons −3.296*** −3.010*** −.251* .0312* .0477** .456***
(−14.47) (−20.40) (−2.27) (2.390) (2.640) (12.790)
N120 120 120 120 120 120
Note. t statistics in parentheses. GLM =Generalized linear model; MANOVA =Multivariate analysis of variance.
*p<.05. **p<.01. ***p<.001.
JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGY—September 2015 1929
DOI: 10.1002/asi

that the growth of citations to Chinese articles also is a
function of the expansion of research output or rising-quality
association with greater international collaboration. Our
study identifies another explanation for China’s fast rise in
research impact from the perspective of forward citations.
We find strong support for our conjecture that there is a
Chinese clubbing effect with respect to citation. Both
descriptive and statistical tests have suggested that highly
cited Chinese papers are more likely than are similar highly
cited U.S. papers to be cited by works from China and from
the same institution or author; that is, substantial citation
differences between U.S. and Chinese highly cited papers
exist at all three levels of internal citation.
There are several compelling reasons for Chinese schol-
ars to internally cite their research. China’s science and
technology evaluation system is undergoing a significant
change, with increasing emphasis on WoS-indexed publica-
tions, top journals based on journal impact factor, and cumu-
lative citations counts (China Education Daily, 2012). This
provides powerful incentives for placement of articles in
journals indexed in the WoS (see discussion in Hvistendahl,
2013) and for seeking citations wherever they are
available—and they are most readily available from internal
Chinese sources. The norms of interpersonal relationships
(guanxi) in China may lead Chinese scholars to cite the work
of their colleagues in the same institute, who they meet
frequently, or leading scholars in their own country, who
have an influence in proposal review and external evaluation
for promotion. It also is plausible that a high concentration
of Chinese research on specific topics, as designated by
Chinese funding agencies, can lead to a higher clubbing
effect in China. Wang (1992) and Tang (2007) indicated that
Chinese researchers may be exceptionally inclined to pursue
or even shift to designated hot topics initiated by funding
agencies. When a growing number of Chinese researchers
work on the same or related research topics, the internal
citations among these authors naturally increase. Our model
includes controls for broad field areas (metadisciplines), and
these coefficients were not significant, but perhaps a more
fine-grained approach to field-effect specification would
yield greater distinctions.
The clubbing effect of citation to Chinese research is not
the whole story, however. The evidence in support of a
clubbing effect does not necessarily invalidate the other
explanations. As expected, our data indicate that U.S.-based
authors are leading international collaborators in China’s
high citation impact nanotechnology research. Our results
indicate that U.S. scholars increasingly cite Chinese nano-
technology research. A recent study conducted by Kostoff
(2012) has suggested that China’s fast-growing citations are
highly associated with the emergence of heavy hitters. We
extend his findings by suggesting that not only is the quan-
tity of China’s heavy hitters growing but this very core of
Chinese heavy hitters is attracting an increasing number of
external citations.
In addition to serving as an indicator of international
impact, citations have been depicted as trails of knowledge
flow among different entities (Chen & Hicks, 2004; Cronin
& Overfelt, 1994; Jaffe, Trajtenberg, & Henderson, 1993).
Our study shows that a larger proportion of citations to
Chinese research are localized within individual, institu-
tional, and country levels. One might argue that a higher
propensity of Chinese scholars to cite research from their
own country and institute may suggest a greater exchange of
localized scientific knowledge in China than in the United
States.
Limitations and Future Study
We acknowledge that our research has several limita-
tions. The year selection and the small sample size limit the
generalizability of our findings. Challenges in cleaning and
disambiguating author names as well as the need to focus on
truly highly cited papers necessitated that we work with a
manageably sized set of seed and citing articles. Future work
could extend this sample size, albeit with the caveat that
some falloff in citation quality occurs the more the sample is
extended. We did not observe that field and temporal vari-
ables had significant influence on clubbing effects. More
fine-grained measurement of both done with larger sample
sizes in future work may yield greater differences. Finally,
this study on the clubbing effect in China versus the United
States is based on secondary data. It would be useful to use
interviews and other primary data methods to further probe
why there are differences in the internal citation propensities
of U.S. and Chinese researchers. It also would be interesting
to further explore author networks of Chinese-born
researchers working in the United States, and the possible
effects of these Chinese-born researchers on increases in
U.S. citations to Chinese research.
Policy Implications
Ideally, the production, accumulation, and exchange of
scientific knowledge proceeds in an open pattern (Popper,
1963), although it has long been recognized that institutional
factors also affect how knowledge advances (Kuhn, 1970).
Undoubtedly, internal citation is an indispensable element in
the research-exploration process. Important work should be
recognized and built upon wherever it is developed. Yet,
self-citations—and by extension, our representation of them
as internal citations—introduce risks to the merit-based cir-
culation of knowledge beyond the original knowledge pro-
ducers (Costas et al., 2010). How should the influence of
internal and localized citations be assessed when evaluating
the research performance and impact of an individual scien-
tist, a research unit, and countries? This question is becom-
ing increasingly important today, as the global research
landscape is evermore populated by emerging economies.
China’s growing influence on science as measured by pub-
lication numbers and citations has attracted worldwide
commentary and discussion. Given the sheer size of the
Chinese research sector, and considering the scale and
growth of Chinese research and development personnel, the
1930 JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGY—September 2015
DOI: 10.1002/asi

question arises: How does the preferential citing of Chinese
papers by other Chinese papers distort quantitative measures
of the impact of national scientific outputs?
The caveats associated with using citation for assessing
research impact have been extensively discussed. Despite
increasing awareness of discredited citation, negative cita-
tion, over-self-citation, and even the selling to Chinese
researchers of authorships in potentially well-cited papers
(Hvistendahl, 2013), a majority of existing studies have
treated citations homogeneously regardless of temporal or
contextual features (Cardillo et al., 2013; King, 2004;
Taylor, Perakakis, & Trachana, 2008). Our study provides
some evidence of the heterogeneous nature of citations from
a national-origin perspective. Depending on what one tries
to measure, simple citation counts themselves may veil
important and influential underlying practices. Simple
evaluation criteria such as publication quantity, journal-
ranking placements, and paper citations are increasingly
adopted by the stakeholders who oversee research funding
and personnel promotion decisions (Lawrence, 2003, 2007).
Our findings add further weight to those who argue for
caution in how quantitative measures of research impact are
interpreted and used—and to be sensitive as to how such
measures may in turn reflect, or indeed influence, particular
research-community practices.
Acknowledgments
We are grateful for the comments and suggestions from
two anonymous reviewers. This research draws on support
from the National Science Foundation of China (NSFC)
under Project #71303147, Shanghai Pujiang Program
13PJC052, the U.S. National Science Foundation (NSF)
through the Center for Nanotechnology in Society (Arizona
State University; Award No. 0531194), and the U.K. Eco-
nomic and Social Research Council under Project
ES/J012785/1. The conclusions contained herein are those
of the authors and do not reflect the views of the NSFC, the
NSF, or the ESRC.
References
Adams, J., King, C., & Ma, N. (2009). Global Research Report-China:
Research and collaboration in the new geography of science. Evidence
Ltd, a Thomson Reuters company. ISBN: 1-904431-22-4.
Aksnes, D.W. (2003). Characteristics of highly cited papers. Research
Evaluation, 12(3), 159–170.
Allison, P.D., & Stewart, J.A. (1974). Productivity differences among sci-
entist: Evidence for accumulative advantage. American Sociological
Review, 39(4), 596–606.
Appelbaum, R.P., Parker, R., & Cao, C. (2011). Developmental state and
innovation: Nanotechnology in China. Global Networks, 11(3), 298–314.
Bonzi, S., & Snyder, H.W. (1991). Motivations for citation—A comparison
of self citation and citation to others. Scientometrics, 21(2), 245–254.
Bookstein, A., & Yitzhaki, M. (1999). Own-language preference: A new
measure of relative language self-citation. Scientometrics, 46(2), 337–
348.
Bound, K., Saunders, T., Wilsdon, J., & Adams, J. (2013). China’s absorp-
tive state: Research, innovation and the prospects for China-UK collabo-
ration. London, England: Nesta.
Cardillo, A., Gómez-Gardeñes, J., Zanin, M., Romance, M., Papo, D., del
Pozo, F., & Boccaletti, S. (2013). Emergence of network features from
multiplexity. Scientific Reports, 3, Article number 1344. DOI: 10.1038/
srep01344.
Chen, C.M., & Hicks, D. (2004). Tracing knowledge diffusion. Sciento-
metrics, 59(2), 199–211.
China Education Daily, July 19, 2012. Reforms and Innovations on research
evaluations. Retrieved from http://paper.jyb.cn/zgjyb/html/2012-07/19/
content_74337.htm
Colizza, V., Flammini, A., Serrano, M.A., & Vespignani, A. (2008). Detect-
ing rich-club ordering in complex networks. Nature Physics, 2(2), 110–
115.
Costas, R., van Leeuwen, T.N., & Bordons, M. (2010). Self-citations at the
meso and individual levels: Effects of different calculation methods.
Scientometrics, 82(3), 517–537.
Cronin, B., & Overfelt, K. (1994). Citation-based auditing of academic
performance. Journal of the American Society for Information Science,
45(2), 61–72.
Egghe, L., & Rousseau, R. (2004). How to measure own-group preference?
A novel approach to a sociometric problem. Scientometrics, 59(2), 233–
252.
Egghe, L., Rousseau, R., & Yitzhaki, M. (1999). The own-language pref-
erence”: Measures of relative language self-citation. Scientometrics,
45(2), 217–232.
Galvez, C., & Moya-Anegon, F. (2007). Approximate personal name-
matching through finite-state graphs. Journal of the American Society for
Information Science and Technology, 58(13), 1960–1976.
Garfield, E. (1979). Citation indexing: Its theory and application in science,
technology, and humanities. New York, NY: Wiley.
Garner, J., Porter, A.L., Borrego, M., Tran, E., & Teutonico, R. (2013).
Facilitating social and natural science cross-disciplinarity: Assessing the
human and social dynamics program. Research Evaluation, 22(2), 134–
144.
Gedik, Y. (2012). Geographical localisation of knowledge spillovers by
Australian patent citations. Economic Papers, 31(2), 173–181.
Georgia Tech (2011). Global Nanotechnology Datasets [see Porter et al.,
2008]. Underlying data retrieved from: http://thomsonreuters.com/
thomson-reuters-web-of-science/.
Glänzel, W., & Meyer, M. (2003). Patents cited in the scientific literature:
An exploratory study of “reverse” citation relations. Scientometrics,
58(2), 415–428.
Glänzel, W., & Thijs, B. (2004a). The influence of author self-citations on
bibliometric macro indicators. Scientometrics, 59(3), 281–310.
Glänzel, W., & Thijs, B. (2004b). Does coauthorship inflate the share of
self-citations? Scientometrics, 61(3), 395–404.
Guan, J.C., & Ma, N. (2007). China’s emerging presence in nanoscience
and nanotechnology—A comparative bibliometric study of several nano-
science “giants. Research Policy, 36(6), 880–886.
Gupta, H.M., Campanha, J.R., & Pesce, R. (2005). Power-law distributions
for the citation index of scientific publications and scientists. Brazilian
Journal of Physics, 35(4a), 981–986.
Hall, B.H., Jaffe, A.E., & Trajtenberg, M. (2001). The NBER patent cita-
tions data file: Lessons, insights and methodological tools (No. w8498).
National Bureau of Economic Research, DOI: 10.3386/w8498.
Hu, G., Carley, S., & Tang, L. (2012). Visualizing nanotechnology research
in Canada: Evidence from publication activities, 1990–2009. The Journal
of Technology Transfer, 37(4), 550–562.
Hu, X., & Rousseau, R. (2013). Are Chinese nanoscience citation curves
converging towards their American counterparts? Malaysian Journal of
Library and Information Science, 18(3), 49–56.
Huang, C., Notten, A., & Rasters, N. (2011). Nanoscience and technology
publications and patents: A review of social science studies and search
strategies. The Journal of Technology Transfer, 36(2), 145–172.
Hvistendahl, M. (2013). China’s publication bazaar. Science, 342(6167),
1035–1039.
Jaffe, A.B., Trajtenberg, M., & Henderson, R. (1993). Georgraphic local-
ization of knowledge spillovers as evidenced by patent citations. The
Quarterly Journal of Economics, 108(3), 577–598.
JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGY—September 2015 1931
DOI: 10.1002/asi

Jin, B., Rousseau, R., Suttmeier, R.P., & Cao, C. (2007). The role of ethnic
ties in international collaboration: The Overseas Chinese Phenomenon.
In D. Torres-Salinas & H.F. Moed (Eds.), Proceedings of the Interna-
tional Society for Scientometrics and Informetrics (ISSI) 2007 (pp. 427–
436). Stockholm: Karolinska University Press.
King, D. (2004). The scientific impact of nations. Nature, 430(6997), 311–
316.
Kostoff, R. (2012). China/USA nanotechnology research output
comparison—2011 update. Technological Forecasting and Social
Change, 79(5), 986–990.
Kuhn, T. (1970). The structure of scientific revolutions (2nd ed.). Chicago,
IL: University of Chicago Press.
Lawrence, P.A. (2003). The politics of publication. Nature, 422(6929),
259–261.
Lawrence, P.A. (2007). The mismeasurement of science. Current Biology,
17(15), 583–585.
Liang, L., Rousseau, R., & Zhong, Z. (2013). Non-English journals
and papers in physics: Bias in citations. Scientometrics, 95(1), 333–
350.
Lin, M.-W., & Zhang, J. (2007). Language trends in nanoscience and
technology: The case of Chinese-language publications. Scientometrics,
70(3), 555–564.
Lotka, A.J. (1926). The frequency distribution of scientific productivity.
Journal of the Washington Academy of Sciences, 16(12), 317–323.
Moed, H.F. (2002). Measuring China’s research performance using the
Science Citation Index. Scientometrics, 53(3), 281–296.
Moiwo, J.P., & Tao, F.L. (2013). The changing dynamics in citation index
publication position China in a race with the USA for global leadership.
Scientometrics, 95(3), 1031–1050.
Noyons, E.C.M., Moed, H.F., & Luwel, M. (1999). Combining mapping
and citation analysis for evaluative bibliometric purposes:A bibliometric
study. Journal of the American Society for Information Science, 50(2),
115–131.
Organisation for Economic Co-operation and Development (OECD).
(2007). OECD Reviews of Innovation Policy: China. Paris.
Onodera, N., Iwasawa, M., Midorikawa, N., Yoshikane, F., Amano, K.,
Ootani, Y., et al. (2011). A Method for eliminating articles by homony-
mous authors from the large number of articles retrieved by author
search. Journal of the American Society for Information Science and
Technology, 62(4), 677–690.
Opsahl, T., Colizza, V., Panzarasa, P., & Ramasco, A.J.J. (2008). Promi-
nence and control: The weighted rich-club effect. Physical Review
Letters, 101(16), 168702–168705.
Papke, L.E., & Wooldridge, J.M. (1996). Econometric methods for frac-
tional response variables with an application to 401(k) plan participation
rates. Journal of Applied Econometrics, 11(6), 619–632.
Persson, O., Glänzel, W., & Danell, R. (2004). Inflationary bibliometric
values: The role of scientific collaboration and the need for relative
indicators in evaluative studies. Scientometrics, 60(3), 421–432.
Phelan, T.J. (1999). A compendium of issues for citation analysis. Scien-
tometrics, 45(1), 117–136.
Pichappan, P., & Sarasvady, S. (2002). The influence of author self-
citations on bibliometric macro indicators. Scientometrics, 54(2), 285–
290.
Popper, K.R. (1963). Conjectures and refutations: The growth of scientific
knowledge. London, England: Routledge.
Porter, A., Youtie, J., Shapira, P., & Schoeneck, D. (2008). Refining search
terms for nanotechnology. Journal of Nanoparticle Research, 10(5), 715–
728.
Porter, A.L., & Rafols, I. (2009). Is science becoming more interdisciplin-
ary? Measuring and mapping six research fields over time. Scientomet-
rics, 81(3), 719–745.
Porter, A.L., & Youtie, J. (2009). Where does nanotechnology belong in the
map of science? Nature-Nanotechnology, 4(9), 534–536.
Redner, S. (1998). How popular is your paper? An empirical study of the
citation distribution. European Physical Journal B, 4(2), 131–134.
Rehn, C., & Kronman, U. (2008).Bibliometric handbook for Karolinska
Institutet. Retrieved from http://ki.se/content/1/c6/01/79/31/bibliometric
_handbook_karolinska_institutet_v1.05.pdf
Ren, S.L., & Rousseau, R. (2002). International visibility of Chinese sci-
entific journals. Scientometrics, 53(3), 389–405.
Shapira, P., & Wang, J. (2010). Follow the money. Nature, 468(7324),
627–628.
Snyder, H., & Bonzi, S. (1998). Patterns of self-citation across disciplines
(1980–1989). Journal of Information Science, 24(6), 431–435.
STATA Technical Bulletin. (July 1999). STB 50. 1-40. Retrieved from
www.stata.com/products/stb/journals/stb50.pdf
Suttmeier, R.P., Cao, C., & Simon, D.F. (2006). China’s innovation chal-
lenge and the remaking of the Chinese Academy of Sciences.. Innova-
tions: Technology, Governance, Globalization, 1(3), 78–97.
Tang, L. (2013). Does “birds of a feather flock together” matter—Evidence
from a longitudinal study on US-China scientific collaboration. Journal
of Informetrics, 7(2), 330–344.
Tang, L., & Shapira, P. (2011a). China–US scientific collaboration in nano-
technology: Patterns and dynamics. Scientometrics, 88(1), 1–16.
Tang, L., & Shapira, P. (2011b). Regional development and interregional
collaboration in the growth of nanotechnology research in China. Scien-
tometrics, 86(2), 299–315.
Tang, L., & Shapira, P. (2012). Effects of international collaboration and
knowledge moderation on China’s nanotechnology research impacts.
Journal of Technology Management in China, 7(1), 94–110.
Tang, L., & Walsh, J.P. (2010). Bibliometric fingerprints: Name disambigu-
ation based on approximate structure equivalence of cognitive maps.
Scientometrics, 84(3), 763–784.
Tang, X.M. (2007). Indigenous innovation through intellectual property
protection. China Science Foundation, 21(4), 210–212, (in Chinese).
Taylor, M., Perakakis, P., & Trachana, V. (2008). The siege of science.
Ethics in Science and Environmental Politics, 8(1), 17–40.
Van Raan, A.F.J. (1998). The influence of international collaboration on the
impact of research results—Some simple mathematical considerations
concerning the role of self-citations. Scientometrics, 42(3), 423–428.
Wang, P.X. (1992). Strengthen the redirection role of science funding.
Advance in Earth Sciences, 7(4), 11–12, (in Chinese).
Youtie, J., Shapira, P., & Porter, A.L. (2008). Nanotechnology publications
and citations by leading countries and blocs. Journal of Nanoparticle
Research, 10(6), 981–986.
Zhang, H., Patton, D., & Kenney, M. (2013). Building global-class univer-
sities: Assessing the impact of the 985 Project. Research Policy, 42(3),
765–775.
Zhou, P., & Leydesdorff, L. (2007). The citation impacts and citation
environments of Chinese journals in mathematics. Scientometrics, 72(2),
185–200.
1932 JOURNAL OF THE ASSOCIATION FOR INFORMATION SCIENCE AND TECHNOLOGY—September 2015
DOI: 10.1002/asi