You are on page 1of 10

Article

Journal of Information Science

A study of interdisciplinarity in 1–10


© The Author(s) 2011
Reprints and permission: sagepub.
information science: using direct co.uk/journalsPermissions.nav
DOI: 10.1177/0165551511407141

citation and co-authorship analysis jis.sagepub.com

Mu-Hsuan Huang
Department of Library and Information Science, National Taiwan University, ROC

Yu-Wei Chang
Department of Library and Information Science, Fu Jen Catholic University, ROC

Abstract
This study uses two bibliometric methods, direct citation and co-authorship, to investigate the interdisciplinary changes in information
sciences during 1978–2007. The disciplines of references and co-authors from five information science journals were analysed.
Furthermore, Brillouin’s Index was adopted to measure the degree of interdisciplinarity. The study revealed that information science
researchers have cited the publications of library and information science (LIS) most frequently. The co-authors of information
science articles are also primarily from the discipline of LIS, but the percentage of references to LIS is much higher. This indicates that
information science researchers mainly rely on publications in LIS, and they often produce scientific papers with researchers from
LIS. The discipline rankings generated by direct citation and co-authorship show a significant consistency via Spearman’s correlation
coefficient test. The interdisciplinary degree of information science has displayed growth. In particular, the degree of interdisciplinarity
for co-authors has grown.

Keywords
co-authorship; direct citation; information science; interdisciplinary analysis

1. Introduction
Information science has been regarded as an interdisciplinary discipline. Borko [1, p. 3] has indicated that ‘information
science is an interdisciplinary science derived from and related to such fields as mathematics, logic, linguistics, psychol-
ogy, computer technology, operations research, the graphic arts, communications, library science, management, and other
similar fields’. When comparing these fields with subfields of information science such as bibliometrics, information
retrieval, scientific communication, webmetrics and patent analysis [2–4], most can be identified as influencing informa-
tion science. Some researchers have attempted to probe the relationship between information science and other disci-
plines. These studies suggest that social sciences may have limited influence on information science. For example, Small
[5] examined the place of information science within social sciences, and their internal and external connection. The
results have shown that information science is isolated from social sciences. Al-Sabbagh [6] analysed references in the
articles published in the Journal of the American Society for Information Science (JASIS) from 1975 to 1985 and found
that the cited publications came from a variety of disciplines. The publications on computer science, library and informa-
tion science (LIS), and general science were cited most frequently. In particular, the percentage of computer science cita-
tions has increased, while the percentage of LIS citations has decreased. Tsay [7] also explored the relationship between
information science and other disciplines via citation analysis and examined the publications in Journal of the American

Corresponding author:
Yu-Wei Chang, Department of Library and Information Science, Fu Jen Catholic University, No. 510, Zhongzheng Rd., Xinzhuang District, New
Taipei City, 24205 Taiwan, ROC.
Email: ywchang@blue.lins.fju.edu.tw

Downloaded from jis.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016


Huang and Chang 2

Society for Information Science and Technology (JASIST) in 1980, 1985, 1990, 1995, 2000, and 2004. She found that the
cited journals came from various disciplines, about half of which were in the field of LIS, followed by science (22.7%)
and social sciences (6.3%). Borgman and Rice [8] investigated the citation relationship between information science and
communication. They found little convergence between the two disciplines.
In fact, there was more research that studied the sources of references from publications on information science and
library science. The hybrid results were slightly different from that only based on the journals of information science. For
example, Buttlar’s [9] citation analysis of 61 LIS dissertations revealed that LIS researchers frequently cited sources in
LIS (49.54%), education (11.45%), computer science (5.72%), health/medicine (3.79%), sociology (3.79%), communica-
tion (1.96%), and business (1.90%). Chikate and Patil [10] analysed references in 27 LIS dissertations and found that the
main source disciplines included LIS, science, medical science, economics, literature, computer science, meteorology,
education, and sociology. Chen and Liang [11] explored disciplinary interflow of LIS studies in Taiwan and discovered
that LIS researchers frequently cited sources in their own discipline (78.12%), followed by social science (13.41%), engi-
neering and technology (1.57%), and history (1.27%). In addition, Sarcacevic [12] pointed out that the interdisciplinary
evolution of information science is far from over. However, no study has examined whether the interdisciplinarity of
information science increases over time.
Although citing literature from different disciplines and the co-authorship of researchers from different disciplines are
common ways of interdisciplinary information transfer [13], prior interdisciplinary research has showed that direct cita-
tion analysis for the disciplinary characteristics of references is the most ordinary technique in bibliometric methods, and
it has been utilized in many fields. However, cited references are determined completely by the authors. In most cases, the
citing authors do not interact with the cited authors. In contrast, co-authorship naturally involves a more collaborative and
interactive relationship than a citing–cited relationship. The relationship of co-authorship implies that at least two authors
are willing to create a paper together. Before publishing their co-authored paper, they may discuss and exchange ideas.
Citing behaviour reveals the information flow across disciplines, while the co-authorship relationship presents the
social network of researchers across disciplines. Owing to the more active nature of the co-authorship relationship, it can
be expected that the number of disciplines for the co-authorship relationship is fewer than for the citing–cited relationship.
With the increasingly prevalence of scientific collaboration, the co-authored paper has become a common type of research
output. Related studies have shown that there is increasing growth in co-authored papers across all disciplines. In informa-
tion science, for example, Lipetz [14] has analysed JASIS, and discovered that its rate of co-authored articles was 4.8% in
1955. But the percentage of co-authored articles reached 32.4% in 1995, with a great range of growth. Chua and Yang [15]
studied multi-disciplinarity in information science during 1988–2007 and found that cross-disciplinary co-author pairs
have become more prevalent. Thus, co-authorship analysis can be properly applied to interdisciplinary studies. However,
quite a few studies used co-authorship to reveal the interdisciplinary nature. Qiu [16] has analysed co-authored articles in
24 LIS journals and only indicated that LIS researchers collaborated with researchers from 10 disciplines outside LIS.
Since most previous studies on interdisciplinarity only used a bibliometric method, especially direct citation, we did
not find any research that discussed the results analysed by different bibliometric methods. At the same time, a question
arises: are the results of interdisciplinarity analysed by direct citation and co-authorship analysis different? Moreover, few
studies so far have examined the interdisciplinary characteristics of information science through more journals. For these
reasons, this study aims to use two bibliometric methods, direct citation analysis and co-authorship analysis, to investigate
the interdisciplinary characteristics in information science, and to compare the interdisciplinary results analysed by
the two bibliometric methods. This study analyses a long-term data set covering the period 1978–2007 for examining
changes in interdisciplinarity and picturing the interdisciplinary trend.
To address these issues, the present study was designed to address these research questions:

1. What disciplines were cited by information science researchers?


2. From which disciplines did the co-authors of information science articles come?
3. What are the differences between the results analysed by direct citation and co-authorship analysis?
4. Does the degree of interdisciplinary in information science increase?

2. Methodology
This study adopted direct citation and co-authorship to probe the interdisciplinary characteristics and changes in informa-
tion science. The disciplines that influenced information science were identified by analysing the discipline distribution of
references and co-authors in direct citation and co-authorship. Brillouin’s Index was used to measure the degree of inter-
disciplinarity of information science for every year in the study period to clearly identify interdisciplinary changes and
trends.

Journal of Information Science, 2011, pp. 1–10 © The Author(s), DOI: 10.1177/0165551511407141

Downloaded from jis.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016


Huang and Chang 3

Table 1.  Five information science journals

Journal title No. of articles No. of information No. of information science


science articles articles with sampling 20%

Information Processing & Management 1466 329 65


JASIST 2224 828 165
Journal of Documentation 550 255 51
Journal of Information Science 1072 484 96
Scientometrics 1966 1966 393
Total 7278 3862 770

2.1. Journal selection


References and co-authors’ affiliations analysed in this study were collected from five information science journals. The
target journals were selected from the journals of information science and library science covered by Journal Citation
Reports of 2008. The impact factors of selected journals had to be in the top ranks, and annual journals were excluded.
Next, the scope of journals had to emphasize ‘information science’. Five selected information science journals are listed
in Table 1.

2.2. Data collection


In this study, only research articles published between 1978 and 2007 were collected. Selected articles had to be
categorized as ‘articles’ by SSCI. Moreover, to reduce the chances of including non-information science articles in
the samples, the authors checked the selected journals and excluded a total of 3412 articles from the discipline of
computer science. Next, 3862 information science articles were sorted first by publication year, and then by journal
name. Twenty per cent of articles were selected via systematic sampling. A total of 770 selected articles were anal-
ysed in this study (see Table 1). There were 17,985 references in the 770 information science articles. Among them,
16,170 references to books and journals were identified for analysis, whereas other categories of references (1815,
or 10.09%) were excluded.
The discipline of each reference was determined by the retrieved classification number, a classification of 30 disci-
plines that is based on the Library of Congress classification. If the classification numbers of references could not be
identified from the Library of Congress, the references were excluded. After the exclusion of 1095 references, there were
a total of 15,075 references (83.82% of 17,985 references) for interdisciplinary analysis. Among them, the classification
numbers of two journals, Research Policy and Scientometrics, were categorized as library and information science instead
of general science.
Within the 770 information science articles with 1440 authors, there were in total 394 co-authored articles (51.17%)
with 1064 authors (73.89% of 1,440 authors). The disciplines for co-authors were judged according to the institutional
affiliation listed in articles and were mapped to the same classification scheme for the disciplines of references. Since the
institutional affiliation data in Web of Science were not recorded in full format, they can only be identified via the original
documents. A total of 102 authors were excluded because their institutional affiliation data were not provided or complete.
As a result, 962 co-authors were identified for interdisciplinary analyses.

2.3. Interdisciplinary indicator


This study adopted Brillouin’s Index as a measure of the annual degree of interdisciplinarity of information science. When
the value is higher, a greater degree of interdisciplinarity is present. Its formula is:

log N !− ∑ (log ni !)
N

where N represents for ‘the total number of the entire references’ and ni represents the quantity of references in i discipline.
When evaluating the interdisciplinary degree of co-authors, N means the whole number of co-authors and ni means the
number of co-authors in i discipline.

Journal of Information Science, 2011, pp. 1–10 © The Author(s), DOI: 10.1177/0165551511407141

Downloaded from jis.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016


Huang and Chang 4

3. Results
As shown in Table 2, during the period 1978–2007, the total numbers of disciplines for reference sources and co-authors
were 29 and 23, respectively. Among 30 disciplines, only the publications in recreation/sport were not cited by informa-
tion science researchers. As for the co-authors of information science articles, seven excluded disciplines were arts, bot-
any, history, law, literature, recreation/sport, and zoology. This reflected that the range of disciplines in cited sources is
broader than in co-authors.

3.1. Distribution of disciplines


3.1.1. Top five disciplines.  Table 3 shows the proportion of the top five disciplines in reference sources and co-authors in
information science. There is a high consistency of the findings for top disciplines generated by the two bibliometric
methods. The top three and fifth-ranked disciplines are the same in the two bibliometric methods. The first discipline is
LIS, followed by general science, computer science, and medicine. Furthermore, the fourth and sixth disciplines for direct
citation analysis are engineering and business/management, respectively. The converse results were found by analysis of
co-authorship. The publications in natural sciences, including computer science, general science, and medicine, were most
cited by information science researchers, which is consistent with the findings of Al-Sabbagh [6] and Tsay [7]. As for
Qiu’s study [16], the results of co-authored articles in 24 LIS journal published between 1972 and 1991 showed that LIS
researchers were most likely to collaborate with researchers in computer science, followed by business/management,
medicine, and engineering. This was also consistent with our findings.
Although there is a quite high consistency in the ranks of the top disciplines generated by direct citation and co-author-
ship, significant differences were found among the percentages of each discipline. First, the percentage of references to
LIS (49.50%) is much higher than the percentage of co-authors from LIS-related institutes (29.21%). The percentage of
co-authors’ affiliations classified in the field of general science (21.10%) is much higher than that of references to general
science (11.75%). Second, the disparity of proportion between the two most cited disciplines in references reached

Table 2.  Disciplines and frequency of direct citation analysis and co-authorship analysis

No. Discipline Direct citation Co-authorship

 1 Agriculture 21 7
 2 Anthropology 38 1
 3 Arts 29 0
 4 Astronomy 24 1
 5 Biology 96 4
 6 Botany 24 0
 7 Business/management 547 85
 8 Chemistry 71 24
 9 Computer science 834 91
10 Earth science 38 8
11 Economics 507 5
12 Education 241 28
13 General science 1771 203
14 General social science 338 23
15 History 42 0
16 Law 75 0
17 LIS 7462 281
18 Linguistics and language 246 8
19 Literature 54 0
20 Mathematics 199 10
21 Medicine 575 80
22 Military science 5 2
23 Philosophy and religion 134 7
24 Physics 176 19
25 Political science 42 7
26 Psychology 364 18
27 Recreation/sport 0 0
28 Sociology 454 9
29 Technology 662 41
30 Zoology 6 0

Journal of Information Science, 2011, pp. 1–10 © The Author(s), DOI: 10.1177/0165551511407141

Downloaded from jis.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016


Huang and Chang 5

Table 3.  Distribution of top five source disciplines by direct citation and co-authorship

Discipline Direct citation Co-authorship

Percentage Rank Percentage Rank

LIS 49.50% 1 29.21% 1


General science 11.75% 2 21.10% 2
Computer science 5.53% 3 9.46% 3
Engineering 4.39% 4 4.26% 6
Medicine 3.81% 5 8.32% 5
Business/management 3.63% 6 8.84% 4
Accumulated percentage of top five 74.98% 76.93%
Disciplines(marked in bold)

37.75%, which is much greater than that between the two main disciplines that co-authors came from (8.11%). This has
revealed that the researchers of information science rely highly on the literature of LIS, while co-authors of information
science articles concentrate on LIS and general science. Third, through direct citation and co-authorship, the percentage
of the top five disciplines amounts to three-quarters of the total percentage. Among the 30 disciplines, only a few have
comparatively greater influence on and interact with information science (see Table 3).

3.1.2. Disciplines with percentage over 1% (excluding the top five disciplines).  Table 4 displays the disciplines with
percentage over 1% but not including the top five disciplines; the proportions of each discipline are close in the two
bibliometric methods. Five disciplines, including psychology, general social science, education, mathematics, and
physics, are the same in two bibliometric methods. As for different disciplines, economics (7th rank) was one of the
top 25% cited disciplines, but only a few (19th rank, 0.52%) co-authors are from that discipline. Similarly, 2.49%
(8th rank) of co-authors are from the discipline of chemistry, but only 0.47% (18th rank) of references are from
chemistry.

3.1.3. Disciplines with percentage under 1%.  As indicated in Table 5, the quantity of disciplines under 1% but not reaching 0%
is high. There were 15 and 11 disciplines in the results of direct citation analysis and co-authorship analysis, respectively.
Information science researchers cite resources from the disciplines of literature, history, arts, law, botany, and zoology, but no
co-authors are from those disciplines. Four disciplines with low percentage generated by direct citation analysis including
arts, history, law, and political science, were consistent with prior studies [6–8]. Six disciplines without co-authors were sel-
dom cited. The insignificantly low share of each discipline leads to the accumulation rates for disciplines in this part at only
4.64 and 6.14%, respectively. This suggests that these disciplines have fairly limited influence on information science and
interactions with information science.

Table 4.  Disciplines with percentage over 1% by direct citation and co-authorship (excluding the top five)

Discipline Direct citation Co-authorship

Percentage Rank Percentage Rank

Engineering 4.39% 4 4.26% 6


Business/management 3.63% 6 8.84% 4
Economics 3.36% 7 0.52% 19
Sociology 3.01% 8 0.94% 13
Psychology 2.41% 9 1.87% 11
General social science 2.24% 10 2.39% 9
Linguistics and language 1.63% 11 0.83% 14
Education 1.60% 12 2.91% 7
Mathematics 1.32% 13 1.04% 12
Physics 1.17% 14 1.98% 10
Chemistry 0.47% 18 2.49% 8
Accumulated percentage of disciplines with 20.3% 16.94%
Percentage over 1% but not including the top
five disciplines (marked in bold)

Journal of Information Science, 2011, pp. 1–10 © The Author(s), DOI: 10.1177/0165551511407141

Downloaded from jis.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016


Huang and Chang 6

Table 5.  Disciplines with percentage under 1% by direct citation and co-authorship

Discipline Direct citation Co-authorship

Percentage Rank Percentage Rank

Economics 3.36% 7 0.52% 19


Sociology 3.01% 8 0.94% 13
Linguistics and language 1.63% 11 0.83% 14
Philosophy and religion 0.89% 15 0.73% 16
Biology 0.64% 16 0.42% 20
Law 0.50% 17 – –
Chemistry 0.47% 18 2.49% 8
Literature 0.36% 19 – –
History 0.28% 20 – –
Political science 0.28% 20 0.73% 16
Anthropology 0.25% 22 0.10% 22
Earth science 0.25% 22 0.83% 14
Arts 0.19% 24 – –
Astronomy 0.16% 25 0.10% 22
Botany 0.16% 25 – –
Agriculture 0.14% 27 0.73% 16
Zoology 0.04% 28 – –
Military science 0.03% 29 0.21% 21
Accumulated percentage of disciplines under 4.64% 6.14%
1% (marked in bold)

3.2.The level of consistency of the discipline rankings generated by two bibliometric methods
The discipline rankings generated by direct citation analysis and co-authorship analysis were somewhat alike. To identify
the differences between the two rankings, Spearman’s correlation coefficient for rank test was used. The result was significant,
with a value of 0.813 (p = 0.01). This means that there was no statistically significant difference between both discipline
rankings for references and co-authors in information science.

3.3. Changes in interdisciplinarity


Based on direct citation and co-authorship analyses for investigating the disciplines influencing information science, only
a few have high rates and show considerable influence. After examining the annual changes in discipline rankings, no
palpable shifts can be discovered. This means that the top disciplines are also those that influence the development of
information science.
As shown in Figures 1 and 2, although the top three disciplines generated by direct citation and co-authorship are the
same, their annual rate changes and tendencies are different. First, the LIS publications were cited most frequently by
information science researchers each year, and the percentage of LIS had a slightly increasing tendency. By contrast, the
percentage of co-authors in LIS was not always the highest, and there was a decreasing tendency. Second, the general
science publications received more citations than computer science publications, except in 1987, 1997, 1999, and 2002.
The percentage of general science had a decreasing tendency. In contrast, the percentage of co-authors in general science
had an obvious fluctuation. It was the first-ranked discipline in 14 years, with a slightly increasing tendency during the
period 1978–2007. Third, there was a rising tendency in computer science as more computer science publications were
cited by information science researchers. By contrast, the percentage of co-authors in computer science has slightly
decreased. In particular, there are 10 discontinuous years in which no co-author from the discipline of computer science
can be found. Based on these findings, there are reverse results for the tendency of the top three disciplines generated by
direct citation analysis and co-authorship analysis. However, the reverse tendency generated by two bibliometric
methods is small.

3.4.The annual number of disciplines


It can be observed from Figure 3 that the annual number of disciplines in reference sources was much higher than
in co-authors. The annual discipline number ranged from 11 to 27. The number of disciplines has obviously

Journal of Information Science, 2011, pp. 1–10 © The Author(s), DOI: 10.1177/0165551511407141

Downloaded from jis.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016


Huang and Chang 7

%
LIS general science computer science
70

60

50

40

30

20

10

0
year
84

88

06
82
78

92

96

00

04
80

86

90

94

98

02
19

19

20
19
19

19

19

20

20
19

19

19

19

19

20
Figure 1.  The annual rates of the top three disciplines generated by direct citation analysis.

%
70
LIS
general science
60
computer science

50

40

30

20

10

0
year
02

04

06
90

94

98

00
8

84

86

88

9
7

19

19

19

19

20

20

20

20
19

19

19

19

19

19

19

Figure 2.  The annual rates of the top three disciplines generated by co-authorship analysis.

increased from 11 disciplines in 1978 to 26 disciplines in 2007, with some variation. By contrast, the annual disci-
pline number for co-authors ranges from 3 to 14. The interdisciplinary degree of co-authors is significantly lower
than that of reference sources. However, the number of disciplines has an increasing tendency over time for both
direct citation and co-authorship. This indicates that information science researchers tended to cite sources from
more disciplines outside LIS, and more researchers from different disciplines became co-authors of information
science articles over time.

Journal of Information Science, 2011, pp. 1–10 © The Author(s), DOI: 10.1177/0165551511407141

Downloaded from jis.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016


Huang and Chang 8

No. of disciplines direct citation co-authorship


30

25

20

15

10

0
78

80

82

84

86

88

90

92

94

96

98

00

02

04

06
year
19

19

19

19

19

19

19

19

19

19

19

20

20

20

20
Figure 3.  The annual number of disciplines analysed by two bibliometric methods.

interdisciplinary direct citation co-authorship


degree
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

year
78

80

82

84

86

88

90

92

94

96

98

00

02

04

06
19

19

19

19

19

19

19

19

19

19

19

20

20

20

20

Figure 4.  Annual change for the interdisciplinary value of information science.

3.5. Degree of interdisciplinarity


Based on the distribution among disciplines of references and co-authors analysed by direct citation and co-authorship,
the degree of interdisciplinarity can be calculated using Brillouin’s Index. The larger the value, the higher the degree of
interdisciplinarity. As seen in Figure 4, the annual degree of interdisciplinarity based on direct citation analysis was higher
than the annual degree of interdisciplinarity based on co-authorship analysis, except in 2005 and 2007. The trends for
interdisciplinary degree based on direct citation analysis and co-authorship analysis are increasing. There is in particular
a greater range of the interdisciplinary value in co-authorship than in cited literature. This reflects that information science
is an interdisciplinary discipline, and its level of interdisciplinarity has been increased.

4. Discussion
A study on both citing literature and co-authorship can show the interdisciplinary natures in a specific discipline. However,
the two different behaviours reflect different levels of interdisciplinarity. Basically, co-authorship develops based on inter-
personal relationships, namely social networks. Before co-publishing, co-authors have to discuss or communicate ideas
with each other. That means co-authorship is different from the decisions authors make themselves about which literature
to cite. More interactive natures exist in co-authorship. In order to compare the differences between the two perspectives
in citing literature and co-authorship, two approaches of direct citation and co-authorship were used to analyse information
science articles. According to the definitions of direct citation and co-authorship, references in all articles were used for
direct citation analysis and authors in co-authored articles were used for co-authorship analysis. So, the figures from the

Journal of Information Science, 2011, pp. 1–10 © The Author(s), DOI: 10.1177/0165551511407141

Downloaded from jis.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016


Huang and Chang 9

two approaches can not be compared directly due to different categories of data. This study focuses on the differences of
the interdisciplinary nature between direct citation and co-authorship. Moreover, this study attempts to identify if the
co-authorship analysis that deals with fewer co-author data can replace direct citation analysis that analyses a lot of
reference data.
The results show that certain similarities between the interdisciplinary nature from direct citation and co-authorship.
First, the study confirms that information science is an interdisciplinary discipline based on the variety of disciplines
in reference sources and co-authors’ affiliations. The growing number of disciplines shows that information science
researchers tend to cite publications from more disciplines and co-authors of information science articles also tend to
come from more disciplines. The ever-increasing interdisciplinary degree of information science has also been
identified. However, the number of publications is much larger than the number of researchers. There are more
information choices for researchers to cite as resources than there are collaborators. Thus, it can be easily understood
why the disciplinary scope of references is larger than that of co-authors. In addition, the level of interdisciplinarity
in co-authorship increased more than in cited publications. This phenomenon may be caused by the trend of scientific
collaboration [17, 18]. As more countries regard collaboration as a scientific policy, the interactions across disciplines
will be fostered directly [19].
Second, the top three and fifth-ranked disciplines generated by direct citation and co-authorship are the same. The top
disciplines in this study also greatly resemble the past research [6–10, 16]. In general, there is a relevant subject relationship
between cited publications and citing papers. The number of citations can reflect the relationship between disciplines.
The publications in discipline A are most cited by researchers in discipline B, indicating that B discipline highly relies on
A discipline. Based on this logic, the main disciplines in reference sources and co-authors may be the same. That is to say,
researchers in the highly related disciplines are the priority for co-authors outside their own disciplines. The inference is
supported by the results of this study.
The main different results between two bibliometric methods are the percentages of top disciplines and their annual rate
changes. The researchers of information science mainly cite LIS publications, and the authors of co-authored papers in
information science also concentrate on LIS. However, the percentage of LIS in references (49.50%) is much higher than
in co-authorship (29.21%). This means that the natures of literature citation and co-authorship are different. From prior
citation studies, it has been found that each discipline mainly relies on its own publications. Similarly, co-authors com-
monly come from the same discipline. However, the lower percentage of LIS of co-authors in information science reflects
that the interdisciplinarity of co-authored articles on information science is higher because more researchers outside LIS
publish more articles on the subject. In addition, scientific collaboration has been a common way to produce knowledge
across disciplines, especially in the natural sciences [20]. The percentage of co-authored papers in the study (51.17%),
which is higher than the results of disciplines in social sciences [20, 21], shows that information science is closer to natural
sciences. The percentage of co-authors from institutes of natural sciences (50.73%) is also much higher than that from
institutes of social sciences and humanities, excluding LIS (20.06%).
Interdisciplinarity is a multifaceted issue and can be presented using different approaches. However, the bibliometric
methods applied to interdisciplinary research seem to be still limited to direct citation. So the results of this study may
have contributed to a greater understand of co-authorship in interdisciplinary research and suggest that different research
methods may lead to different results.

5. Conclusion
This study has confirmed that information science is an interdisciplinary discipline. In the results of direct citation analy-
sis, information science researchers cited publications from 29 disciplines. Information science researchers most
frequently cited publications of LIS, followed by general science, computer science, engineering, and medicine. The find-
ings are consistent with prior studies. As for the results of co-authorship analysis, co-authors of information science arti-
cles were from 23 disciplines. The co-authors were mainly from LIS, followed by general science, computer science,
business/management, and medicine. It is evident that information science has a close relationship with natural sciences.
Among the diversity of disciplines, information science researchers rely heavily on LIS publications, and most of their
co-authors are also from LIS. However, it should be noted that the percentage share of reference sources in LIS is appar-
ently larger than that of co-authors in LIS. LIS sources were cited most frequently each year, while co-authors were not
primarily from LIS each year. Since no discipline has seen significant changes in its percentages and rankings during the
30-year period, it is evident that the top disciplines are the ones that have more influence on the interdisciplinary develop-
ment of information science.
Based on analyses of direct citation and co-authorship, although there was a significant consistency in the discipline
rankings opposite results were obtained in the trends of percentages of the top three disciplines in the two bibliometric

Journal of Information Science, 2011, pp. 1–10 © The Author(s), DOI: 10.1177/0165551511407141

Downloaded from jis.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016


Huang and Chang 10

methods. It can be observed that information science researchers tended to cite more publications from LIS and computer
science and more co-authors were from the discipline of general science.
The degree of interdisciplinarity in reference sources and co-authors had exhibited an increasing trend during 1978–
2007, meaning that information science researchers tend to cite more publications from outside the discipline of LIS and
to write collaborative papers with researchers in diverse disciplines. In particular, the growing trend in co-authorship is
more obvious than that in cited publications.

References
  [1] Borko H. Information science: what is it? American Documentation 1968; 19: 3–5.
  [2] Janssens F, Glänzel W, Moor BD. A hybrid mapping of information science. Scientometrics 2008; 75: 607–631.
  [3] White HD, Griffith BD. Author cocitation: a literature measure of intellectual structure. Journal of the American Society for
Information Science 1981; 32: 163–171.
  [4] White HD, McCain KW. Visualizing a discipline: an author co-citation analysis of information science, 1972–1995. Journal of
the American Society for Information Science 1998; 49: 327–355.
  [5] Small H. The relationship of information science to the social sciences: a co-citation analysis. Information Processing and
Management 1981; 17: 39–50.
  [6] Al-Sabbagh IA. The evolution of the interdisciplinarity of information science: a bibliometric study (Unpublished manuscript,
Florida State University, Tallahassee, FL, 1987)
  [7] Tsay MY. Journal bibliometric analsis: a case study on the JASIST. Malaysian Journal of Library & Information Science 2008;
13: 121–139.
  [8] Borgman CL, Rice RE. The convergence of information science and communication: a bibliometric analysis. Journal of the
American Society for Information Science 1992; 43: 397–411.
  [9] Buttlar L. Information sources in library and information science doctoral research. Library & Information Science Research
1999; 21: 227–245.
[10] Chikate RV, Patil SK. Citation analysis of theses in library and information science submitted to University of Pune: a pilot study.
Library Philosophy and Practice 2008. Available at: http://unllib.unl.edu/LPP/chikate-patil.html (accessed 23 July 2010)
[11] Chen KH, Liang CF. Disciplinary interflow of library and information science in Taiwan. Journal of Library and Information
Studies 2004; 2: 31–55.
[12] Saracevic T. Information science. Journal of the American Society for Information Science 1999; 50: 1051–1063.
[13] Pierce SJ. Boundary crossing in research literatures as a means of interdisciplinary information transfer. Journal of the American
Society for Information Science 1999; 50: 271–279.
[14] Lipetz BA. Aspects of JASIS authorship through five decades. Journal of the American Society for Information Science 1999; 50:
994–1003.
[15] Chua AYK, Yang CC. The shift towards multi-disciplinarity in information science. Journal of the American Society for
Information Science and Technology 2008; 59: 2156–2170.
[16] Qiu L. A study of interdisciplinary research collaboration. Research Evaluation 1992; 2: 169–175.
[17] Leydesdorff L, Wagner C. International collaboration in science and the formation of a core group. Journal of Informetrics 2008;
2: 317–325.
[18] National Science Board. Science and Engineering Indicators 2008. Available at: http://www.nsf.gov/statistics/seind08/(accessed
21 February 2011)
[19] Wagner CS, Leydesdorff L. Network structure, self-organization, and the growth of international collaboration in science.
Research Policy 2005; 34: 1608–1618.
[20] Lariviere V, Gingras Y, Archambault E. Canadian collaboration networks: a comparative analysis of the natural sciences, social
sciences and the humanities. Scientometrics 2006; 68; 519–533.
[21] Moody J. The structure of a social science collaboration network: disciplinary cohesion from 1963–1999. American Sociological
Review 2004; 69: 213–238.

Journal of Information Science, 2011, pp. 1–10 © The Author(s), DOI: 10.1177/0165551511407141

Downloaded from jis.sagepub.com at PENNSYLVANIA STATE UNIV on March 4, 2016

You might also like