Empirical research of advantage disciplines of C9 League*

Transcription

1

2

3 Empirical research of advantage disciplines of C9 League* Shu FANG 1, Haiyun XU 1, Zenghui YUE 1,2, Hongpei LI 1,2, Yan QI 1,2 & Bingxu LEI 1,2 1 Chengdu Branch,, Chengdu , China. 2 University of, Beijing , China Received: Oct. 10, 2013 Revised: Jan. 10, 2014 Accepted: Jan. 20, 2014 Abstract Purpose: Based on the case of the top nine elite universities in China (i.e. C9 League), we evaluate the university discipline advantages in terms of outputs and influence of scientific articles. Combining with the discipline collaboration analysis, we provide suggestions on discipline development and future collaboration for C9 League members. Design/methodology/approach: First, we carry out a correlation analysis. Then, we obtain the new combination evaluation indicators through factor analysis on seven primary indicators from InCite platform. Afterwards, we assign the weighted values to different indicators in order to get the evaluation scores and the ranking for each C9 League member. Then, we obtain the bibliographic data of the C9 League members for extracting the category classifications from the Web and transforming them into Degree-Granting and Personal Trainning Courses Catalogue (DPTCC). According to the original collaboration frequencies among C9 League members, we carry out the analysis on two-mode collaboration networks. Finally, we propose some suggestions on the future collaboration and development. Findings: There is correlation relationship between the outputs of scientific papers and influence indicators, so it is necessary to use the various indicators to acquire the more scientific evaluation. Meanwhile, league members are different in the discipline development level, and have inadequate collaboration with each other, which indicates that there exists the great potential for further collaborations. Research limitations: Despite the relative high quality SCI papers, preference for different disciplines of SCI and different discipline features for C9 members probably exert negative influence on evaluation results. Practical implications: We provide a research thought and an operational approach to the disciplines evaluation, which would promote more methods for scientific discipline evaluation and reasonable management for universities. * This work is supported by the West Light Foundation of Science (Grant No: Y4C ). Corresponding author: Haiyun Xu ( xuhy@clas.ac.cn). CJLIS Vol. 6 No. 4, 2013 pp

4 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 Originality/value: There are more and more articles on the universities evaluation indexes, but rarely on the relationship among the indicators. Therefore, our approach would be helpful for the evaluation of university disciplines. Keywords index Discipline evaluation; University evaluation; Scientific assessment; Evaluation 1 Introduction Discipline evaluation can support the universities to discover their own advantages and disadvantages, to improve the discipline management and make the strategies of development. It can also help to select the cooperative partners to strengthen the cooperation and learn from each other. On the other hand, students and their parents are chiefly concerned with the discipline evaluation which can help them choose the appropriate schools [1]. With Chinese national strength improving, Chinese higher education is also constantly flourishing. Constructing the world-class universities becomes Chinese people s great dream. C9 League is short for China 9 University League, including nine most prestigious universities, which have the greatest potential to be the world-class universities in China. On the way to become the world-class universities, discipline construction is undoubtedly the top priority task for any university. By possessing one or more world-class level disciplines, it could be qualified as the first-class one. While constructing the world-class disciplinary, the disciplinary evaluation is the principal way to grasp the level of academic development, to guide the discipline development direction and to promote the discipline development. Therefore, it is significant to carry out the discipline advantage evaluation on C9 League. Chen et al. [2] pointed out that the discipline evaluation is a comprehensive process of evaluating the scientific research and personnel training activities. Zhao [3] indicated that through discipline inspection and evaluation, university managers can gain wider experience, identify existing problems, reinforce the management, promote the communication and improve their work, so as to promote discipline development. Normally, discipline evaluation is conducted totally by the experts who give the evaluation scores in terms of their own personal experience and knowledge. Nowadays, with the rapid development of university disciplines, hardly any expert promises he can have a comprehensive understanding of each discipline and its branches. In this situation, quantitative analysis methods, such as statistical analysis and bibliometric methods, are more objective and more effective, which could be used as the useful supplement and reference to the expert evaluation. In fact, there are already various quantitative evaluation indicators and practical methods, several of which have even been used in the university evaluation practice. 80

5 Empirical research of advantage disciplines of C9 League Wainer et al. [4] correlated the peer evaluations performed in late 2009 by the disciplinary committees of CNPq (a Brazilian funding agency) with some standard bibliometric measures for 55 scientific areas. The highest correlations for each area ranged from 0.95 to 0.29, although there were areas with no significantly positive correlation with any of the metrics. QS World University Rankings [5], World University Rankings [6] and Academic Ranking of World Universities(ARWU) [7] are three most influential discipline rankings. Each of these three rankings established a comprehensive evaluation index and evaluation methods. Such comprehensive evaluation index provides a good reference for other relevant research and practice, and makes quantitative indicators more acceptable. However, it does not completely rely on the expert evaluation, so it has far-reaching implications for university evaluation. The Britain Guardian has eight indicators in its discipline ranking indicators [8], all these evaluation practice helps the university evaluation with more impersonality and maneuverability, and makes the quality evaluation become the real practice. With the increasing indicators and large index systems, how to manage the various indicators and how to combine a variety of indicators to get a better result are becoming more and more concerned nowadays. Dong et al. [9] made use of Analytic Hierarchy Process to evaluate the academic level of disciplines in universities. Jiang et al. [10] combined absolute indicators with relative indicators and compared with advantage indicators respectively in order to evaluate the outputs of papers on Humanities and Social in Science and Engineering universities in China. Li et al. [11] adopted Normalized Citation Impact (NCI) to evaluate the natural sciences, life and medical sciences, engineering and technological sciences of the colleges and universities of 985 Project Innovation Platform. Since the NCI has an normalized process, it can measure the relative gaps among the universities [12]. Zhang et al. [13] proposed a new method to calculate the weights of indexes by Analytic Hierarchy Process of systems and conduct the comprehensive evaluation. Wu [14] analyzed the influence of university disciplinary construction through the measurement of webometric indicators. The results revealed the difference in the influence of physics between Chinese and American universities and proposed some advice on the development of this discipline. Gao et al. [15] selected six indexes and established a comprehensive evaluation model for discipline based on the factor analysis. Wang et al. [16] started with the definition of metaevaluation and improved the oversea meta-evaluation checklist, and an empirical study of discipline evaluation was presented based on the example of key discipline evaluation of Hebei Province in China. All those studies above evaluate the influence of university disciplines from different perspectives but lack of the analysis on the correlations among the Project Innovation Platform Profile. Shu FANG et al. 81

6 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 indicators. Although each indicator has its unique characteristic, the indicators are not independent with each other but with variable correlations between them. Meanwhile, the quantity and influence of science outputs play an important role in the evaluation of academic influence. With the science publications constantly growing, a more precise relationship between the outputs and the influence is vital to improve the discipline evaluation. With the consideration above, we will discuss the advantages of university disciplines with their outputs and influence of SCI papers in order to explore the evaluation method for the output and impact of discipline. Then, we analyze and evaluate the discipline advantages of C9 League on the basis of outputs and influence of the SCI papers and their collaboration for further improvements in future. 2 Methodology 2.1 Research objects This paper selects C9 League [17] as the evaluation object, C9 League is the nine top universities in China, which are Peking University (PKU), Tsinghua University (THU), University of Science and Technology of China (USTC), Zhejiang University (ZJU), Nanjing University (NJU), Fudan University (FDU), Shanghai Jiaotong University (SJTU), Xi an Jiaotong University (XJTU) and Harbin Institute of Technology (HIT). The Russell Group and Australia Go8 elite universities establish an academic alliance, which aims to make a good use of the advantages of each university for further developments. The establishment of C9 League is to imitate the ivy league. As the leading universities in China, each member of the C9 League has its own advantages. 2.2 Data sources and indicators selected Discipline classification is prerequisite for discipline evaluation. The universities in China mainly adopted the discipline classification of the 2011 revised DPTCC, which is used for the enrollment, training, degree granting of masters and doctors, discipline construction and education statistical classification, etc. [18]. The 12 discipline categories are Philosophy, Economics, Law, Education, Literature, History, Natural Science, Engineering, Agriculture, Medicine, Art and Management. The evaluation data are collected from the institutional comparisons module of ISI InCites platform. In comparison with Web of Science, the evaluation function of InCites is more comprehensive and powerful. Since this dataset is in the worldwide domain, it is used to generate the metrics for comparing institution performance. InCites. 82

7 Empirical research of advantage disciplines of C9 League Based on address criteria, the subsets from this dataset can be created so that the academic influence of specific institution in its academic sector can be analyzed more deeply [19]. Furthermore, the report of InCites s institutional comparative analysis contains a variety of indicators for discipline evaluation of the C9 League universities. However, DPTCC division and ISI disciplines (Web of category) are not identical. Then, we make a conversion vocabulary from category of ISI to DPTCC, in order to achieve the evaluation on 12 disciplines of C9 League by ISI data. In this paper, we select 7 most commonly used indicators, which are presented and described in Table 1. Shu FANG et al. Table 1 InCites indexes of institutional comparison Indexes Description Function Document number Number of documents in Web of Science (WOS) database from an institution Times cited Total number of citations from WOS database % documents cited Number of cited documents divided by the total number of documents included in WOS database Cites per document (impact) % documents in discipline area % documents cited relative to area Impact relative to discipline area Average number of citations received per paper or the total number of citations divided by the total number of WOS database Number of articles in a discipline area produced by a country, territory, or institution divided by the total number of documents in the discipline area included in WOS database Percentage of cited documents for a country, territory or institution in one discipline area divided by percentage of cited papers of the world in the same discipline area. Percentage of cited documents for a country, territory or institution in one discipline area divided by percentage of cited papers of the world in the same discipline area To reflect the output of research articles, the greater the document number, the greater of the institutions research output ability. The higher of the times cited, the greater their academic influence The higher of the % documents cited, the greater influence of the proportion of papers. To reflect the influence of institutional research amount, the larger this indicator, the bigger influence of the institution scientific output. Same as the above To reflect the quality and influence of institutional research in a certain disciplinary relative to other institutions, the greater the value, indicating the greater influence. Same as the above 2.3 Evaluation methods and steps In Fig. 1, it shows the main contents, methods and procedures of this paper. The structure of this paper can be divided into four steps: 83

8 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 Disciplines evaluation Step 1 Acquisition of indicator data from InCite Acquisition of bibliographic data from WOS Acquisition of category from Web of Science Transform to DPTCC Collaboration analysis Step 3 Summary statistics Correlation analysis between indicators Correlative analysis Factor analysis Evaluation score Step 2 Results and suggestions Step 4 Fig. 1 Main content and technology roadmap. Step 1: To obtain the document number and times cited of 9 universities from 2001 to 2010, and obtain a descriptive statistics of the C9 League. Step 2: To get the data of 7 evaluation indicators in Table 1. Carry out the correlation analysis and the factor analysis on 7 primary indicators to generate new evaluation indicators. Assign the weighted values to the different indicators and obtain the evaluation score for each member of C9 League. Step 3: To obtain the bibliographic data of 9 universities from in the database of Web of Science, extract the web of category classifications, and make a conversion to DPTCC. Then we carry out discipline collaboration analysis on the C9 League members in 12 disciplines. Step 4: To according to the evaluation index scores and the collaboration network, we make recommendations for the academic development and cooperation to 9 universities at last. 3 Results and discussions 3.1 Descriptive statistics In the following presentation, we provide an overview to the C9 League members in the 12 different disciplines. We obtain the data of the C9 League members during the period from 2001 to 2010, which are from ISI InCite platform for comparison. In Fig. 2, it shows the characteristics of the document number for 9 universities in the twelve disciplines. The document number of 9 universities included in Web of Science shows a clear unbalanced distribution among different disciplines. 84

9 Empirical research of advantage disciplines of C9 League Shu FANG et al. Engineering % Natural science % Medicine % Others 3.002% Education 0.320% Agriculture 0.822% Economics 0.287% Law 0.193% Literature 0.148% Philosophy 0.063% History 0.023% Management science 1.140% Art 0.004% Fig. 2 Document number and ratio of C9 League disciplines. According to the document number, the disciplines of C9 League members can be divided into three classes (Table 2). The three classes show extreme imbalance, overwhelming superiority in Natural Science, Engineering and Medicine but few number in other disciplines. Table 2 Three classes of disciplines according to the document number Class Disciplines Percentage (%) 1 Natural Science, Engineering and Medicine 97 2 Management Science, Agriculture, Education, Economics and Law Literature, Philosophy, History and Art 0.24 In Fig. 3, it shows the differences of the document number and the times cited among 9 universities. Obviously, there is disparity in C9 League both for document number and the times cited. ZJU, THU and ZJU rank in the top three in both of Documents Times Cited PKU THU ZJU SJTU FDU USTC NJU HIT XJTU Fig. 3 Documents number and times cited of C9 League ( ). Note: PKU, Peking University; THU, Tsinghua University; ZJU, Zhejiang University; SJTU, Shanghai Jiao Tong University; FDU, Fudan University; USTC, University of Science and Technology in China; NJU, Nanjing University; HIT, Harbin Institute of Technology; XJTU, Xian Jiaotong University. 85

10 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 document number and times cited, these three universites have long enjoyed a good reputation in China. Although ZJU has the maximum document number, its times cited falls behind PKU and THU. PKU has the maximum times cited of papers and THU ranks the second, which means these two universities have better academic outputs influence. XJTU is in the last position both in document number and times cited. As is shown in Figs. 4 and Fig. 5, we construct two 2-mode networks by Netdraw tool: % documents in discipline area universities and % documents cited relative to area universities. The size of a node is proportional to the value of % documents in discipline area (Fig. 4) and % documents cited relative to area (Fig. 5). Obviously, it shows a roughly complementary network structure in Figs. 4 and 5. More specific, for a member of C9 League, when the % documents in discipline area is bigger, the % documents cited relative to area is smaller. Although the discipline with a large quantity of papers has more research outputs, it does not mean the maintenance and promotion of influence. In contrast, it perhaps leads to a relative decline in the influence. The disciplines of History and Art with relatively few papers have larger cites per document (Impact), and their scientific achievements have better influence per article. Due to the unbalance development of different disciplines, all C9 League members have outstanding performance in Natural Science, Engineering and Medicine. In fact, we can observe the dominating position of THU and ZJU in Natural Science and Engineering. It can also be observed that all the League members are relatively weak in Art (Fig. 4). Agriculture Economics Engineering ZHEJIANG UNIV FUDN UNIV PEKING UNIV History Law XIAN JIAOTONG UNIV Literature UNIV SCI & TECHNOL CHINA Philosophy HARBIN INST TECHNOL NANJING UNIV Management Science SHANGHAI JIAOTONG UNIV TSINGHUA UNIV Art Natural Science Medicine Education Fig. 4 Universities co-occurrence network according to % documents in discipline area. 86

11 Empirical research of advantage disciplines of C9 League Shu FANG et al. Medicine Literature History Art Agriculture UNIV SCI & TECHNOL CHINA SHANGHAI JIAO TONG UNIV Natural Science Management Science PEKING UNIV NANJING UNIV Law FUDN UNIV ZHEJIANG UNIV XIAN JIAOTONG UNIV Education Economics HARBIN INST TECHNOL Engineering TSINGHUA UNIV Philosophy Fig. 5 Universities co-occurrence network according to % documents cited relative to area. 3.2 Factor analysis of indicators Literally, except for document number and % documents in discipline area, the other five indicators in Table 1 are all related to citation. SPSS is used to carry out the correlative analysis of these indicators (Table 3). % documents Index cited relative to area % documents cited relative to area % documents in discipline area Impact relative to discipline area Documents number Table 3 Correlation between indicators (cosine similarity) % documents in discipline area Impact relative to discipline area Documents number Times cited % documents cited Cites per document Times cited % documents cited Cites per document Based on the results of the correlation analysis in Table 3, it can be found the correlations with each other indeed exist. For example, the cosine correlation coefficient between % documents cited relative to area and % documents cited is 87

12 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 up to 0.462, which belongs to the moderate correlation (0.3 correlation coefficient 0.5). % documents in discipline area shows the output force of an institution in certain disciplines, and % documents cited relative to area shows the influence of institutions in certain disciplines. The relevance between % documents in discipline area and % documents cited relative to area is (Table 3), which belongs to the weak correlation. SPSS software is used to carry out the factor analysis on seven indicators, in order to determine the suitability of the factor analysis with KMO value and Bartlett sphericity test. KMO test is proposed by Kaiser Meyer Olkin, the value of which is between 0 and 1. The higher value means more common factor between the variables, that is to say, it will be more suitable for the factor analysis. Bartlett sphericity test is used to test whether each variable is relevant. If the matrix is identical, it is unsuitable for the factor analysis. Generally, the significant level P < 0.05 indicates the presence of meaningful relationship between the original variables. In this analysis, the KMO value is 0.607, which indicates the factor analysis is better. Bartlett sphericity test value is , P = < 0.001, which proves that the factor analysis can be performed. Two principal components are obtained by the factor analysis, which shows 71% of the total variance can be explained by the two principal components (Table 4). Component Total Initial Eigenvalues % of variance Table 4 Total variance explained Cumulative (%) Extraction sums of squared loadings Total % of variance Cumulative (%) Extraction sums of squared loadings Total % of variance Cumulative (%) Then, we use the rotation factor to obtain the estimated load matrix (Table 5) and the component score coefficient matrix (Table 6) by the Varimax rotation method. In Table 5, it shows that seven indicators mentioned above could be extracted into two principal components, one of which is more involved in the output factors (including % documents in discipline area, document number, times cited and % Cohen, J. Statistical power analysis for the behavioral sciences [M].Routledge Academic,

13 Empirical research of advantage disciplines of C9 League documents cited), another one of which mainly reflects the impact/influential factor (including % documents cited relative to area, impact relative to discipline area and cites per document). Shu FANG et al. Table 5 Load factor of factor analysis Indicator Principal component 1 Principal component 2 % documents cited relative to area % documents in discipline area Impact relative to discipline area Document number Times cited % documents cited Cites per document Table 6 Component score coefficient matrix Indicator Principle component 1 Principle component 2 % documents cited relative to area % documents in discipline area Impact relative to discipline area Document number Times cited % documents cited Cites per document Evaluation results according to disciplines According to the coefficient matrix, the scores of the research output factor and the research impact factor of C9 League can be calculated respectively. Considering the high volume of scientific research output with relatively low influence in China, This status quo has complicated reasons, for example, China is a new member with fast growth in the scientific world. Actually, without the large amounts of outputs, we cannot obtain high influence in a very short time, neither can the C9 League. Therefore, we set the weighted coefficient values of the research output factor and the research impact factor as 0.7 to 0.3. In this way, the total scores of the discipline advantages of C9 League can be calculated. Then, the strengths and the weaknesses of disciplines of C9 League are divided into three levels: relative leading, medium, and relatively weak (including the case of lack of some disciplines). The process of classification calculation consists of two steps: Step 1: Calculate the average value of a discipline for all C9 League members, the universities above the average value are considered to be the relatively leading types. 89

14 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 Step 2: Calculate the average value of the universities below the average value in Step 1, the universities above the new average value are considered to be the medium types, the rest members are considered to be the relatively weak ones. This classification method can better illustrate the characteristics of the data distribution which evaluate the disciplines advantages of the universities, and can ensure that the universities having similar indicators are divided into one group which is more reasonable. The evaluation results are presented in Table 7. Table 7 Discipline Advantage Evaluation ranks for C9 League Disciplines Relative leading Medium Relatively weak Law ZJU, XJTU, PKU, USTC, FDU SJTU, HIT, THU NJU Engineering THU, ZJU, SJTU PKU, HIT, USTC NJU, FDU, XJTU Management Science PKU, THU, XJTU, SJTU USTC, HIT NJU, FDU, ZJU Education PKU, FDU, USTC, NJU, XJTU THU SJTU, ZJU HIT Economics PKU, USTC, THU, XJTU NJU, SJTU FDU, ZJU, HIT Natural Science PKU, THU, ZJU, USTC, NJU FDU, SJTU HIT, XJTU History USTC, ZJU, PKU, THU XJTU FDU, NJU, SJTU Agriculture ZJU, THU, USTC, HIT, PKU FDU, NJU SJTU, XJTU Literature XJTU, SJTU, USTC, PKU FDU, HIT, THU NJU, ZJU Medicine PKU, FDU, SJTU, ZJU NJU, USTC, THU XJTU, HIT Art FDU, NJU USTC, PKU THU Philosophy SJTU, XJTU PKU, HIT, THU, NJU FDU, ZJU, USTC PKU is the first well-known national university with a longstanding good reputation in China, which has 9 leading disciplines without any relatively weak discipline. Traditionally, THU had a dominating position in engineering in China, a remarkable position in Management, Economics, Science, Agriculture, etc., but a weak position in Art. After the department adjustments in 1952, THU had no liberal arts, neither was its science discipline. Since the late 1980s, THU has been devoted to developing a comprehensive university. Nowadays, in addition to maintaining the advantages of engineering, it has gained a leading position in Management Science, Economics, etc. In comparison with THU, there exists unbalanced devel opment in different disciplines in ZJU, which has a remarkable position in Law, Engineering and Agriculture, but relative weak in Management Science and Literature. In order to develop a world-class university, ZJU needs more strength to develop the weak subjects and makes sure to keep the balance development among various disciplines. Nearly all 12 disciplines in USTC get relative high score. Since disciplines are distributed in all three levels in Table 7, so it shows uneven disciplines development in SJTU and XJTU. Some universities have obvious advantages in certain disciplines, such as FDU and NJU in art field, and HIT in agriculture field. 90

15 Empirical research of advantage disciplines of C9 League Calling for special attention, our proposed two-step ranking method is not a relative one, neither are the ranking results. Especially in Literature, Philosophy, History and Art, the four disciplines take up 0.24% altogether. Then, there exists a big gap in Natural Science, Engineering and Medicine, which is up to 97% in total (Table 2). Shu FANG et al. 3.4 Discipline collaboration of C9 League Discipline collaboration network The ultimate goal of the discipline evaluation is to promote the development of university disciplines and realize the powerful combinations of the league members with the complementary advantages. The development of university disciplines relies on the interdisciplinary research projects to a large extent. In the great science era, the discipline branches are becoming increasingly complicated. Then, the research projects are more and more difficult to carry out by a single institution. Therefore, there is a sharp increase in cooperation among universities. Although the research cooperation among universities has a variety of ways, the co-authored articles are the most common and important way. Thus, the strength of interdisciplinary collaboration among universities is also available as an aspect of discipline advantage. We obtain the bibliographic data of 9 universities during from Web of Science, and extract classifications from web of category. Since DPTCC classification and ISI disciplines (web of category) are not identical, therefore, we make a conversion from category of ISI to DPTCC. According to the original collaboration frequencies between C9 League members respectively, we proceed discipline collaboration analysis between the C9 League members in 12 disciplines. The density of network reflects the overall collaboration level of the C9 League universities in various disciplines. As shown in Table 8, the density of C9 League research collaboration network is measured in different disciplines by the software Ucinet. Table 8 The density of C9 League research collaboration network Discipline Density Discipline Density Law 0.06 History 0 Engineering Agriculture 1.28 Management Science 1.25 Literature 0.17 Education 0.14 Medicine Economics 0.86 Art 0.44 Natural Science Philosophy

16 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 The highest network density for C9 League is in Natural Science, which indicates the highest level of scientific collaboration. Engineering and Medicine also have strong associate collaboration, there is also a certain degree of cooperation between Management Science and Agriculture. Meanwhile, there are scarce partnerships in other disciplines. Furthermore, there is no collaboration in History among C9 League members up to now Network visualization of discipline collaboration According to the original collaboration frequencies among C9 League members, we realize the visualization of the collaboration network by Netdraw tool [20] (Fig. 6). Figure 6 shows a two-mode map, one mode signifies the cooperation frequencies of C9 League members (illustrated in red square), another one is the 12-discipline (illustrated in blue circle). For both types of modes, the bigger size signifies more cooperation frequency. The lines connecting these modes signify that there exists the disciplinary collaboration between two universities, the thicker the lines, the more collaboration in the field between two universities, that is to say, they have the greater collaboration intensity. Literature XJTU&HIT SJTU&USTC ZJU&HIT ZJU&XJTU NJU&XJTU FDU&XJTU Art NJU&USTC PKU&HIT PKU&USTC Education HIT&USTC THU&USTC NJU&HIT THU&FDU FDU&HIT THU&XJTU XJTU&USTC FDU&USTC ZJU&USTC THU&NJU ZJU&SJTU THU&SJTU Natural Science Engineering Medicine THU&ZJU THU&HIT ZJU&NJU FDU&SJTU NJU&SJTU PKU&ZJU PKU&THU ZJU&FDU PKU&FDU SJTU&HIT SJTU&XJTU PKU&XJTU PKU&SJTU Economics FDU&NJU PKU&NJU Management Science Agriculture Law Philosophy Fig. 6 Disciplinary collaboration network of C9 League. Obviously, the most scientific collaboration of C9 League members can be found in Natural Science, Engineering and Medicine, which is the core of the whole collaboration network. However, the collaboration in other disciplines decreases to a large extent, such as in Management Science, Agriculture, Literature 92

17 Empirical research of advantage disciplines of C9 League and Economics, which are the edge nodes of the network. Furthermore, Philosophy, Education, Law, History and Art have the least collaboration in the network, which are illustrated in relative tiny nodes and far away from the core of the network. In addition, SJTU, FDU, USTC, PKU and THU collaborate actively. In contrast, the other C9 League members hardly collaborate. On the other hand, the collaboration behavior also demonstrates the strong geographical characteristics, such as PKU and THU, FDU and SJTU, which have the maximum times of the collaboration in Nature Science. This means, both of them have better development in Nature Science, especially for PKU & THU. In addition, PKU & THU is neighbored to each other in Beijing municipality and FDU & SJTU are both in Shanghai municipality. Meanwhile, PKU and THU also have the maximum times of the collaboration in both economics and engineering. In Agriculture and Management, SJTU and FDU are relatively weak according to the discipline advantage evaluation ranks in Table 7. But because of the positive cooperation activities of discipline, there exists relatively high frequency in scientific cooperation network. In Art and Education, it shows rare cooperation relations. But ZJU has more frequent scientific cooperation despite under-development in art field. The discipline advantages of C9 League suggest that the league members have different advantage disciplines and their scientific collaboration with each other in 12 disciplines is also different. Universities conducting more frequent scientific collaboration with other universities have more obvious discipline advantages and relatively balanced discipline advantages. Meanwhile, universities with relative bigger gap in development of certain disciplines rarely have any scientific cooperation with others. For example, USTC has an obvious advantage in the field of Nature Science among the C9 League members, while XJTU is the weakest one of C9 League members in Nature Science. However, the frequency of scientific collaboration between two universities is in low level (8 times). In Medicine field, FDU is ranked the 2 nd place while HIT is in the last among C9 League members. In fact, there is no scientific cooperation between these two universities until now. Therefore, the universities in weaker disciplines could proactively cooperate with the universities with advantages in corresponding disciplines, in order to improve the influence and the overall construction of their own disciplines. Shu FANG et al. 4 Conclusions This paper has explored the discipline advantage from the perspective of quantity and influence of the academic output included in SCI. Through a correlation analysis and factor analysis on seven primary indicators from InCite platform, we got the new combination evaluation indicators, assigned weights to different indicators, 93

18 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 got the evaluation score and rank classifications for each C9 League member. In addition, we carry out two mode collaboration networks analysis according to the original collaboration frequencies between C9 League members. We found various correlation relationships between scientific output and influence indicators, so it is very necessary to organize the various indicators to get a more scientific evaluation. Meanwhile, League members are difference in the discipline development level and inadequate collaboration to each other, which indicates great potential for further collaboration. This paper provided a research thought and operational approach to disciplines evaluation, which would substantially promote more scientific and reasonable discipline evaluation and management for universities. With relying much on the data from Web of Science, it is feasible to evaluate the discipline advantage by the output and influence of papers abroad, but the influence of domestic papers in Chinese is neglected. Meanwhile, the discipline classification criterion is made only according to the top level from hierarchal structure of discipline catalogue, which is relatively simple and rough. It will be more conducive to make a more detailed analysis on the secondary even the third level disciplines catalogue. In addition, it is not enough to evaluate the disciplines merely from the perspective of article output, it is also necessary to conduct related researches on faculties, such as experts, scholars and academic teams, and other scientific output such as patents, etc., so that discipline advantage can be measured more comprehensively and accurately. Author contributions S. Fang (fangsh@clas.ac.cn) designs the framework of the research and the outlines of the manuscript, and supervises the manuscript writing; H.Y. Xu (xuhy@mail.clas.ac.cn, corresponding author) and Z.H. Yue (yuezenghui@mail.las.ac.cn) collect and analyze the data, draw the figures, write and revise the manuscript; H.P. Li (lihongpei@mail.las.ac.cn) and Y. QI (qiyan@mail.las. ac.cn) conduct the literature survey. B.X.Lei (leibx@mail.las.ac.cn) downloads the data for scientific cooperation section. References 1 Qiu, J.P., Wen, F.F.,& Du, H. The improvement and enlightenment of Chinese university evaluation-analysis on the methods and results of Chinese universities and subjects evaluation. Evaluation & Management (in Chinese), 2011,(1): Retrieved on August 23, 2013, from 2 Chen, H.F, & Sun, L.J. A comparative study on discipline evaluation of Chinese and foreign universities (in Chinese). University (Academic), 2012, (8), Retrieved on August 23, 2013, from 94

19 Empirical research of advantage disciplines of C9 League 3 Zhao, S.Y. The establishing and the application of the universities discipline evaluation system. Northeastern University, Retrieved on August 23, 2013,from from cnki.net/kns/brief/default_result.aspx. 4 Wainer, J., & Vieira, P. Correlations between bibliometrics and peer evaluation for all disciplines: the evaluation of Brazilian scientists. Scientometrics, 2013, 96(2), Retrieved on August 23, 2013, from /fulltext.html. DOI: /s QS world university rankings: Methodology. Retrieved on August 27, 2013, from The essential elements in our world-leading formula. Retrieved on September 1, 2013, from methodology. 7 Academic ranking of world universities: Methodology. Retrieved on September 1, 2013, from 8 Comparison of the three UK university ranking. Retrieved on September 1,2013, from Dong, Y.L., & Ji, S.J. Study on method to evaluate the academic level of discipline in universities by AHP. Science and Technology Management Research (in Chinese), 2011, 31(5): Retrieved on August 23, 2013, from Html. 10 Jiang, C.L, & Liu, Z.Y. A quantitative analysis on the humanity & social science discipline paper structure of research oriented universities of science and technology. Science Research Management (in Chinese), 2007, 28(4): Retrieved on August 23, 2013, from Li, F., Miu, Y.J., & Qi, W. Academic influence evaluation of project 985 sponsored universities disciplines based on normalized citation impact (in Chinese). Science of Science and Management of S. & T., 2011, 32(10): Retrieved on August 23, 2013, from Li, F.P., & Wang, H.L. Efficiency evaluation in key disciplines scientific research based on DEA and SFA IEEE 18th International Conference on Industrial Engineering and Engineering Management (IE&EM 2011), 2011, Retrieved on August 23, 2013, from DOI: / ICIEEM Zhang, X.S, & Cheng, H.F. Fuzzy comprehensive evaluation of discipline competitiveness based on AHP. In Proceedings of Advancing Knowledge Discovery and Data Mining Technologies, 2009, Retrieved on August 23, 2013, from cai/article/viewarticle/ Wu, Y.Y. Quantitative evaluation of the discipline development in the university webometric analysis of physics. Information Science, 2010, 28(6), Retrieved on August 23, 2013, from 15 Gao, Y.R, Guo, J.X, & Pang, S.N. Integrated evaluation of key provincial disciplines in colleges and universities based on factor analysis method. Proceedings of the 2009 Second Shu FANG et al. 95

20 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 International Conference on Future Information Technology and Management Engineering (FITME 2009), 2009, Retrieved on August 23, 2013, from org/xpls/abs_all.jsp?arnumber= &tag=1. DOI: /FITME Wang, Y.F, & Qu, F. Discipline evaluation application study based on meta-evaluation theory. Science Research Management (in Chinese), 2009, 30(1), Retrieved on August 23, 2013, from 17 Universities of C9 League. Retrieved on August 23, 2013, from xwyyjsjyxx/xwsytjxx/yxmd/ shtml. 18 Degree-granting and personnel training courses catalogue (2011). Retrieved on August 23, 2013, from html. 19 Research evaluation and objective analysis of your people, programs and peers. Retrieved on August 20, 2013, from 20 Analytic Technologies. Retrieved on August 23, 2013, from Netdraw/netdraw.htm. 96

21 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 Brief News The 2014 annual meeting of Global Research Council will be held in Beijing The Annual Meeting of Global Research Council 2014 will take place in Beijing hosted by the (CAS) and the Natural and Engineering Research Council of Canada (NSERC). The 2 nd Annual Global Meeting of the Global Research Council (GRC) has ended in Berlin on 29 May Over three days, the heads of about 70 science and research councils around the world as well as high-ranking guests from science and research, science administration and research policy attended the meeting. The topics of discussion and endorsement were an Action Plan towards Open Access, a high-level Statement of Principles for Research Integrity as well as new statutes for the GRC. The outcomes of the meeting s discussions as well as the endorsed documents were presented at a press conference on 29 May 2013 in the Berlin-Brandenburgische Akademie der Wissenschaften in Berlin. DFG-President Professor Peter Strohschneider stressed the relevance of Open Access to publications as a main paradigm of scientific communication in the following years. The participants agreed that sharing research publications openly is a means to increase the quality of research communication and thus of research itself. The endorsed Action Plan includes activities by which GRC member organisations can foster the open exchange of research results with a high degree of flexibility. Therefore the plan specifies three basic principles: Encouragement, awareness rising, and support for researchers that wish to provide their results in Open Access. The implementation requires engaging a number of stakeholders, for instance, scientists and scholars themselves, universities, science organisations, libraries, and publishers. The endorsed Principles for Research Integrity point out that researchers and institutions themselves remain ultimately responsible for undertaking research with integrity on the one hand. On the other hand, research funding agencies have an obligation to ensure that supported research corresponds to the highest standards possible, as CNPq-President Professor Glaucius Oliva stated. Hosted jointly by the German Research Foundation (DFG) and the National Council of Technological and Scientific Development (CNPq) of Brazil, the 2 nd Annual Global Meeting was the largest of its kind so far and represented some 80 percent of the world s public, non-ministerial research potential. (Available at CJLIS Vol. 6 No. 4, 2013 pp

22 Chinese Journal of Library and Information Science Vol. 6 No. 4, 2013 Subscription Chinese Journal of Library and Information Science (CJLIS, Chinese Title: 中国文献情报 ) ISSN , CN 11570/G2; Price: $US 199/Volume Chinese Journal of Library and Information Science (CJLIS) is the first and only Englishlanguage academic journal in the field of Library and Information Science (LIS) that has been published in Mainland China so far. It aims to establish a platform for students, researchers and professionals in LIS and archive science to engage in intellectual dialogs and to share best professional practices of their experience. CJLIS adopts a double-blind peer-review process for each published article. The Editorial Board of CJLIS is composed of well-known scholars both from national and international LIS fields. The main aim of CJLIS is to serve librarians, information professionals, researchers and practitioners in the LIS field for their various purposes and keep them well informed of the latest development of LIS field both in China and the world. CJLIS mainly publishes original research articles, including research papers, library practice papers, progress reports, research reviews and perspectives. For subscription, you can use the following channels: 1) Via EBSCO International, Inc (USA) 2) To order at any post office inside China, with a distribution code ) To contact with the Editorial Office Address: No.33 Beisihuan Xilu, Zhongguancun, Haidian District, Beijing , P.R. China Tel: or ext Fax: chinalibraries@mail.las.ac.cn Website: CJLIS Vol. 6 No. 4, 2013 pp ) Via Beijing China Science Journal Publishing Co., Ltd. Address: Room 301, Building B, Desheng International Center, No. 83 Deshengmenwai Street, Beijing , P.R. China Tel: ext Fax: ext sunzhipeng@mail.sciencep.com

23

24