Houqiang Yu

Research on the cross-citation relationship of core authors in scientometrics

Generally speaking, citation relationship among authors can be divided into 3 types: co-citation, coupling and cross-citation. Since author co-citation analysis was first introduced in 1982, it has been widely applied to study discipline structure, research state and research trends. Afterwards, conception of author bibliographic-coupling analysis was put forward and related empirical studies provided a method for mapping active authors in a research field for a more realistic picture of the current state of its research activities. Additionally, if one of author A’s papers has a citation from one of author B’s, there is cross-citation relationship between A and B. However, studies based on author cross-citation relationship mainly describe citation behaviors themselves using citation identity and citation image; they rarely involve any implicit knowledge communication, author research correlation or discovering academic communities. Author cross-citation analysis infers to both citing and cited phenomenon, which can be roughly correspond to citation identity and citation image. The study will further explore the author cross-citation relationship with core authors in scientometrics field as study object in order to provide reference for development of scientometrics field and in-depth application of citation analysis.

Comparative study on structure and correlation among author co-occurrence networks in bibliometrics

This paper introduces author-level bibliometric co-occurrence network by discussing its history and contribution to the analysis of scholarly communication and intellectual structure. The difference among various author co-occurrence networks, which type of network shall be adapted in different situations, as well as the relationship among these networks, however, remain not explored. Five types of author co-occurrence networks were proposed: (1) co-authorship (CA); (2) author co-citation (ACC); (3) author bibliographic coupling (ABC); (4) words-based author coupling (WAC); (5) journals-based author coupling (JAC). Networks of 98 high impact authors from 30 journals indexed by 2011 version of Journal Citation Report-SSCI under the Information Science & Library Science category are constructed for study. Social network analysis and hierarchical cluster analysis are applied to identify sub-networks with results visualized by VOSviewer software. QAP test is used to find potential correlation among networks. Cluster analysis results show that all the five types of networks have the power for revealing intellectual structure of sciences but the revealed structures are different from each other. ABC identified more sub-structures than other types of network, followed by CA and ACC. WAC result is easily affected and JAC result is ambiguous. QAP test result shows that ABC network has the highest proximity with other types of networks while CA network has relatively lower proximity with other networks. This paper will provide a better comprehension of author interaction and contribute to cognitive application of author co-occurrence network analysis.

Context of altmetrics data matters: an investigation of count type and user category

Context of altmetrics data is essential for further understanding value of altmetrics beyond raw counts. Mainly two facets of context are explored, the count type which reflects user’s multiple altmetrics behaviors and user category which reflects part of user’s background. Based on 5.18 records provided by Altmetric.com, both descriptive statistics and t test result show significant difference between number of posts (NP) and number of unique users (NUU). For several altmetrics indicators, NP has moderate to low correlation with NUU. User category is found to have huge impact on altmetrics count. Analysis of twitter user category shows the general tweet distribution is strongly influenced by the public user. Tweets from research user are more correlated with citations than any other user categories. Moreover, disciplinary difference exists for different user categories.

Global science discussed in local altmetrics: Weibo and its comparison with Twitter

Local altmetrics is currently an integral part of the altmetrics landscape. This paper aims to investigate the characteristics of microblog altmetrics of the Chinese microblog platform, Weibo, to shed light on cultural differences and draw attention to local altmetrics in developing countries. The analysis is based on 4.4 million records provided by Altmetric.com. Data collected are from March 2014 to July 2015. It is found that Weibo users discuss global science, more actively compared with several international altmetrics sources. Statistical results show strong evidence of the immediacy advantage of metrics based on Weibo as well as Twitter and the general altmetrics over citations. Distribution of Weibo altmetrics on the article level, source level and discipline level are highly skewed. Overall, compared with Twitter, Weibo altmetrics present similar distributions, with some minor variations. To better understand how and why Weibo users discuss global scientific articles, the top weiboed articles, sources and disciplines are identified and further explored. Our content analysis shows that the common motivation of scientific weibos is to disseminate or discuss the articles because they are interesting, surprising, academically useful or practically useful. Conclusion of articles is the most frequently mentioned element in scientific weibos. In addition, different from Twitter, Weibo users have a preference for traditional prestigious journals.

Does the average JIF percentile make a difference

Average journal impact factor (JIF) percentile is a novel bibliometric indicator introduced by Thomson Reuters. It’s of great significance to study the characteristics of its data distribution and relationship with other bibliometric indicators, in order to assess its usefulness as a new bibliometric indicator. The research began by analyzing the meaning of average JIF percentile, and compared its statistical difference with impact factor. Based upon factor analysis, the paper used multivariate regression and quantile regression to study the relationship between average JIF percentile and other bibliometric indicators. Results showed that average JIF percentile had changed the statistical characteristic of impact factor, e.g. improved the relative value of impact factor, having smaller variation coefficient and distribution closer to normal distribution. Because it’s non-parametric transformation, it cannot be used to measure the relative gap between journals; Average JIF percentile had the highest regression coefficient with journal impact, followed by timeliness and lastly the citable items; The lower the average JIF percentile, the higher the elastic coefficient of journal impact; When average JIF percentile was extremely high or extremely low, citable items were not correlated with the average JIF percentile at all; When average JIF percentile was low, elastic coefficient of timeliness was even higher; Average JIF percentile was not a proper indicator for multivariate journal evaluation; Average JIF percentile had both the advantages and disadvantages of impact factor, and thus had the same limitation in applying as the impact factor.

Who, what, why? An exploration of JoVE scientific video publications in tweets

This paper investigates how and why scientific video articles are communicated on Twitter. We use video articles published in the Journal of Visualized Experiments (JoVE) as our objects of study. We harvested tweets from October 2011 to November 2015 that contained one or more JoVE links. These tweets “citing” JoVE articles were analyzed both statistically and qualitatively. In this paper, we present the distribution of these tweets, with a closer look at the affordance use of Twitter including hashtags and mentions. In addition, we conducted a content analysis of the sampled Twitter accounts and tweets. We present the coding schemes and results of both Twitter user accounts and tweets text. In addition to the analysis of the coding results, we discuss the content of the tweets with particular attention to issues including the video/visual feature mentioned, the role of Twitter bots, and self-promotion of different stakeholders in the Twitter communication of JoVE video publications.

Analysis on research activity and impact of authors in Chinese information science based on citation relationship

– The purpose of this paper is to further explore the co-citation and bibliographic-coupling relationship among the core authors in the field of Chinese information science (IS), to expose research activity and author impact, and to make induction analyses about Chinese IS research patterns and theme evolution. , – The research data include 8,567 papers and 70,947 cited articles in the IS field indexed by Chinese Social Sciences Citation Index from 2000 to 2009. Author co-citation analysis, author bibliographic-coupling analysis, social network analysis, and factor analysis were combined to explore co-citation and bibliographic-coupling relationships and to identify research groups and subjects. , – Scholars with greatest impact are different from the most active scholars of Chinese IS; there is no uniform impact pattern forming since authors’ impact subjects are scattered and not steady; while authors’ research activities present higher independence and concentration, there is still no steady research pattern due to no deep research existing. Furthermore, Chinese IS studies can be delineated by: foundation or extension. The research subjects of these two parts, as well as their corresponding/contributing authors, are different under different views. The general research status of core authors is concentrated, while their impact is broad. , – The combined use of some related methods could enrich the development and methodology research of the discipline, and the results establish a reference point on the development of IS research.

Is there Lingua Franca in informal scientific communication? Evidence from language distribution of scientific tweets

Language distribution in scientific communication reflects the influence of different languages on science in global perspective. The study, based on over 450 thousand scientific tweets of all publications indexed by Scopus in June 2015, reveals the language distribution in informal scientific communication. Moreover, this result is compared with the language distribution in formal scientific communication reflected in scientific publications. Results show: (1) The language of scientific tweets is concentrated in English (91%), Japanese (2.4%) and Spanish (1.7%), while the language of scientific publications is concentrated in English (90.6%), Chinese (5%) and German (1.1%). (2) Both scientific tweets and scientific publications present disciplinary differences in language distribution, reflecting the different amount of attention that authors of different languages have on certain disciplines. (3) Except Saudi Arabia, investigated countries all over the world, regardless of whether their native language is English or not, all have English scientific tweets in the dominant position. For the vast majority of these countries, the native language scientific tweets only rank the second position. (4) Overall, 26% of tweeters use more than one language to tweet scientific products, while 49% of scientific tweeters tweet everything in English only. The results indicate that English has undoubtedly become the lingua franca in informal scientific communication.