1 Introduction

Recently, the push for academic publications is seen at its peak. Scholars have noted that university rankings, funding, and even the prestige of a certain field of study are all interconnected with the number of academic publications and their subsequent citations by later articles [1]. Furthermore, the ubiquitous nature of the internet has also inspired research related tools such as the Thomson Reuters’ ISI Web of Science (WOS) website, which provides a window of access to published articles of interest [2, 3].

Tandem with such prevalent usage of the WOS, many studies have emerged discussing the actual contributions of such tools [4]. More specifically, the practical implications of bibliometric researches and studies that are able to provide various insights to both novice and experts [5]. Two practical application of bibliometric studies are the evaluation and comparison of research performance either across universities or countries [6, 7] and research trend analysis of a specific field of study [8–10].

The concept of bibliometric originated from the Greek word for book biblios and measure metron [11]. Broadus [12] noted that the word bibliometric first appeared in a study by Fairthorne [13]; wherein the words bibliometric description was used. With the increased in bibliometric researches, the question regarding how data is dispersed or distributed within the citation databases has become of great interest [14]. Various distribution laws have been incorporated to such studies, such as the Pareto distribution [15], Bradford distribution [16], and many other variations and combinations. Many bibliometric theorists have proposed that within a specific field of study, 80 % of the articles are contributed by 20 % of the field’s source journals [17]. Hence, the objective of this study is to determine if such phenomenon also exists in the pool of mechanical science researches in Taiwan.

Since 1993, contributions to the WOS database have been gradually changing with strong emphasis coming from Asian countries; Taiwan included [18]. In Taiwan, engineering researches are among the highly cited papers within WOS database [19]. However, majority of the contributions are papers coming from twelve key universities in Taiwan. This also holds true in the areas of mechanical science and engineering [20]. Furthermore, graduate students also played an important role in increasing the number of contributions [21]. With regards to the research trends in mechanical science, one seminal study on mechanical engineering papers’ bibliometric information reveals that technological advancement played an important role in shifting the entire research focus of the field [22]. Currently, there seems to be a gap in the literature regarding the overall implications of the mechanical science and engineering researches in the WOS database. Hence, a study focusing on the distribution characteristics of the mechanical science and engineering researches in Taiwan is provided.

2 Research Method

The current study is designed as a descriptive research; wherein the focus of the study is placed on the conditions or relationships that exist, practices that prevail, beliefs, points of views, or attributes that are held, processes that are ongoing, effects that are currently being felt [23]. To summarized, the bibliometric database was accessed through the use of the WOS website. The database included the various citation indexes or bibliographic indexes which allow users to search for articles included in certain categories. The database consisted of the Science Citation Index (SCI), the Social Science Citation Index (SSCI), and the Arts and Humanities Citation Index (A&HCI); which are some of the most commonly used tools for bibliometric researches [24]. For the current study, a total of 9,576 bibliometric entries from articles in the field of Mechanical Sciences are downloaded. Besides the normal bibliometric information (author/s, year, source, title, abstract, and keywords), other facts including the total number of words and pages, author’s affiliations, funding information, geographical location, number of citations are also downloaded and analyzed.

3 Results and Discussions

After the data is downloaded, analysis is then undertaken with regards to their publication year, while correlation analysis of the articles’ age with author counts, number of references, number of citations, authors’ supplied keywords, and publisher supplied keywords. Such analysis is undertaken with the intention of better understanding how bibliographic information is affected by the passing of time.

Table 13.1 shows the yearly frequency of publications with their corresponding article’s field of studies. Data shows that majority of the articles (around 85 %) are related to three major fields such as engineering with 47.32 %, thermodynamics with 24.63 %, and mechanics with 13.85 %. More importantly, Table 13.1 also shows the rate of increase/decrease in yearly publication frequency. Here data points out that there is an overall average increase of 0.045 % starting from 1996 till 2012.

Table 13.1 Publications by year and field of studies
Full size table

To further understand the nature of the articles, Table 13.2 shows the yearly article distribution with their corresponding subject areas. Data shows similar distributions tendencies as of Table 13.1 with the three major subject areas as mechanical engineering with 60.24 %, mechanics with 59.98 %, and thermodynamics with 24.61 %. Although both tables shows similar trends, Table 13.2 provides a more disperse but clearer bibliometric information.

Table 13.2 Publications by year and subject areas
Full size table

With regards to the bibliometric information, Tables 13.3 and 13.4 shows the various summary. Table 13.3 shows that 19.70 % of the articles are accomplished through projects funded majority by the Taiwan National Science Council; a government research agency with 16.60 %. In addition, data suggests that majority of the studies; around 4 out of 5 are not funded. At first, this seems to indicate an unhealthy situation. Further analysis shows that project funding starting from 2008 have been steadily increasing. Table 13.4 shows the summary of the bibliometric information. Data shows that the maximum number of authors is 12, which when compared to other major fields seem to be low. Furthermore, correlation analysis shows that the title and abstract word counts significantly increases in newer articles. In addition, the number of references also increases in younger articles. In sum, result also shows that mechanical science researches follow the majority trend of older articles having the most number of citations.

Table 13.3 Article funding information
Full size table
Table 13.4 Bibliometric summary
Full size table

As for the article distribution analysis, Fig. 13.1 shows the Pareto distribution graph of the articles. In terms of publication analysis, when the publications are arranged by means of source journal contributions from highest to lowest; Pareto distribution denotes that 20 % (43) of the journals should be responsible for 80 % of the articles and the remaining 80 % (173) of the journals should be responsible for 20 % of the articles [25]. However, in the current study; 20 % of the journals represent around 72.03 % (6898) articles, while the remaining 80 % of the journals represent around 27.97 % (2678) of the articles.

Fig. 13.1
figure 1

Pareto distribution graph

Full size image

Figure 13.2 shows the Bradford-Zipf plot; denoting the exponential diminishing returns of article contributions [26]. This is computed by dividing the contributing journals into three equal groups. Figure 13.2 show that the first 33 % (72) of the journals represents 85.70 % (8,207) articles. The next 34 % (72) of the journals represents 11.92 % (1141) articles and the remaining 33 % (72) of the journals represents 2.28 % (228) of the articles (Fig. 13.3, Table 13.5).

Fig. 13.2
figure 2

Bradford-Zipf’s plot

Full size image
Fig. 13.3
figure 3

Revised Welch vitality curve

Full size image
Table 13.5 Top 12 keywords
Full size table

With regards to the Welch’s Vitality Curve; this is demonstrated by having the 20-70-10 concept. Wherein, the top 20 % of the journal is the leading (representing) journals of the field, while the following 70 % is the adequate majority with the remaining 10 % of the journals considered as negligible [27]. For the current study, a revised Welch Vitality Curve is proposed, wherein the distribution is 10-70-20 with the top 10 % being the leading journals of the field. An important factor of distribution analysis is determining the least number of journals with the most prominent impact in the field; hence, Table 13.6 shows the citations per article breakdown across the different distribution schemes. Result shows that the revised Welch Vitality Curve has provided the most likely suitable condition. Wherein, a researcher will need only to looked into 22 major journals and be able to gather the article with the highest impact in the field. In essence, saves time and effort, while not sacrificing the quality of the work.

Table 13.6 Comparison of article and journal distributions
Full size table

Additional information is provided in Table 13.5, wherein the all-time top 12 keywords used by authors are listed. While the top 5 contributing journals are the following: International Journal of Heat and Mass Transfer, Journal of Sound and Vibration, Journal of Micromechanics and Micro-engineering, International Communications in Heat and Mass Transfer, and Journal of Mechanics. These 5 journals represent almost 25 % of the article pool in the entire Mechanical Science researches in Taiwan.

4 Conclusion

The objective of this study is to determine how knowledge in terms of journal publications is distributed within the field of mechanical science and engineering in Taiwan. Many bibliometric theorists have proposed the use of Pareto distribution; wherein 80 % of the articles are contributed by 20 % of the field’s source journals. Result of the bibliometric analysis shows that articles are somewhat distributed in a Pareto like form. However, results can be better explained through the use of the Revised Welch Vitality Curve. Wherein, the distribution is 10-70-20 with the top 10 % being the leading journals of the field, while the following 70 % is the adequate majority with the remaining 20 % of the journals considered as negligible. Furthermore, bibliometric result shows that mechanical science researches follow the majority trend of older articles having most number of the citations with newer articles having a significant higher number of authors and references. It is urged that future researchers test the distribution model, to confirm whether the Revised Welch Vitality Curve also exists in other field of studies.