What are the publication patterns of the documents about “library science” indexed in WOS during the 2007 to 2016 decade?

Analysis of the 2007-2016 decade of research indexed as “library science” in the WOS database reveals divergent and expected trends in publications. In total, the 219 retrieved items classified by WOS as journal articles were included in the sample. 63 different journals were represented in this sample, including internationally-focused library science-oriented journals such as Information Research and Libri, but also including regionally focused journals such as African Journal of Library Archives and Information Science. Journals with five or more articles in our dataset are shown in Table 1.

We were interested to see that the two journals with the highest number of LIS articles indexed in WOS were neither Canadian nor United States-based, but rather journals published in Mexico (Investigación Bibliotecológica) and the Netherlands (Scientometrics), with Brazil (Perspectivas em Ciência da Informação, Informação & Sociedade) and Spain (Profesional de la Información) also having a strong showing in the dataset. Other countries of publication include the United States (Journal of the Association for Information Science and Technology, Library Trends, College & Research Libraries, Library Quarterly, Journal of the Medical Library Association), the United Kingdom (Journal of Academic Librarianship) and Malaysia (Malaysian Journal of Library and Information Science).

What is the range of languages of publication of the documents indexed in WOS between 2007 and 2016?

Of the 219 journal articles, 148 articles, just over two thirds, were published in English. Spanish was the second most published language with 46 publications (21 percent). The pie graph in Figure 2 shows this visually. Together, English and Spanish make up almost 90 percent of the content represented in the 219 journal articles, and there is representation of indexing articles in journals published in other languages including Portuguese and German.

Source: Own elaboration, 2019

FIGURE 2. WOS INDEXED ARTICLES, BY LANGUAGE 

That Spanish language publications make up over 20 percent of the indexed publications retrieved from WOS with the “library science” search may indicate a positive step in access to non-English publications in the field, although citation counts still skew heavily toward English publications. In 2017, ^{Olmeda-Gómez, Ovalle-Perandones and Perianes-Rodríguez} reported that Spanish language articles and authors affiliated with Spanish institutions indexed in WOS rose significantly from 1984 to 2014 in the “Information Science and Library Science” category as WOS commenced indexing more Spanish language professional and research journals. However, they did not further limit their findings to “library science” specifically so no direct comparisons can be made with their findings. When we looked specifically at the journals indexed, we saw that all the journals indexed publish at least some articles in English, although quite a few of the journals publish in multiple languages. This is the case with Profesional de la Información (Spanish and English), Perspectivas em Ciência da Informação (Spanish, English, Portuguese and French), Investigación Bibliotecológica (Spanish, English and Portuguese), Informação & Sociedade (Spanish, English, Portuguese and occasionally French).

Half of the journals indexed were published by university presses, including Investigación Bibliotecológica (National Autonomous University of Mexico), Library Trends (Johns Hopkins University), Perspectivas em Ciência da Informação (Federal University of Minas Gerais), Informação & Sociedade (Federal University of Paraiba), Library Quarterly (University of Chicago) and Malaysian Journal of Library & Information Science (University of Malaysia).

Which are the leading countries and institutions that produced the works indexed as “library science” between 2007 and 2016?

Perhaps the most striking feature of the data are the countries and institutions of the 408 authors of the 219 journal articles, as shown in Table 2. The authors affiliate with 45 countries around the world and 205 institutions in total. The United States dominates the list with authors from 79 institutions. The next closest country to that figure is Brazil with authors at 14 institutions and Germany and Taiwan coming in next with 10 institutions each.

United States-affiliated researchers authored more than one-third of these “library science” articles. One reason for this is most likely due to there being more LIS researchers publishing in the United States than any other country worldwide (^{Jabeen, Yun, & Rafiq et al., 2015}). Another reason is that WOS may have stronger alliances with American journal publishers since it is based in North America. However, LIS researchers from Australia do not typically limit their submissions to only Australian journals nor Swedish researchers only to European journals, etc., so even though WOS may possibly index more American-published journals, it does not follow that these journals necessarily contain more American content or articles written by Americans.

The map in Figure 3 presents the locations of authors, reflecting the dominance of northern hemisphere influence on these Library Science publications. This global north-south divide has been highlighted in other works overviewing global disparities in research and publishing (e.g., Jeffrey, 2014) so is not surprising. This map also shows how European countries, while perhaps not having more than three indexed “library science” publications per country in WOS, when mapped, show a reasonably even regional influence to that of the United States.

Source: Own elaboration, 2019

FIGURE 3.  GLOBAL VIEW OF AUTHORSHIP 

Which journals are indexed as representative of “library science” in WOS between 2007 and 2016?

The number of indexed articles per journal is related to the high periodicity of its journals. Most of the journals indexed are quarterly publications. Only Journal of the Association for Information Science and Technology (monthly) formerly known as the Journal of the American Society for Information Science and Technology, Journal of Academic Librarianship (bi-monthly) and College & Research Libraries (bi-monthly) are more frequent in publication schedules. The number of articles indexed in WOS per journal do not necessarily correlate with the SCImago Journal Rank (From now on and whenever mentioned, we shall refer to SJR as the abbreviated name of the indicator) or Quartile rankings. For example, the journal with the most indexed articles, Investigación Bibliotecológica, is ranked in the fourth quartile, with a 2016 SJR of 0.149, but has 31 articles in WOS, and an average of 0.68 citations, while the journal Scientometrics is in second listing, with 19 articles indexed, 7.26 average citations, and a 2016 SJR of 1.099, ranked in the first quartile. These figures are shown in Table 3.

As WOS does not disclose specific criteria for their indexing practices, we can only guess that indexing decisions are based on factors such as topics covered, methods employed and/or a mission to provide balance in provision of content from various countries, regions, institutions, and in a range of languages.

While it can take five to twenty-five years to truly be able to map and measure research impact (^{Ebrahim, 2013}; ^{Penfield, Baker, & Scoble et al., 2013}), in terms of the citations of the individual articles in our dataset, the numbers of citations per article indicated that 57 percent of the 219 journal articles had already been cited elsewhere within the ten-year period. These figures are shown in Figure 4. One 2011 article published in the Journal of the Association of Information Science and Technology was cited a total of 71 times while 42 percent of the sample had not yet been cited even once. In relation to this analysis, it is important to note that older articles have more time to be read and cited by the scientific community. For example, in 2016 47 articles were published, which have less dissemination time to achieve an increase in their citations.

Source: Own elaboration, 2019

FIGURE 4. NUMBER OF CITATIONS PER ARTICLE 

These journal-level metrics seem to be reflective of ^{Callaway’s (2016}) warning that journal impact factors are deceptive. Callaway writes that most of the articles published in a given journal have lower citation rates than the journal in which they are published, as the journal citation rank is distorted by very few highly-cited articles. To add to this argument, our data also seem to indicate that journal impact factor does not necessarily predict number of articles indexed from in an aggregate research publication database such as WOS from a given journal.

What are the authorship patterns for “library science” publications indexed in WOS between 2007 and 2016?

In our review of authorship patterns, we particularly wanted to focus on the trend of co-authorship and collaboration that has been reportedly growing in the social sciences (^{Endersby, 1996}). Research collaboration has also been linked to the maturity of a discipline or research system and the complexity of the scholarly networks reflects relationships between not only the authors but also their institutions and potentially their countries of affiliation (^{Ding, Rousseau, & Wolfram, 2014}; ^{Shin, Lee, & Kim, 2013}). The pie graph in Figure 5 shows that most of the publications in our sample (64.6 percent) came from single institutions -- that is, the publication had only one author or, if more than one author, all are listed as being at the same institution (i.e., internal collaboration). Domestic collaborations of authors working with individuals at different institutions within the same country were the second highest count of collaborations on these publications, at 28.3 percent. International collaborations were a very small piece of this sample of publications with only 7 percent of the publications including authors collaborating from different countries in the world.

Source: Own elaboration, 2019

FIGURE 5. INTERNAL AND EXTERNAL AUTHORSHIP COLLABORATIONS 

The small quantity of international collaborations is interesting in light of ^{Han et al. (2014}) review of international collaboration in LIS publications wherein they found international collaboration to have increased significantly from 2000 to 2011 and inferred that the trend would continue rising in following years. Factors such as geography, politics and language can affect international collaboration decisions and evidence of this seems to be seen in the choice of collaborators (^{Subramanyam, 1983}).

That said, recent studies have shown that the number of papers written by two or more authors are increasingly outnumbering the number of articles written individually (^Jeffery ,2014; ^{Todeschini, & Baccini, 2016}). In our dataset, we see that there is indeed a strong showing of collaborative publishing in the WOS library science articles. As Figure 6 shows, the authorship patterns of publications (including number of authors per article) in library science shows that single-authored papers still rank proportionately first: of the 219 journal articles in our study, 90 articles (41 percent) were written by a single author.

Source: Own elaboration, 2019

FIGURE 6 AUTHORSHIP PATTERNS 

Additionally, it is important to note that the authorship trend throughout the decade indicated (Figure 7) shows a predominance of publications with a single author and two authors. The temporal evolution is relevant to understand how the predominance of production of articles with a single author is distributed over the years. Under this perspective, it is possible to affirm that the first five years have high predominance of individual research, while in the following years research in co-authorship becomes relevant.

Source: Own elaboration, 2019

FIGURE 7 AUTHORSHIP TREND 

Which authors are most indexed and cited and what are their most important contributions?

The ten authors with the highest number of publications in our dataset primarily affiliate with institutions in Europe and the United States, with two authors affiliating from institutions in Asia (China) or Latin America (Mexico). These authors are shown in Table 4 with their affiliations and rankings in our dataset.

Of the twenty-six total publications written by these ten most-represented authors, nine of the papers were single-authored. All others were multiple-authored, with eight of these top ten authors collaborating in pairs with each other, including Thelwall and Levitt, Bornmann and Leydesdorff, Sugimoto and Ni, and Garoufallou and Vassilakaki. Three of these four collaborations are domestic in affiliation (Thelwall and Leavitt: England, Sugimoto and Ni: USA, and Garoufallou and Vassilakaki: Greece) and the fourth (Bornmann and Leydesdorff) is regional (Germany and Netherlands). Voutssas (Mexico) was the only scholar with all single-author publications; other authors collaborated with one or more author in their publications included in the dataset.

In terms of citation counts, the ten most-cited articles in our dataset are shown in Table 5. Nine were multi-authored articles and one was single-authored. Of the nine multi-authored publications, five were intra-institutional collaborations and four were international collaborations. This finding resonates with previous bibliometric studies such as ^{Beaver (2004}), ^{Godin & Gingras (2000}), and ^{Hunter & Leahey (2008}) which emphasize the increased epistemic authority, impact value, and institutional prestige associated with collaborative research.(See Appendix I)

[InsertarTable 5. Most Cited Articles and Type of Collaboration aqui]

While interdisciplinary collaborations hold unique challenges (e.g., different methodological approaches, communication, literature bases, foundational theoretical paradigms, value systems, language and terminology), the literature indicates that interdisciplinary collaboration is becoming increasingly important for success in academia, particularly for attaining grant funding (^{Gooch, 2005}). This is an area that needs more study for the LIS field as the nature of the LIS discipline is ripe with opportunities for collaborations with other disciplines.