Borgman and Furner (2002) highlighted the importance of choosing the publication venue for scholars, “matching the topic of the paper with the scope statement of the journal and the topicality of previous articles published.” However, authors and their publications determine the creation of intellectual knowledge and social identity of those same journals where they publish their work. This is because by publishing in a certain journal, scholars become active participants in the group of authors associated with the journal. It is then that their academic identity is added to that of the other authors who have also published there, contributing to the thematic configuration of that publication venue. In the literature there are several definitions of the academic identity of an author. For example, White (2001) defined an author’s citation identity as the set of authors that an author cites. Assuming an author has been cited at all, White and McCain (1998) also defined an author’s citation image as the set of all authors with whom an author has been co-cited.
In this scenario, the grouping of authors according to the shared academic journals can provide evidence for the existence of social and intellectual communities of scholars. Price and Beaver (1966) introduced the concept of “invisible college”, highlighting the importance of the social grouping of scholars. The basic phenomenon behind invisible colleges is that in the most active and competitive research fields there seems to exist an “in-group” of academics. Authors in this subgroup contribute materially to research on the same topic and often publish their work in similar publication venues. As suggested in (Price & Beaver, 1966), the problem is that although it is relatively easy to find a well-known scholar in a chosen field of research, it is considerably more difficult to select a group of authors who make up the majority of a single invisible college. The basic difficulty is capturing and dissecting that group of scholars within a specific field of research or discipline (Price & Beaver, 1966).
In this paper, following these ideas, we propose a clustering of authors according to their journal publication profile. Using this approach, the similarity between authors is computed based on the shared academic journals. Therefore, the basic assumption will be that authors who published in the same journals are more likely to share similar conceptual frameworks and interests (and therefore, they are more likely to be part of the same invisible college) than those who never publish in the same journals (Garcia et al., 2012; Ni et al., 2013; Price & Beaver, 1966). Thus, in our study, the mention of the name of a journal evokes all the articles that have been published by that journal and consequently represents a “conceptual marker,” (Ni et al., 2013). Our model is based on this idea: if the journal in which scholars published their work represents a conceptual marker, it follows that authors can be identified by the journals in which they publish their research (Garcia et al., 2012; Ni et al., 2013). In this way authors belonging to a research area can differentiate themselves by identifying an underlying group of journals. The authors will thus be grouped both socially and intellectually, the latter being the norm in citation studies.
In disciplines with high degrees of single authorship, hyper-authorship behavior does not determine a good descriptor of the interaction structure between the authors of the discipline. Therefore, in this scenario, collaborative networks based on hyper-authorship behavior are not appropriate. Alternatively, our approach identifies similarities between authors based on similarities in their journal publication profiles, and consequently, it does not matter whether or not authors interact with their community through co-authorship. Therefore, we follow (Minguillo, 2010)’s fundamental premises that “journals act as platforms of interaction and membership for scientific fields.” He stated that “analysis of the relation between authors and journals makes it possible to see how communication among scientists overall forms the structure of highly specialized and well-controlled scientific (sub)fields influenced by the reputational system and cognitive limitations.”
Using this premise, our research will investigate the degree to which authors (who published in information science and library science journals) are related to each other, based on their journal publication profile (i.e., their research output). Therefore, the grouping of authors using journal coupling (i.e., the shared academic journals) provides a new approach to examine the social and intellectual coherence of the considered set of authors.
In this study, after presenting our approach, we find the clusters of authors (or invisible colleges) in the information science and library science (IS&LS) category of the Web of Science Core Collection. Therefore, our results may be of interest to IS&LS scholars. This is because these results provide a new lens for grouping authors, making use of the authors’ journal publication profile and journal coupling. Furthermore, extending our approach to the study of the structure of other disciplines would possibly be of interest to historians of science as well as scientometricians.
To demonstrate the structure of scientific domains, several methods based on co-occurrence and coupling have been proposed. On the one hand, co-citation, co-wording, and co-authorship methods, are based on the co-occurrence of elements (Qiu et at., 2014; Small, 1973; Tijssen & Van Raan, 1994; White & McCain, 1998). For example, co-citation establishes a co-occurrence relationship between cited references. On the other hand, Kessler (1963) proposed the concept that “two documents are related if they share the same set of citations.” Bibliographic coupling is based precisely on this concept and, thus, researchers want to observe how many common references two articles share. Later Small and Koenig (1977) also proposed using bibliographic coupling of journals. In our study, the similarity of authors is based on this type of coupling assumption. We then claim that two authors are related if they share similar publication venues. This is journal coupling of authors. We used journals to discover relationships among authors. In journal coupling of authors, the number of shared publication venues among authors can be used to measure their relationships.
In relation to research works that connect authors through journals, Garcia et al. (2012) presented a novel methodology to map academic institutions based on their journal publication profiles. Using a sample of Spanish universities as a case study, Garcia et al., (2012) mapped the study sample according to the overall research output of each institution and in different disciplinary contexts. Furthermore, Robinson-Garcia et al., (2013) presented a descriptive analysis of Spanish universities according to their journal publication profile in five scientific domains during the period 2007-2011. In their study, two universities had a similar journal publication profile if they published in a high number of common journals. This idea led them to the possibility of mapping universities and thus offering an enriched view of the Spanish higher education system. Extending the idea of bibliographic coupling, Ni et al., (2013) proposed a metric called author-publication-place coupling that identifies similarities between journals based on the similarities of their author profiles. The use of this method in information science and library science journals provided evidence of four distinct subfields, namely, management information systems, specialized information and library science, library science-focused, and information science-focused research.
Therefore, our study introduces a novel approach for clustering authors on the basis of the common publication in academic journals. A map of authors can then be used to visualize the relationships between these authors using journal coupling. This tool will make it possible for individual scholars to see how they relate to other authors. Also, it may be relevant for studies exploring the scholars in a discipline or area, or, more in general, within the research management context (Noyons, 2004).
When visualizing the structure of science, the authors’ maps are commonly visualized as node-edge diagrams, similar to those used in network science. On one hand, the representation of the authors is carried out by placing each of them in a two-dimensional space (Klavans & Boyack, 2009). On the other hand, relationships between authors are represented by the explicit linking of pairs of authors (Klavans & Boyack, 2009). Henry Small and colleagues (Griffith et al., 1974; Small & Koenig, 1977; Small & Garfield, 1985) were the first to use bibliometric methodologies to map all of science. Their approach was simple: focus their attention on highly co-cited articles. In their developments, they defined these highly co-cited articles as pairs of references that coexist in bibliographies at least five times in a year. Using a different level of granularity than articles when visualizing the structure of science, specifically publication venues, Narin et al., (1972) chose academic journals as the basic unit for mapping science. To do this, Narin et al. (1972) first divided the journals into a certain number of groups and then used a visualization algorithm to generate a layout of the previously obtained groups. More recently, different techniques have been used to create discipline-based maps. For this, the large amounts of data that are currently available on scientific publications have been used. In addition to new visualization tools. Using this approach, Moya-Anegón et al., (2004, 2007) created several discipline-based maps. They used the Thomson Scientific disciplinary classification system. Leydesdorff and Rafols (2009) represents another relevant example of discipline-based maps.
In information science (IS) and library science (LS), different bibliometric analyzes have been carried out to study the structure of the research area. For example, Saracevic, (1999) seeing “the two as separate but closely related fields.” Analyzing the structure of library and information science (LIS), Moya-Anegón et al., (2006) found LS, together with information management, “being included or excluded depending on what level the co-citation analyzes” were done. However, according to a set of citation analyzes performed in (Astrom, 2010), “the patterns in the citation data support the concept of a joint LIS field with information science and library science being the two main subfields.” In our study, we determine the structure of information science and library science using journal coupling.
Several previous studies identified the most influential journals in the LIS field. For example, Vinkler, (2019) identified the core journals in scientometrics using the frequency of articles in academic journals in the elite publication subgroups of Price medalists. Nixon (2014), Walters and Wilder (2015), and Weerasinghe (2017) also found the most important journals for determining the structure of the research area using different approaches, mainly reflecting reputational factors based on bibliometric indicators or expert surveys. In a more recent paper, Safón and Docampo (2023) found trendy journals which are the most read by scholars who are publishing within the scope of a consolidated discipline. In the following sections, we also analyze the most relevant journals to determine the structure of information science and library science when using journal coupling. Our results are consistent with those of previous studies that identify which journals are truly influential within the discipline.
In this article we present an automatic system for grouping academics from a research area. The first choice in our model is related to the research production units that we will use as author descriptors. Here, given the availability of large amounts of data on scientific publications, we have used academic journals to represent the scholars’ research output. With this choice, instead of using more complex descriptors, we will avoid the problems shown in (Salton & Buckley, 1988). However, this will force us to use as descriptors the journals in which the author has published, which may not determine a complete identification of an author’s scientific results. This balance between precision and complexity will allow us to develop an operational representation of academics’ research output.
Taking all this into account, and with the aim of improving the use of academic journals as descriptors, it will be mandatory to differentiate the most important journals from those that are less important when grouping authors based on their shared academic journals, using journal coupling. Therefore, in our modeling we are going to introduce journal weights that allow us to make distinctions between those venues in which an academic has published, based on their value as a descriptor when using journal coupling. This will be shown in the following sections, where we present the generation of effective weighting factors attached to journals that act as author descriptors.
Therefore, in our problem, for each author in a discipline or research area, we record the academic journals in which these authors published their articles. Next, using the list of academic journals in which scholars have published their research results, we construct a journal-by-author matrix. In this matrix, each row contains the weights of the individual journals for each of the authors considered (see Section 3.1). These weights must take into account that journals in which a large proportion of academics have published are a poor indicator of the similarity between two specific authors. That is why in our model, as shown in Section 3.1, we will calculate the weights associated with journal descriptors using the inverse frequency method (Salton & Buckley, 1988).
Based on this journal-by-author matrix, we next measure the similarity between two authors using the approach for calculating the similarity between two documents proposed by (Ahlgren & Colliander, 2009). However, the similarity between authors can be obtained using two different approximations as briefly described below (see Section 3.2 for more details). Specifically, first-order similarities can be obtained by measuring the similarity between columns in a journal-by-author matrix. In this first-order approach, one focuses on the direct similarity between two authors. However, we can also obtain similarities by measuring the similarity between columns in this first-order author-by-author similarity matrix (see Section 3.2). This approach produces a new author-by-author similarity matrix, populated with second-order similarities. This second-order approach finds that two authors are similar by detecting that there are other scholars such that both authors are similar to each of them. In a scientometric context,Ahlgren and Colliander (2012), and Janssens (2007) found that the second-order approach works better than the first-order.
There is only one step left to execute if we want to obtain the clusters of authors that arise from the second-order similarities. Clustering analysis is then used to group the authors based on the similarity values (see Section 3.3). More precisely, we used hierarchical clustering and the complete linkage method to cluster analysis (Everitt et al., 2001). In this article, we illustrate the performance of our approach using a set of 302 authors who have published articles in the IS&LS category of the Web of Science Core Collection (see Section 4). For this group of scholars, we found four main invisible colleges that are made up of authors who frequent the same journals, and therefore, share similar conceptual frameworks.
All the software used in this analysis and the raw data collected are available at https://github.com/rosadecsai/grouping_authors.git.
We assume a given set of authors A = {ai} that we want to study to find invisible colleges in a research area. In our study, the basic assumption is that relationships between the research output of different authors in A can be found by comparing the academic journals in which these authors published their work. Therefore, two authors in A are related if they share similar publication venues (journal publication profiles). Our grouping of authors is then based on journal coupling.
In our model, J = {jm} represents the list of journals where authors in A published their work. We then constructed a journal-by-author matrix W = {wm,i}, with wm,i being the weight for representing the ai’s research output using as descriptor the journal jm. Therefore, the author ai can be represented using a vector of descriptors and journal weights as follows (see (Salton & Buckley, 1988) for further details):
In our model, each journal receives a weight that represents its level of relevance as a descriptor of the author. The journal weights wm,i can vary their value between 0 and a maximum value. This determines a greater degree of discrimination between the journals of J that intervene as descriptors. Thus, the most relevant journals as descriptors of an author receive a greater weight. On the contrary, academic journals that are less relevant as author descriptors will receive lower weight assignments, close to 0, or even a value of 0 if the author has not published any work in the journal. In this approach, therefore, there is no threshold for considering a journal to be relevant.
An author may have published several articles in different academic journals. However, when we compare this author with the rest in A, not all of these journals will be equally relevant when acting as descriptors of the author’s scientific production. In that sense, the best descriptors will be those journals that are truly capable of distinguishing certain authors from the rest in A. This implies that the best journals jm for representing the research output of author ai should have high journal frequencies (i.e., journals in which ai frequently published their work) but low overall frequencies across authors in A.
In our model, freqm,i represents the number of papers that author ai published in journal jm. Since a journal in which a large proportion of scholars published their work often is a bad indicator of similarity between authors, it is reasonable to weight a journal jm in accordance with how frequently different authors in A published their work in this journal. In our model, we used the inverse frequency factor to perform this function (Salton & Buckley, 1998):
Therefore, in our model, the value of journal jm as descriptor of author ai (wm,i) is the product of the journal frequency and the inverse frequency factor (Ahlgren & Colliander, 2009; Salton & Buckley, 1988):
Now, the similarity between authors in A is determined based on the number of shared journals between these authors. That is, the similarity value is measured using journal coupling. Recall that, following (Garcia et al., 2012; Ni et al., 2013), the underlying assumption of our approach is that authors who publish in the same journals are more likely to share similar conceptual frameworks, and thus to be part of the same invisible college, than those who never publish in the same venues.
From equation (1), the similarity between authors ai and aj in A could be calculated using the vector product formula. However, this formula does not take into account that the weights of the journals must depend to some extent on the values of the weights assigned to the other journals in the same vector. In our model, we solve this problem by using a length normalized journal-weighting system. However, introducing this normalization, as demonstrated in (Baeza-Yates & Ribeiro-Neto, 1999), the value of the similarity between two authors turns out to be the cosine of the angle between the two journal vectors that represent authors ai and aj:
This first-order approach measures the similarity between two authors directly using their journal vector representation. However, alternatively, a second-order strategy finds that two authors are similar by detecting that there are other scholars in A such that both authors are similar to each of them. Thus, using equation (4), a second-order similarity matrix is defined as follows (see (Ahlgren & Colliander, 2009) for more details):
The last stage of our approach (before interpreting the results obtained) will be to group the authors in A using agglomerative hierarchical clustering. In this effort, we first obtain dissimilarity values between authors in A. For its calculation we use the second-order similarity values obtained in the equation (5). The dissimilarity value is obtained by subtracting the given similarity value from 1.
In our clustering analysis, we will initially have as many clusters as there are individual authors in A. That is, each author determines a single cluster. However, subsequently and at each stage of the agglomerative grouping, the closest clusters according to a measure of distance between them, join together forming a new, larger group of authors. Agglomerative hierarchical clustering stops when a single author group remains, which will consist of all authors of A.
In our analysis, the distance measure between clusters is obtained using the complete linkage method proposed in (Everitt et al., 2001). Therefore, the distance between two clusters is obtained by calculating the maximum distance between pairs of authors, the first of them belonging to one cluster and the second author to the other cluster. To do this, at each stage of the agglomerative clustering, the distance between two groups of authors is calculated as the maximum second-order dissimilarity between two authors, one from the first group and one from the second.
In our study, we used the SciPy package (Virtanen et al., 2020) to perform the complete linkage clustering based on second-order dissimilarities between authors in A. The components at each iterative step of the agglomerative clustering are always a subset of authors or groups of authors. Hence, a tree diagram, or dendrogram will be used to represent the grouping of authors. At a given level of the clustering, the clusters that exist above and below a grouping threshold are obtained simply using horizontal slices of the tree. Thus, in the next section, dendrograms illustrate the agglomerative hierarchical clustering of authors based on journal coupling. However, what will be the cutoff threshold defined in the agglomerative hierarchical clustering? In our experiments we used the maxclust criterion in fcluster of the SciPy package (Virtanen et al., 2020): fcluster (Z, nu mclust, criterion=‘maxclust’). Using the maxclust criterion, we find an optimal distance between each pair of points which are going to be in the same cluster. All the software used in this analysis is available at https://github.com/rosadecsai/grouping_authors.git.
In this section, we illustrate the grouping of authors using journal coupling. The results shown in this paper are from a controlled exercise. We analyzed 302 authors who had published in the IS&LS category of the Web of Science Core Collection. These were authors who published an article between 2022 and 2024 or who published a highly cited article between 2014 and 2024.
We downloaded the complete list of academic journals in which they had published all their works. For each author, we retrieved all the scientific journals in which this author had published his/her articles. We then used the cosine measure to calculate the similarity between authors (both first and second order).
Figure 1 illustrates the dendrogram of IS&LS authors according to their journal publication profile (using journal coupling as described above). To obtain this dendrogram of IS&LS authors, we used the complete linkage method for clustering the 304 IS&LS authors, using second-order dissimilarities. Based on the maxclust criterion of the SciPy package (Virtanen et al., 2020), we found four distinct clusters according to similarities in their research output. Authors who frequent the same journals are more likely to be part of the same invisible college (Price & Beaver, 1966). This is because by publishing in the same journals, they share similar conceptual frameworks. Thus, from the four distinct clusters in Figure 1 using the maxclust criterion, we found four invisible colleges in the list of authors. Figure 2 illustrates a visualization in VOSviewer of these four invisible colleges. This map is a node-edge diagram, which places each author on the plane and links pairs of scholars using their corresponding author-author similarities.
Dendrogram of IS&LS authors according to their journal publication profile.
Visualization in VOSviewer of a map for IS&LS authors using journal coupling.
The four invisible colleges using journal coupling of authors were: (i) “Information Systems” invisible college (see Figure 3); (ii) “Business and Information Management” invisible college (see Figure 4); (iii) “Quantitative Information Science” invisible college (see Figure 5); and (iv) “Library Science” invisible college (see Figure 7). From the dendrogram of the agglomerative hierarchical clustering (see Figure 1), we see that the invisible colleges of “Quantitative Information Science” and “Information Systems” determine two very strong groupings. However, the invisible colleges of “Library Science” and “Business and Information Management” are relatively less strong clusters and show a more diversified journal publication profile (see Figures 3, 4, 5, and 7). Tables 1–4 illustrate the 100 journals of higher weight to determine each invisible college based on journal coupling.
Visualization in VOSviewer of the ‘Information Systems’ invisible college.
Visualization in VOSviewer of the “Business and Information Management” invisible college.
Visualization in VOSviewer of the “Quantitative Information Science” invisible college.
| Information Systems | |||||
|---|---|---|---|---|---|
| Rank | Journal Abbreviation | Weight | Rank | Journal Abbreviation | Weight |
| 1 | INFORM SYST RES | 153.88 | 51 | J GLOB INF MANAG | 6.33 |
| 2 | INFORM SYST J | 122.69 | 52 | J ORGAN END USER COM | 6.24 |
| 3 | J MANAGE INFORM SYST | 117.14 | 53 | ELECTRON COMMER RES | 6.24 |
| 4 | J ASSOC INF SYST | 110.90 | 54 | J BUS ETHICS | 6.24 |
| 5 | EUR J INFORM SYST | 104.67 | 55 | EUR J OPER RES | 6.24 |
| 6 | MIS QUART | 75.66 | 56 | MANAGE SCI | 6.04 |
| 7 | P ANN HICSS | 71.39 | 57 | INT FED INFO PROC | 5.75 |
| 8 | INFORM MANAGE-AMSTER | 52.36 | 58 | J ELECTRON COMMER RE | 5.75 |
| 9 | BUS INFORM SYST ENG+ | 49.91 | 59 | UROLOGY | 5.55 |
| 10 | J STRATEGIC INF SYST | 38.82 | 60 | PERS INDIV DIFFER | 5.55 |
| 11 | J INF TECHNOL-UK | 38.12 | 61 | BEHAV BRAIN RES | 5.55 |
| 12 | DECISION SCI | 36.04 | 62 | J ENDOUROL | 5.55 |
| 13 | INT J INFORM MANAGE | 30.78 | 63 | DES ISSUES | 5.55 |
| 14 | DECIS SUPPORT SYST | 29.06 | 64 | CHINA ECON REV | 5.55 |
| 15 | INFORM MANAGE | 26.34 | 65 | TELECOMMUN POLICY | 5.55 |
| 16 | J UROLOGY | 26.34 | 66 | LECT NOTES BUS INF P | 5.47 |
| 17 | DATA BASE ADV INF SY | 25.65 | 67 | J ORG COMP ELECT COM | 4.89 |
| 18 | PROD OPER MANAG | 24.95 | 68 | INT J HUM-COMPUT ST | 4.89 |
| 19 | ELECTRON MARK | 22.18 | 69 | IFIP ADV INF COMM TE | 4.85 |
| 20 | OMEGA-INT J MANAGE S | 19.41 | 70 | J INFORM TECHNOL | 4.16 |
| 21 | MIS Q EXEC | 19.41 | 71 | PERS PSYCHOL | 4.16 |
| 22 | COMMUN ACM | 19.27 | 72 | RES TECHNOL MANAGE | 4.16 |
| 23 | ORGAN SCI | 18.02 | 73 | J TRADIT CHIN MED | 4.16 |
| 24 | MIT SLOAN MANAGE REV | 15.25 | 74 | ECON MODEL | 4.16 |
| 25 | ELECTR J INF SYS DEV | 14.56 | 75 | J MARKETING RES | 4.16 |
| 26 | INT J ELECTRON COMM | 13.86 | 76 | J ASIAN NAT PROD RES | 4.16 |
| 27 | IEEE T ENG MANAGE | 13.52 | 77 | WORLD ECON | 4.16 |
| 28 | ADV MANAG INFORM SYS | 13.17 | 78 | INT J PROD RES | 4.16 |
| 29 | J BUS RES | 13.17 | 79 | IEEE T SYST MAN CY C | 4.16 |
| 30 | GROUP DECIS NEGOT | 13.17 | 80 | CHIN MED-UK | 4.16 |
| 31 | IND MANAGE DATA SYST | 12.48 | 81 | J INFECT DIS | 4.16 |
| 32 | IT PROF | 12.48 | 82 | J DECIS SYST | 4.16 |
| 33 | L N INF SYST ORGAN | 12.48 | 83 | CHANDOS ASIAN STUD | 4.16 |
| 34 | COMPUT SECUR | 12.48 | 84 | INT J ACCOUNT INF MA | 4.16 |
| 35 | PAC ASIA J ASSOC INF | 11.78 | 85 | J POLYNESIAN SOC | 4.16 |
| 36 | WIRTSCHAFTSINF | 11.09 | 86 | J ORGAN EFF-PEOPLE P | 4.16 |
| 37 | ELECTRON COMMER R A | 10.07 | 87 | ACM TRANS MANAG INF | 4.16 |
| 38 | SMALL GR RES | 9.70 | 88 | FEBS LETT | 4.16 |
| 39 | INFORM SYST MANAGE | 9.70 | 89 | BIOCHEM INT | 4.16 |
| 40 | ORGAN BEHAV HUM DEC | 9.70 | 90 | J SMALL BUS MANAGE | 3.47 |
| 41 | DATA BASE | 9.70 | 91 | J OPER RES SOC | 3.47 |
| 42 | J INF TECHNOL | 9.01 | 92 | ACAD MANAGE J | 3.47 |
| 43 | J OPER MANAG | 9.01 | 93 | INF SYST E-BUS MANAG | 3.47 |
| 44 | INFORM ORGAN-UK | 9.01 | 94 | J AM MED INFORM ASSN | 3.47 |
| 45 | J APPL PSYCHOL | 8.32 | 95 | INFORM TECHNOL DEV | 3.47 |
| 46 | IEEE T PROF COMMUN | 7.62 | 96 | ANN OPER RES | 3.47 |
| 47 | COMPUT HUM BEHAV | 7.19 | 97 | INFORM SOC | 3.47 |
| 48 | J MARKETING | 6.93 | 98 | FRONT PSYCHOL | 2.88 |
| 49 | J ACAD MARKET SCI | 6.93 | 99 | J GLOB INF TECH MAN | 2.88 |
| 50 | PEDIATR INFECT DIS J | 6.93 | 100 | BIOMED PHARMACOTHER | 2.77 |
| Business and Information Management | |||||
|---|---|---|---|---|---|
| Rank | Journal Abbreviation | Weight | Rank | Journal Abbreviation | Weight |
| 1 | J BUS RES | 172.59 | 51 | REV MANAG SCI | 19.41 |
| 2 | J KNOWL MANAG | 135.86 | 52 | BENCHMARKING | 19.41 |
| 3 | INT J INFORM MANAGE | 92.92 | 53 | SPRINGERBRIEF BUS | 19.41 |
| 4 | PROD PLAN CONTROL | 85.95 | 54 | J CONSUM BEHAV | 19.41 |
| 5 | J HOSP MARKET MANAG | 70.70 | 55 | INT J PROD ECON | 18.99 |
| 6 | INT J CONTEMP HOSP M | 70.01 | 56 | J TECHNOL TRANSFER | 18.71 |
| 7 | J TRAVEL RES | 69.31 | 57 | IEEE T ENG MANAGE | 18.41 |
| 8 | ANN OPER RES | 67.24 | 58 | INNOV PUBLIC SECT | 18.02 |
| 9 | TECHNOL FORECAST SOC | 66.74 | 59 | SERV IND J | 17.33 |
| 10 | IND MANAGE DATA SYST | 64.46 | 60 | INT J E-BUS RES | 16.64 |
| 11 | ANN TOURISM RES | 59.61 | 61 | ECON RES-EKON ISTRAZ | 16.64 |
| 12 | INT J HOSP MANAG | 59.61 | 62 | J STRATEG MARK | 16.64 |
| 13 | J SUSTAIN TOUR | 52.68 | 63 | TECHNOVATION | 16.64 |
| 14 | INT J INDIAN CULT BU | 51.29 | 64 | J DESTIN MARK MANAGE | 16.64 |
| 15 | INT J ELECTRON GOV R | 49.21 | 65 | PUB ADMIN INF TECH | 15.94 |
| 16 | INT J PROD RES | 49.21 | 66 | CREAT INNOV MANAG | 15.25 |
| 17 | BUS STRATEG ENVIRON | 48.52 | 67 | ENTERP INF SYST-UK | 15.25 |
| 18 | COMPUT HUM BEHAV | 46.03 | 68 | INT J TOUR RES | 15.25 |
| 19 | J INTELLECT CAP | 44.36 | 69 | J UNIVERS COMPUT SCI | 15.25 |
| 20 | INFORM SYST MANAGE | 44.36 | 70 | EUR MANAG J | 15.25 |
| 21 | J ENTERP INF MANAG | 40.90 | 71 | TOUR ANAL | 15.25 |
| 22 | INT J ENTREP BEHAV R | 40.20 | 72 | INT J INNOV LEARN | 13.86 |
| 23 | GOV INFORM Q | 39.70 | 73 | RESOUR CONSERV RECY | 13.86 |
| 24 | IND MARKET MANAG | 38.12 | 74 | INT MARKET REV | 13.86 |
| 25 | EUROMED ACAD BUS CON | 37.43 | 75 | INT J CONSUM STUD | 13.86 |
| 26 | INT J MOB COMMUN | 36.04 | 76 | EUR MANAG REV | 13.86 |
| 27 | INT J BANK MARK | 33.96 | 77 | INT J RETAIL DISTRIB | 13.86 |
| 28 | J RETAIL CONSUM SERV | 33.66 | 78 | J BUS IND MARK | 13.86 |
| 29 | TOUR REV | 33.27 | 79 | J HOSP TOUR TECHNOL | 13.86 |
| 30 | BRIT FOOD J | 33.27 | 80 | INT J OPER PROD MAN | 13.86 |
| 31 | INT ENTREP MANAG J | 31.88 | 81 | ELECTRON MARK | 13.17 |
| 32 | TRANSFORM GOV-PEOPLE | 31.19 | 82 | EUR J INFORM SYST | 13.17 |
| 33 | P ANN HICSS | 30.50 | 83 | MANAGE DECIS | 12.95 |
| 34 | EUR J INT MANAG | 29.11 | 84 | J SERV MARK | 12.48 |
| 35 | PSYCHOL MARKET | 29.11 | 85 | INT J LOGIST MANAG | 12.48 |
| 36 | J HOSP TOUR RES | 27.73 | 86 | BRIT J MANAGE | 12.48 |
| 37 | EUR J MARKETING | 27.73 | 87 | PUBLIC MANAG REV | 12.48 |
| 38 | BUS PROCESS MANAG J | 27.73 | 88 | INT J EMERG MARK | 12.48 |
| 39 | TOTAL QUAL MANAG BUS | 26.34 | 89 | J PROD BRAND MANAG | 12.48 |
| 40 | ADV THE PRAC EMER MA | 26.34 | 90 | SUSTAINABILITY-BASEL | 12.37 |
| 41 | J HOSP TOUR MANAG | 26.34 | 91 | EUR J INNOV MANAG | 11.78 |
| 42 | J TRAVEL TOUR MARK | 26.34 | 92 | INT J INNOV MANAG | 11.09 |
| 43 | ROUTLEDGE STUD MARK | 24.95 | 93 | FOOD QUAL PREFER | 11.09 |
| 44 | INNOV TECH KNOWL MAN | 24.95 | 94 | IFAC PAPERSONLINE | 11.09 |
| 45 | INFORM MANAGE-AMSTER | 23.88 | 95 | CORP SOC RESP ENV MA | 11.09 |
| 46 | ASIA PAC J TOUR RES | 23.57 | 96 | INT J ENTREP VENTUR | 11.09 |
| 47 | J INNOV KNOWL | 22.18 | 97 | COMPUT IND ENG | 11.09 |
| 48 | TOUR MANAG PERSPECT | 20.79 | 98 | J ORGAN BEHAV | 11.09 |
| 49 | TOURISM MANAGE | 20.43 | 99 | J INT MANAG | 11.09 |
| 50 | CURR ISSUES TOUR | 20.10 | 100 | J INT CONSUM MARK | 11.09 |
| Quantitative Information Science | |||||
|---|---|---|---|---|---|
| Rank | Journal Abbreviation | Weight | Rank | Journal Abbreviation | Weight |
| 1 | J INFORMETR | 366.67 | 51 | FRONT INFORM TECH EL | 4.16 |
| 2 | PRO INT CONF SCI INF | 234.28 | 52 | ACM-IEEE J CONF DIG | 4.16 |
| 3 | PROF INFORM | 99.12 | 53 | INVESTIG BIBLIOTECOL | 4.16 |
| 4 | QUANT SCI STUD | 74.86 | 54 | CAN J SOCIOL | 4.16 |
| 5 | RES EVALUAT | 42.98 | 55 | ACTA PHYS POL A | 4.16 |
| 6 | MATH COMPUT MODEL | 34.66 | 56 | INT CONF BIG DATA | 4.16 |
| 7 | J DATA INFO SCI | 31.19 | 57 | THEOR CHEM ACC | 4.16 |
| 8 | J AM SOC INFORM SCI | 29.11 | 58 | NAT HUM BEHAV | 4.16 |
| 9 | SPRINGER HBK | 29.11 | 59 | HIGH EDUC | 4.03 |
| 10 | REV ESP DOC CIENT | 27.03 | 60 | INFORM RES | 4.03 |
| 11 | SCI PUBL POLICY | 26.34 | 61 | DATA KNOWL ENG | 3.47 |
| 12 | ONLINE INFORM REV | 16.40 | 62 | LIBRI | 3.47 |
| 13 | J DOC | 14.67 | 63 | HUM SOC SCI COMMUN | 3.47 |
| 14 | COLLNET J SCIENTOMET | 14.56 | 64 | COMUNICAR | 3.47 |
| 15 | CURR SCI INDIA | 13.86 | 65 | RES POLICY | 3.45 |
| 16 | EMBO REP | 13.86 | 66 | ELECTRON LIBR | 3.16 |
| 17 | LEARN PUBL | 13.17 | 67 | J SUPERCOMPUT | 2.77 |
| 18 | ASLIB J INFORM MANAG | 12.95 | 68 | J ORTHOP SURG RES | 2.77 |
| 19 | P ASIST ANNU | 11.09 | 69 | ECON POLIT-ITALY | 2.77 |
| 20 | J CHEM PHYS | 11.09 | 70 | CHIMIA | 2.77 |
| 21 | ELIFE | 9.70 | 71 | DATA | 2.77 |
| 22 | SCI ENG ETHICS | 9.70 | 72 | J LIBR INFORM STUD | 2.77 |
| 23 | J SCIENTOMETR RES | 9.70 | 73 | INT J HEALTH POLICY | 2.77 |
| 24 | ASLIB PROC | 9.70 | 74 | PHYS LETT B | 2.77 |
| 25 | NATURE | 9.21 | 75 | INFORM VISUAL | 2.77 |
| 26 | CLIMATE | 8.32 | 76 | MATDEMATICS-BASEL | 2.77 |
| 27 | REV ESP SALUD PUBLIC | 8.32 | 77 | UNIV PSYCHOL | 2.77 |
| 28 | J ECON SURV | 6.93 | 78 | J EVOL ECON | 2.77 |
| 29 | BMC BIOINFORMATICS | 6.93 | 79 | J ALZHEIMERS DIS | 2.77 |
| 30 | CRIMINOLOGIE | 6.93 | 80 | PAC J MATH | 2.77 |
| 31 | PRESSE MED | 6.93 | 81 | STUD CLASS DATA ANAL | 2.77 |
| 32 | LIBR INFORM SCI RES | 6.33 | 82 | RES HIGH EDUC | 2.77 |
| 33 | J INTELL FUZZY SYST | 6.24 | 83 | ACTA PAEDIATR | 2.77 |
| 34 | LIBR INFORM SCI SER | 6.24 | 84 | ANN I H POINCARE B | 2.77 |
| 35 | MEAS-INTERDISCIP RES | 5.55 | 85 | INT J STROKE | 2.77 |
| 36 | IEEE INT CON MULTI | 5.55 | 86 | ESTUD MENSAJE PERIOD | 2.77 |
| 37 | CAN J INFORM LIB SCI | 5.55 | 87 | PALGR COMMUN | 2.77 |
| 38 | HIGH EDUC Q | 5.55 | 88 | SOC STUD SCI | 2.77 |
| 39 | Z EVAL | 5.55 | 89 | EVALUATION REV | 2.77 |
| 40 | J ORGANOMET CHEM | 5.55 | 90 | PHYS REV LETT | 2.77 |
| 41 | IEEE INT CONF FUZZY | 5.55 | 91 | PSICOTHEMA | 2.77 |
| 42 | PUBLICATIONS | 5.55 | 92 | IEEE INT C BIO BIO W | 2.77 |
| 43 | J LIBR INF SCI | 5.55 | 93 | IETE TECH REV | 2.77 |
| 44 | MALAYS J LIBR INF SC | 5.18 | 94 | MED CLIN-BARCELONA | 2.77 |
| 45 | COLL RES LIBR | 4.85 | 95 | J TECHNOL TRANSFER | 2.77 |
| 46 | J KOREAN MED SCI | 4.16 | 96 | TECHNOVATION | 2.77 |
| 47 | INT CONF CONTEMP | 4.16 | 97 | EDUC XX1 | 2.77 |
| 48 | MINERVA | 4.16 | 98 | AIP CONF PROC | 2.77 |
| 49 | KNOWL ORGAN | 4.16 | 99 | INFORM POL | 2.77 |
| 50 | FEMS MICROBIOL LETT | 4.16 | 100 | DOC BIBL | 2.77 |
| Library Science | |||||
|---|---|---|---|---|---|
| Rank | Journal Abbreviation | Weight | Rank | Journal Abbreviation | Weight |
| 1 | INT J GEOGR INF SCI | 198.24 | 51 | HEALTH AFFAIR | 18.02 |
| 2 | J AM MED INFORM ASSN | 120.61 | 52 | COLLECT BUILD | 18.02 |
| 3 | LEARN PUBL | 113.68 | 53 | PEDIATRICS | 17.33 |
| 4 | PROF INFORM | 92.19 | 54 | APPL CLIN INFORM | 17.33 |
| 5 | CHANDOS INF PROF SER | 83.18 | 55 | IEEE J-STARS | 16.64 |
| 6 | J LIBR ADM | 67.93 | 56 | NPJ DIGIT MED | 16.64 |
| 7 | ASLIB PROC | 58.92 | 57 | PUBLIC LIBR Q | 16.64 |
| 8 | J LIBR INF SCI | 57.53 | 58 | APPL GEOGR | 15.25 |
| 9 | REV BELGE PHILOL HIS | 56.84 | 59 | AGR FOREST METEOROL | 15.25 |
| 10 | EDINB STUD CLASS ISL | 48.52 | 60 | JAMA-J AM MED ASSOC | 15.25 |
| 11 | AM HIST REV | 44.36 | 61 | MALAYS J LIBR INF SC | 15.25 |
| 12 | LIBR MANAGE | 44.36 | 62 | J DOC | 14.96 |
| 13 | EM QUESTAO | 42.98 | 63 | JAMA NETW OPEN | 13.86 |
| 14 | J AUST LIB INF ASSOC | 37.43 | 64 | GISCI REMOTE SENS | 13.86 |
| 15 | T GIS | 37.43 | 65 | INTERLEND DOC SUPPLY | 13.86 |
| 16 | LIBR HI TECH | 36.54 | 66 | INFORM SOC-ESTUD | 13.86 |
| 17 | ISPRS INT J GEO-INF | 34.66 | 67 | INT J CLIN MONIT COM | 13.86 |
| 18 | LANDSCAPE URBAN PLAN | 33.27 | 68 | DIGIT LIBR PERSPECT | 13.86 |
| 19 | J GEN INTERN MED | 31.88 | 69 | AUST ACAD RES LIBR | 12.48 |
| 20 | SYNTH REACT INORG M | 30.50 | 70 | BMJ QUAL SAF | 12.48 |
| 21 | COMPUT ENVIRON URBAN | 29.11 | 71 | NEW REV ACAD LIBR | 12.48 |
| 22 | SPECULUM | 27.73 | 72 | JOURNALISM | 12.48 |
| 23 | J ASIAN AFR STUD | 27.73 | 73 | NEW LIB WORLD | 11.09 |
| 24 | HEALTH INFO LIBR J | 26.34 | 74 | MUSLIM WORLD AGE CRU | 11.09 |
| 25 | IFLA J-INT FED LIBR | 26.34 | 75 | JAMA INTERN MED | 11.09 |
| 26 | LIBRI | 25.65 | 76 | ANN EMERG MED | 11.09 |
| 27 | IEEE T GEOSCI REMOTE | 24.95 | 77 | LANDSCAPE ECOL | 11.09 |
| 28 | REV IBERI-AM CIENC I | 24.95 | 78 | ISLAM | 11.09 |
| 29 | GLOB KNOWL MEM COMMU | 23.59 | 79 | LECT NOTES GEOINF CA | 11.09 |
| 30 | INT J REMOTE SENS | 23.57 | 80 | HEALTH SERV RES | 11.09 |
| 31 | AFR J LIBR ARCH INFO | 23.57 | 81 | EVID BASED LIB INF P | 11.09 |
| 32 | QUAL QUANT METHODS L | 23.57 | 82 | RDBCI-REV DIG BIB CI | 11.09 |
| 33 | J ACAD LIBR | 23.01 | 83 | PHOTOGRAMM ENG REM S | 11.09 |
| 34 | LIBR RESOUR TECH SER | 22.18 | 84 | INF DISCOV DELIV | 10.93 |
| 35 | J TRANSP GEOGR | 20.79 | 85 | REMOTE SENS ENVIRON | 10.40 |
| 36 | S AFR J INFORM MANAG | 20.79 | 86 | PORTAL-LIBR ACAD | 10.40 |
| 37 | ANN INTERN MED | 20.79 | 87 | ONLINE INFORM REV | 10.07 |
| 38 | LIBR REV | 20.79 | 88 | ASLIB J INFORM MANAG | 9.78 |
| 39 | B SCH ORIENT AFR ST | 20.79 | 89 | COLLECT CURATION | 9.70 |
| 40 | ANN AM ASSOC GEOGR | 20.79 | 90 | CAT CLASSIF Q | 9.70 |
| 41 | AM J MANAG CARE | 20.79 | 91 | SCI CHINA SER D | 9.70 |
| 42 | INFORM DEV | 20.71 | 92 | HEALTHCARE-J DEL SCI | 9.70 |
| 43 | J INFORM OPTIM SCI | 19.41 | 93 | J SCHOLARLY PUBL | 9.70 |
| 44 | NEW ENGL J MED | 19.41 | 94 | SERIALS REV | 9.70 |
| 45 | ELECTRON LIBR | 19.27 | 95 | PERFORM MEAS METR | 9.70 |
| 46 | S AFR J LIBR INF | 18.71 | 96 | JAMIA OPEN | 9.70 |
| 47 | PERSPECT CIENC INF | 18.02 | 97 | BIBLIOS | 9.70 |
| 48 | ENCONTROS BIBLI | 18.02 | 98 | INT J APPL EARTH OBS | 9.70 |
| 49 | PEDIATR RES | 18.02 | 99 | ARCH INTERN MED | 9.70 |
| 50 | DESIDOC J LIB INF TE | 18.02 | 100 | IEEE T NEUR NET LEAR | 9.70 |
For example, for the invisible college of “Information Systems”, the academic journals with the highest weights for representing the research output of authors are five (see Table 1): (1) INFORM SYST RES; (2) INFORM SYST J; (3) J MANAGE INFORM SYST; (4) J ASSOC INF SYST; and (5)EUR J INFORM SYST. These five journals publish studies of the highest quality in the field of information systems. Their articles promote knowledge about the design, management, use, assessment and impacts of information technologies by individuals, groups, organizations, societies and nations for the improvement of economic and social well-being. To this end, they integrate technological disciplines with social, contextual and management issues. Distinguished scholars within this invisible college would be, for example, Dennis, Davinson, Chan, Sarker, Grover, Pan, Benitez, and Lowry (see Figure 3). Table A1 (in Appendix) illustrates the complete list of authors in this invisible college of “Information Systems” using journal coupling.
However, for the invisible college of “Business and Information Management”, the academic journals with the highest weights for representing the authors’ research output are six (see Table 2): (1) J BUS RES; (2) J KNOWL MANAG; (3) INT J INFORM MANAGE; (4) PROD PLAN CONTROL; (5)J HOSP MARKET MANAG; and (6) INT J CONTEMP HOSP MANAG. These journals aim to publish research that examines a wide variety of business decision contexts, processes, and activities. These articles also focus on the challenge for information management and the emerging needs of the industry. This includes the management of activities that generate changes in the behavioral patterns of customers, people and organizations, as well as information that leads to changes in the way people use information to participate in knowledge-focused activities. Distinguished scholars within this invisible college of “Business and Information Management” would be, for example, Pandey, Dwivedi, Dennehy, Meissner, Kar, Khare, Pappas, Papadopoulos, and Popa (see Figure 4). Table A2 (in Appendix) illustrates the complete list of authors in this invisible college of “Business and Information Management” using journal coupling.
For the invisible college of “Quantitative Information Science”, the academic journals with the highest weights for representing the research output are four (see Table 3): (1) J INFORMETR; (2) PRO INT CONF SCI INF; (3) PROF INFORM; and (4) QUANT SCI STUD. This group of authors included several scholars who are among the recipients of the Derek de Solla Price Memorial Medal, the first and the most important international prize in scientometric studies, e.g., Bornmann, Thelwall, Waltman, Rousseau, and Egghe (see Figure 5). In this group we also found some of the most significant and influential LIS scholars during the time examined, e.g., Lariviere, Costas, Sugimoto, D’Angelo, Abramo, Wouters, Zahedi, Delgado Lopez-Cozar, Moya-Anegón, and so on (see Figure 5). The academics in this cluster have a high level of production and more links with the rest of the authors who seem to surround them. Table A3 (Appendix) illustrates the complete list of authors in this invisible college of “Quantitative Information Science” using journal coupling.
In Figure 6, we can see how high levels of similarity correspond with similar journal publication profiles: Bornmann’s and Rousseau’s research output is very similar (similarity = 0.99) and highly focused on four main journals (SCIENTOMETRICS, J INFORMETR, J ASSOC INF SCI TECH/J AMSOC INF SCI TEC, PRO INT CONF SCI INF) which contain more than 64% of the total production for both authors. In the same Figure 6, we found that the similarity between Bornmann and Stuart is much smaller (similarity = 0.66). However, for both authors, seven journals contain more than 43% of their production (see Figure 6).
Detail of journal similarities among Bornmann, Rousseau, and Stuart according to the IS&LS category of the Web of Science Core Collection.
For the invisible college of “Library Science”, the academic journals with the highest weights for representing the authors’ research output are eight (see Table 4): (1) INT J GEOGR INF SCI; (2) J AM MED INFORM ASSN; (3) LEARN PUBL; (4) PROF INFORM; (5) CHANDOS INF PROF SER; (6) J LIBR ADM; (7) ASLIB PROC; and (8) J LIBR INF SCI. These journals publish articles that reflect all aspects of library and information science. They focus on the practices, perspectives and tools of management, information technology, education and other areas of libraries. This includes the collection, organization, preservation and dissemination of information resources. They also cover the political economy of information, as well as the design, implementation and use of geographic information, medical in-formation and other special libraries for monitoring, prediction and decision making. Distinguished scholars within this invisible college of ‘Library Science’would be, for example, Zipf, Ho, Nicholas, Yuvaraj, Holley, Verma, Jamali, Rodriguez-Bravo, and Boukacem-Zeghmouri (see Figure 7). Table A4 (In Appendix) illustrates the complete list of authors in this invisible college of “Library Science” using journal coupling.
Visualization in VOSviewer of the “Library Science” invisible college.
For the invisible college of library science, Figure 8 illustrates some details of journal similarities among Batool, Warraich, and Onyancha according to the IS&LS category of the Web of Science Core Collection. We can see how high and low levels of similarity correspond with certain journal publication profiles. For example, Batool’s and Warraich’s research output is similar (similarity = 0.93) and highly focused on seven main journals which contain 53% and 45% of the total production, respectively. In the same Figure 8, we found that the similarity between Batool and Onyancha is smaller (similarity = 0.63). However, only five journals still contain 44% and 38% of their production, respectively (see Figure 8).
Detail of journal similarities among Batool, Warraich, and Onyancha according to the IS&LS category of the Web of Science Core Collection.
Figure 9 illustrates the map of invisible colleges according to their journal publication profile in IS and LS. This map is a node-edge diagram, which places each invisible college on the plane and links pairs of invisible colleges using their second-order similarities. On one hand, the invisible college “Quantitative Information Science” has the highest production by author (represented by node size) and a moderately strong link to Library Science. On the other hand, Information Systems and Business & Information Management are (relatively) less productive invisible colleges that show similarities with each other in certain fields of activity. They have weak links to Quantitative Information Science and Library Science (see Figure 9).
Map of invisible colleges according to their journal publication profile in IS and LS.
However, what are the academic journals that demonstrate that IS and LS determine a single field of research, since authors from all the invisible colleges publish in these journals? Are there academic journals of this type and what are they? Table 5 shows the list of academic journals in which authors from all the invisible colleges have published. In this table we see that authors from all invisible colleges have published in Plos One and Lect Notes Compt Sc. Academic journals such as Expert Syst Appl, Internet Res, Technol Forecast Soc, or Telemat Inform also appear on this list. But what are the specific journals of the research area in which authors from all the invisible colleges publish? There are seven main journals that stand out in this regard and they are J Am Soc Inf Sci Tec/J Assoc Inf Sci Tech, Scientometrics, Int J Inform Manage, Inform Process Manag, Inform Technol Peopl, Inform Syst Front, and J Inf Sci (see Table 5). These journals tell us about the existence of a unique research field, which appears to encompass several subfields of research (i.e., the detected invisible colleges).
| Journal Abbreviation | Inform Syst | Bus & Inform Manag | Quant Inform Sci | Libr Sci |
|---|---|---|---|---|
| ADV MATER RES-SWITZ | 7 | 3 | 2 | 10 |
| BEHAV INFORM TECHNOL | 20 | 39 | 1 | 2 |
| COMM COM INF SC | 4 | 9 | 8 | 14 |
| COMMUN ACM | 67 | 12 | 1 | 2 |
| COMMUN ASSOC INF SYS | 123 | 22 | 1 | 8 |
| COMPUT EDUC | 2 | 13 | 1 | 3 |
| ENVIRON SCI POLLUT R | 1 | 5 | 1 | 20 |
| EXPERT SYST APPL | 2 | 82 | 3 | 28 |
| IEEE ACCESS | 1 | 8 | 3 | 11 |
| IEEE T ENG MANAGE | 47 | 72 | 1 | 3 |
| INFORM PROCESS MANAG | 10 | 5 | 79 | 42 |
| INFORM SCIENCES | 2 | 3 | 2 | 12 |
| INFORM SYST FRONT | 25 | 207 | 1 | 11 |
| INFORM TECHNOL MANAG | 1 | 7 | 2 | 1 |
| INFORM TECHNOL PEOPL | 35 | 79 | 3 | 7 |
| INT J ENV RES PUB HE | 5 | 7 | 2 | 15 |
| INT J HUM-COMPUT INT | 2 | 28 | 2 | 3 |
| INT J INFORM MANAGE | 107 | 366 | 3 | 34 |
| INT J MED INFORM | 2 | 4 | 1 | 27 |
| INTERNET RES | 55 | 81 | 4 | 8 |
| J AM SOC INF SCI TEC | 20 | 7 | 252 | 18 |
| J ASSOC INF SCI TECH | 9 | 10 | 256 | 35 |
| J CLEAN PROD | 2 | 78 | 1 | 116 |
| J COMPUT INFORM SYST | 10 | 65 | 1 | 5 |
| J COMPUT-MEDIAT COMM | 1 | 3 | 3 | 7 |
| J ENVIRON MANAGE | 2 | 7 | 1 | 34 |
| J INF SCI | 3 | 5 | 61 | 66 |
| J MED INTERNET RES | 9 | 10 | 3 | 24 |
| LECT NOTES ARTIF INT | 3 | 3 | 4 | 10 |
| LECT NOTES COMPUT SC | 13 | 158 | 26 | 67 |
| NEW MEDIA SOC | 1 | 5 | 3 | 12 |
| P NATL ACAD SCI USA | 3 | 1 | 4 | 7 |
| PLOS ONE | 6 | 12 | 124 | 72 |
| SCI REP-UK | 3 | 6 | 5 | 19 |
| SCIENTOMETRICS | 2 | 7 | 846 | 41 |
| TECHNOL FORECAST SOC | 5 | 249 | 7 | 6 |
| TELEMAT INFORM | 3 | 74 | 3 | 17 |
| TOURISM MANAGE | 1 | 71 | 1 | 21 |
Our results based on journal coupling show four main invisible colleges in LIS (i.e., Library Science, Quantitative Information Science, Information Systems, and Business and Information Management). These results are consistent with those obtained in (Astrom, 2010) using co-citation maps, as well as the co-occurrence of shared references between IS and LS authors. Using citation data from publications in LIS journals, Astrom (2010) found that IS and LS determine two main subfields in a joint field of LIS, where a further division into more specialized research areas can also be found when analyzing data from IS citations. On the contrary, LS research areas were less visible in citation analyses. Thus, what are the academic journals that determine the existence of subfields in LIS? They are journals in which authors from a certain invisible college publish very frequently and in which the authors from other colleges have never published. For example, in the invisible college of Information Systems, there are four journals that stand out and determine a subfield of research due to their weight in the discrimination of IS and LS authors. They are Inform Syst Res, Inform Syst J, J Manage Inform Syst, and Eur J Inform Syst (see Table 1).
In the invisible college of Business and Information Management, there are two main academic journals that determine the existence of this research subfield, due to the frequency of publication in them (see Table 2): J Bus Res and J Knowl Manage.
In the invisible college where the most significant and influential LIS academics were found (Quantitative Information Science), the journals with the highest weights when it comes to discriminating this research subfield are fundamentally two (see Table 3): J Informetr and Pro Int Conf Sci Inf. Furthermore, Prof Inform and Quant Sci Stud also played a prominent role, the latter being a recently created journal (see Table 3).
Finally, three academic journals are the most important in determining the invisible college of Library Science, specifically, Int J Geogr Inf Sci, J Am Med Inform Assn, and Learn Publ (see Table 4). This tells us about the importance that special libraries have in this subfield, more precisely, geographic information and medical information.
Our results regarding the most relevant journals to determine the structure of information science and library science, are consistent with those of previous studies. For example, Vinkler (2019) identified among the core journals in scientometrics the following journals: Scientometrics; Journal of the American Society for Information Science and Technology; Information Processing and Management; Journal of Information Science; Research Policy; Library Trends; and Research Evaluation. Our results on the most important journals for determining the structure of the research area also coincide with those presented in (Nixon, 2014; Walters & Wilder, 2015; Weerasinghe, 2017). In a more recent paper, Safón and Docampo (2023) found that “the top 10 journals that most influenced papers published between 2021 and March 2022 were: Scientometrics; International Journal of Information Management; Journal of the Association for Information Science and Technology; Quantitative Science Studies; MIS Quarterly; Information and Management; Information Processing and Management; Journal of the Association for Information Systems; Journal of Informetrics; Journal of Academic Librarianship.” All these journals are, according to our approach based on journal coupling, of great importance either to define LIS as a field of research (see Table 5) or to determine the structure of IS and LS into invisible colleges (see Tables 1-4).
However, our approach and the results obtained not only inform the structure of information science and library science, but also provide an adaptable methodology that can be used in other areas of research and extended to answer additional research questions. For example, to the problem of finding potential reviewers in complex evaluation processes. Ultimately, this paper served as a proof of concept for finding invisible colleges in a research area, using journal coupling. Regarding limitations and further research, a first limitation of our study is that the results shown in this paper are from a controlled exercise. We analyzed 302 authors who published in the IS&LS category of the Web of Science Core Collection. In future work, we will explore alternative approaches to identifying the authors that will be used to determine the structure of the research area, and compare the results with those reported here.
Second, the analysis performed using journal coupling excludes books, book chapters, and conference papers. This is a significant omission, since contributions other than articles remain important within information science and library science. This could be an interesting point of analysis for future work.
Third, although related to the previous limitation, in this paper, academic journals alone were used for research output representation. However, as suggested in (Garcia et al., 2012), “sets of journals cannot provide complete identifications of research output.” In future work, we will explore the incorporation of more complex entities for author representation.