Quantifying the internationality and multidisciplinarity of authors and journals using ecological statistics

Abstract

Authors or journals often claim internationality or multidisciplinarity based on assertion or qualitative evidence, while scientometric studies employ sophisticated analyses or software beyond the resources of occasional users to assess these concepts. This paper demonstrates how statistics used to describe ecological communities can be applied to bibliometric data to quantify internationality or multidisciplinarity for individuals and journals, enabling tests of statistical significance using graphical user interface freeware accessible to even occasional users. Margalef Richness, diversity and evenness or equitability can be calculated to indicate whether papers or citations come predominantly from a small group of countries or disciplines, or are more widely distributed. Tests of statistical significance for differences in Margalef richness, diversity or evenness between authors or journals enable testing of diverse hypotheses including, for example: differences in internationality or multidisciplinarity between authors or between journals; or changes over time in these variables for authors or journals (perhaps in response to career changes or changes in editorial policy). Quantifying internationality and multidisciplinarity in an accessible way for many potential users, with the possibility of statistical hypothesis testing, is a significant advance over assertion and qualitative description on the one hand or conceptually and practically complex analysis on the other.

Appendices

Appendix 1: Procedure for profiling international citations or subject area citations of an author or journal in WoSCRS

Before undertaking a search, it is prudent to check the year range of the WoS subscription to ensure that it embraces the years over which the target author published, or the years over which a journal profile is desired (for more detail, refer to http://clarivate.libguides.com/woscc/coverage).

Procedure for authors

In brief, the procedure involves conducting a Cited Reference Search for the chosen author in the WoS (Core Collection), retrieving the citations to the work of the chosen author, and then profiling the countries of affilation for the authors in the citations. The steps below apply to the current version of the WoS (Core Collection), but interface changes could occur.

• 1. In the WoS (Core Collection) choose the CRS option from the dropdown menu of search options and enter the search term for the desired author.

• 2. Scan the output and select all entries that are publications of the target author. This may be difficult when the searcher has no knowledge of the target author’s full curriculum vitae (Jacsó 2007; Calver et al. 2013). There is no need to locate and aggregate cases where the same publication has multiple entries because of errors by citing authors—the country of origin of the citing authors will be identified correctly.

• 3. Once all citations attributed correctly to the target author have been selected, clicking the finish search button displays all the citations to the work of the target author. Clicking the ‘analyze results’ button on the screen advances to the next stage. There is no need to set the checkboxes of all the references displayed before clicking analyse results.

• 4. The results analysis screen indicates the number of papers citing a work by the target author. Set the ‘display options’ to 500 results with a minimum record count of one, and the ‘sort by’ option to ‘record count’. Then in the ‘rank the records by this field’ box, select ‘countries/territories’ for an international analysis or ‘Web of Science categories’ for a multidisciplinary analysis. Click the analyse button. The percentages in the resulting output indicate the percentage of records (citations to the target author’s work) that include at least one author from the countries displayed, or one citation from the subject category displayed.

Some records will contain authors from two or more countries or two or more subject categories, so percentages across all countries or categories will sum to a value greater than 100 and the sum of all the countries or Web of Science categories will exceed the number of citing papers. The analysis also counts each country attributed to a record only once. For example, if a paper had three authors from Canada and one author from New Zealand, it would be counted as one Canadian author and one New Zealand author. Error messages at the bottom of the screen will indicate if any records did not contain data in the country or subject category field, or if the display options chosen did not include all possible records. There is also an option to export the results as a text file.

Procedure for journals

• 1. Even if a journal is not listed in WoS, citations to it from sources covered/processed by WoS can still be retrieved via a CRS, so an evaluation can proceed.

• 2. The procedure is similar to that for evaluating an individual author. Given the long spans of publication of some journals, it may be appropriate to restrict the search to specific years. For example, examining the international profile of successive decades may be useful in detecting if the international profile of a journal is changing (see the example in Calver and Bryant 2008).

Appendix 2: Procedure for profiling international citations or subject area citations of an author in Scopus

In brief, the procedure involves combining the results of a standard search and a secondary documents search for the chosen author, retrieving the citations to the work of the chosen author, and then profiling the countries of affiliation for the authors in the citations. The steps below apply to the current version of Scopus, but interface changes could occur.

Despite the greater journal coverage in Scopus compared to WoS (Jacsó 2005), until recently Scopus did not claim to have accurate data earlier than 1996 (although there is on-going expansion of coverage of the earlier literature to 1970 (Elsevier 2016)). Therefore Scopus may not be suitable for assessing the international or multidisciplinary citations of authors with papers earlier than 1996.

• 1. Use any of the Scopus search options to list and select the publications of the target author. Then click the ‘view cited by’ link to display all the citations to these documents. Select all these citations and transfer them to the ‘My List’ store.

• 2. Return to the original list of the target author’s publications and click the ‘secondary documents search’ to display all citations from documents in Scopus to any of the target author’s publications not in Scopus, broadening the range of citations retrieved. Scan the output carefully to select only those documents actually by the target author (a good knowledge of the relevant CV is essential). Some publications may have duplicate entries because of small variations made by authors when citing, but it is not necessary to aggregate these because the country of the citing authors is all that is needed. Once the documents belonging to the target author are selected, the ‘view cited by’ link can be clicked to display the citing papers. The citing documents can then be selected and transferred to ‘My List’, which will then contain all the citations from the primary search and the secondary documents search.

• 3. Selecting the entries in ‘My List’ and clicking ‘analyze results’ displays an output page where ‘country’ can be selected to display the countries of the citing authors, or ‘subject category’ to display the subject area of the citations. These data can be interpreted in the same way as the output from a WoSCRS. The full data set can be downloaded as a.csv file for opening in spreadsheet software.

Procedure for journals

• 1. Scopus does not have complete data earlier than 1996 [although there is a project to extend the range to1970 (Elsevier 2016)], so the range of years that can be searched is restricted.

• 2. The procedure is similar to that for evaluating an individual author. Instead of entering an author name, the full journal title is entered and the search is set to ‘Source title.

• 3. It is prudent to include a check for secondary documents, which may return papers from the journal cited incorrectly or published earlier than 1996.