2.3.2 Searching for relevant articles and assessment of the relevance and quality of the articles found

Searches were performed in six research databases: EBSCO (Business Source Complete, Psychology and Behavioral Sciences Collection, PsycINFO and GreenFILE), Ex Libris PRIMO, JSTOR, Scopus, Web of Science (Web of Science Core Collection) and ProQuest. Search strings were used to identify relevant articles with a combination of the following keywords: (female OR women OR woman OR gender) AND (advanc* OR promot* OR causes OR reasons) AND (board of directors OR compan* OR corporate board). For some search queries additional keywords were added, which were: (achiev* AND diversity) OR underrepresentation OR participation. The keywords were used for a search in the title and/or abstract of articles, except for the Ex Libris PRIMO database, where a search was performed in the title or elsewhere in the text.

The inclusion criteria that were used to select relevant articles are:

The inclusion criteria under (1) and (2) were applied to ensure the selection of high-quality articles which can also be understood by the authors. The quality of the articles was assessed by the authors. If the authors regarded an article of bad or mediocre quality the journal’s Scopus cite score 2017 and ranking in the Academic Journal Guide 2018 (AJG)1 were also considered as an indication of quality. The large majority (91.94 percent) of the collected articles are published in impactful journals included in the ABS list and almost half of the articles (48.80 percent) are published in journals with an ABS score above or equal to 3. Almost three-quarters of the collected articles (72.55 percent) have a cite score above or equal to 2. Figure 2.1 shows the number of articles published per year after 1999 (criterion (4)). Criterion (5) is included to select studies that cover public and private limited liability companies because across jurisdictions these corporate forms are mostly structured in similar ways. This assists in generating consistent results.

The keywords search and the application of inclusion criteria produced a total of 44 relevant articles that were selected for this review. Additional articles were identified by searching in the reference lists of the 44 selected articles using a snowball technique. This resulted in 7 additional articles that were selected as relevant for the review and brought the total number of selected articles to 51. As the scope of this review is broad and database search algorithms may be flawed, the authors do not claim that all relevant articles were identified.