Open access scientific journals: an analysis of the DOAJ catalogue

Introduction

Open access scientific journals are a complex theme, demanding the study of various elements: publishers and publishing countries; whether article processing charges are required or not; the value placed on an article as an information-bearing unit; continuity of publication; and the interoperability of the support systems for the technological infrastructure of publishing (Larivière et al., 2015a, 2015b; Beasley, 2016; Brown, 2016; Pinfield et al., 2016; Peters et al., 2016; Björk, 2017; Ponte et al., 2017; Tella and Kwanya, 2018).

The DOAJ is a global directory of open access scientific journals that are peer reviewed and which are freely available to be read by anyone with an internet connection without an embargo period (Grace, 2013).

By requiring open access immediately available upon publication, the DOAJ emerges as an essential voice in discussions of modalities of access, particularly in what pertains to hybrid journals or articles with embargo periods, establishing the very scope of the definition of open access. An analysis of the set of journals with international criteria that are selected and accepted thus allows us to discuss how science is communicated, according to agreements and rules, and with interoperability and accessibility for the development of science.

The recognition accorded DOAJ journals results from the inclusion of journals in the directory, which serves as a stamp of quality for open access scientific journals. The constant updating of its commitment has consolidated a preference for use of its archives over other options for scientific publishing (Ramírez et al., 2017).

The DOAJ brand is thus linked to a commitment to the immediate availability of scientific journals in open access, so long as these journals meet its indexation standards. This commitment boosts the international reputation of the open access scientific journal and has come to be accepted by a range of databases, such as the Scientific Electronic Library Online (SciELO), Scopus, and Web of Science, since, as Tananbaum et al. (2016) observe, quality as measured by traditional indexes and the reputation of a journal still matter to authors, serving as parameters through which journals are evaluated.

The journals that constitute the directory are classified into two categories according to the number of (basic and certified) indexation criteria that they meet: DOAJ (journals with basic criteria, which, in this study, we call ‘No Seal’) and DOAJ Seal, which include additional parameters (DOAJ certification). Taking DOAJ as representative of the open access titles worldwide, we ask: what are the aspects of scientific journals that constitute open access?

The present research aims to analyse aspects (publishers, country of publication, citation indexes, and article processing charges) and their inter-relations in the set of journals that make up the Directory of Open Access Journals (DOAJ) to understand how the expansion of the directory within the international landscape of scientific journals.

Literature Review

The crisis of journals in the 1980s and 1990s was caused by expensive subscription fees paid to commercial publishers as well as by the emergence of the internet, both of which stoked the open access movement (Guédon, 2001; Suber, 2012; Miguel et al., 2016). The movement’s principles were established in three declarations: the Budapest Open Access Initiative in 2002; and the Bethesda and Berlin declarations in 2003 (Suber, 2009; BOAI, 2012).

According to Couto and Ferreira (2019) and Piwowar et al. (2018), open access is defined as follows:

1. Legal open access:
   • open – users can read and reuse the literature for automated tracking, archiving, or other purposes;
   • free – users are only allowed to read the material;
   • gold – the content is published in an open access journal and all articles are directly accessible on the journal’s website, without an embargo period;
   • green – the content is archived in a repository;
   • platinum/diamond – publication is free of article processing charges fees or any charges to the authors and publications do not undergo an embargo period (Weller, 2014; Sotudeh and Estakhr, 2018);
   • hybrid – content is published in a subscription journal which offers the author an open access option upon payment of an article processing charge;
   • embargo – content is published in a subscription journal, but it is free to read after an embargo period;
   • academic social networks – articles are shared by authors through their academic social networks, such as ResearchGate and Academia.edu;
2. Illegal open access – articles shared in websites such as Sci-Hub and LibGen.

Hybrid journals are linked to publishers that own subscription journals. Through open access, these publishers have increasingly moved from subscription models to open access, but they generally apply article processing charges, which amounts to double-charging. This fact makes them less attractive to funding agencies that require immediate open access publishing (Solomon et al., 2016; Weller, 2014; Okada, 2018; Sotudeh and Estakhr, 2018).

The move from subscription to open access journals points to issues of the relevance of the journal’s content and its reach for readers that do not have access to large university libraries. Publishers sometimes provide an exemption of article processing charges for a period, as a strategy for conquering new markets and establishing new journals (Solomon et al., 2016; Furnival and Silva-Jerez, 2017).

Pinfield et al. (2016) have studied the total costs of open access publications in 23 institutions in the United Kingdom, showing that, from 2006 to 2013, UK publishers published more open access articles with article processing charges fees in comparison to their North American counterparts.

Solomon and Björk (2012) compared open access journals that require article processing charges to those that do not. Their analysis indicates that article processing charges are not associated with the quality of a journal, but are rather a feature of commercial publishers and open access journals published in the United States, the United Kingdom, the Netherlands, and Germany, which are the four leading countries in scientific publishing. Article processing charges cover a significant part of the costs of publishing and their values are higher for journals with high citation indexes (Solomon, 2013; Gadd et al., 2018; Asai, 2020).

Scientific publishers that publish journals can be: for profit (commercial); non-profit; university; scientific societies and research institutions (Crow and Goldstein, 2003; Morris, 2007; Ware and Mabe, 2015; Gadd et al., 2018). Fyfe et al. (2017) argue that institutions that co-publish journals with third parties should consider whether the co-editor’s mission and business strategy are a good fit for the academic mission of the society.

Lowering the costs of a journal through voluntary labour, use of open-source software management, such as the Open Journal Systems (OJS), and with the support of universities and other organizations, is a viable approach to publishing platinum route open access journals, particularly conspicuous in Latin America due to portals such as SciELO and Radlyc. In this model, journals are subsidized by governments, universities, and teaching and research organizations and institutions, making open access a sustainable practice even in the absence of author fees (Miguel et al., 2011; Aguado-López and Becerril-García, 2014; Weller, 2014; Tananbaum et al., 2016).

Started in 2003, DOAJ is a non-profit organization with open access infrastructure services managed by the Community Interest Company, based in the UK, called IS4OA. This directory is furthermore supported by communities composed of individuals, libraries, consortiums, aggregators, and service providers. In 2020, it had a monthly average of 713 applications for indexation (DOAJ, 2016; Mitchell, 2018; Olijhoek et al., 2015; Morrison, 2017; DOAJ, 2020a).

The directory was created to increase the visibility of open access journals and to promote their use, with the stated mission to:

increase the visibility, accessibility, reputation, usage, and impact of quality, peer-reviewed, open access scholarly research journals globally, regardless of discipline, geography or language. DOAJ will work with editors, publishers, and journal owners to help them understand the value of best practice publishing and standards and apply those to their own operations. DOAJ is committed to being 100% independent and maintaining all of its services and metadata as free to use or reuse for everyone. (Mitchell, 2018, pp.1).

In 2014, DOAJ adopted new criteria for indexation, which led to the removal of 3,776 titles from its catalogue. Most were excluded from the directory due to the failure of editors to complete the reapplications forms for indexation, which accounted for the removal of 2,851 journals; a further 243 were excluded because they did not publish articles in the preceding year or published less than 5 articles; 125 were removed because they ceased publication, and 122 were removed because their websites were unavailable (Noorden, 2014; Baker, 2016; Nassi-Calò, 2016; Marchitelli et al., 2017).

The criteria stipulated in 2014 were divided into five layers (Olijhoek et al., 2015; DOAJ, 2019): (1) basic information – title of the journal, International Standard Serial Number (ISSN), publisher, name of the main contact, e-mail address, information concerning article processing charges, information article submission requirements, information on the first issue published in open access, description of the journal, and obligatory language for publication; (2) quality and transparency of the editorial process – editorial board, peer review, journal aims and scope, plagiarism policy, statement of open access, content immediately available; (3) scope of the openness of journal content – all content has to be immediately available upon publication; (4) content licensing – article licensing and authorizations for readers to read, download, copy, distribute, print, research or link to complete articles; and (5) copyright and permissions – information on author rights and permissions.

On top of these criteria, a certificate of good practice, the DOAJ Seal, is granted to those journals that meet seven additional criteria (DOAJ, 2020b): digital preservation; digital object identifier, license type (CC BY, CC BY-SA, CC BY-NC, and BY-NC-SA); upload article metadata to the DOAJ; information on license available in articles; copyright and publication rights; and self-archiving policy.

In August of 2020, 10% of the journals in the directory were certified, i.e., were DOAJ Seal. In 2020, the DOAJ specified criteria for journals in the arts and humanities, clinical case reports, annals for events with attending data, pre-prints, student commissions, and publications that migrated from a subscription/hybrid model to an immediate open access model (DOAJ, 2020c).

The certification criteria for a DOAJ Seal underscore accessibility, openness, access, discovery, reuse, and the author rights of the scientific information made available by the directory (DOAJ, 2019; Xin, 2017). Initiatives such as the DOAJ represent significant gains in the system of scientific communication as a common good and serve the building, ordering, circulation, use, and reuse of the scientific record. This is clear when we analyse what is produced and the conditions under which it is elaborated, that is, what makes a phenomenon such as this possible.

Methodological procedures

The DOAJ was chosen because it is internationally acknowledged to be the main directory for open access scientific journals, and it is accepted by the two main private citation databases used in this investigation: Scopus and the Web of Science.

The research is descriptive, with a quantitative and inferential approach that provides information on the size, composition, and characteristics required for a macro-understanding of a given context (Best and Kahn, 1998). In this specific case, the context is that of global open access publications.

Data were obtained from the DOAJ platform in July 2017. Our initial search returned 9,478 journals, 473 of which were discarded for lacking information on the value of the article processing charges, resulting in a sample of 9,005 journals in the study. These journals were published by 4,156 publishers in 111 countries. In November of 2020, the 9,005 journals included in this study remained indexed in the DOAJ and accounted for 58% of the directory’s catalogue.

Data collection (Title of journal, Seal or No Seal Classification, publisher, country of publisher, and article processing charge value) was registered on spreadsheets using the journal metadata in DOAJ. After data was collected, they were standardized and uncluttered. The conversion of the monetary value of the article processing charge to the US dollar was done manually according to values obtained from the Federal Reserve System.

Journal data from the DOAJ were complemented with information obtained from the Web of Science and Scopus, considering that, as Miguel et al. (2011) point out, open access information systems complement, but do not replace traditional bibliographical databases. The DOAJ does not provide citation data, for which metric data from the Journal Citation Reports and SCImago Journal Rank were used.

Data gathering from the Journal Citation Reports and SCImago Journal Rank was done manually in the Web of Science platform, maintained by Clarivate Analytics, and SCImago. The 9,005 journals were individually consulted in the platforms through their print and electronic International Standard Serial Numbers, as well as through official and alternative journal titles.

The aspects considered (publisher, country of publisher, article processing charges, Journal Citation Reports, and SCImago Journal Rank) were gathered and classified as per the criteria adopted in this study, into the DOAJ No Seal and DOAJ Seal categories. The DOAJ No Seal category does not exist in the directory and refers to those journals in the directory which have not received Seal certification. DOAJ No Seal attends to the 5 basic criteria that all journals in the DOAJ must conform to (basic information, quality, and transparency of the editorial process, scope of coverage, licensing of content/copyright, authorizations, and business model).

Journals certified as Seal meet the 5 basic DOAJ criteria and an additional seven criteria: the openness of content through creative commons licenses, interoperability of journal and article content, author and publication rights, digital preservation policies, self-archiving policies, metadata, and permanent identifiers for articles.

Statistical analysis includes calculations of absolute, relative, and median frequencies, standard deviations, normality tests (so that the adequate test could be used in variable relationships), and medians of central tendency. The confidence interval was 95%. For categorical variables (titles, publishers, countries) we used chi-squared and Mann-Whitney tests (Agresti, 2012).

Results and Discussion

Table 1 shows that more than half of the DOAJ Seal journals require article processing charges (66.52%), while DOAJ No Seal journals have a lower percentage of article processing charges fees (25.8%). There is a significant statistical difference regarding article processing charges (p-value <0.001) for DOAJ Seal journals when compared to DOAJ No Seal, and the chance of a DOAJ Seal journal article processing charges was shown to be 5.72 (4.94; 6.63) times higher than the chances of a DOAJ No Seal journal charging fees. In brief, there is a correlation and greater chance of a DOAJ Seal journal applying article processing charges.

Article processing charges can lead to a new crisis in journal publishing, amplifying and maintaining tightening the grip of commercial publishers situated in central countries in science publishing rankings. Guédon (2017) observes an increase in the value of article processing charges and the number of journals that apply it, a fact which veers from the primary aim of the open access movement against the high costs of subscription journals.

Table 2 shows the number of open access journals by article processing charges, publisher, and according to whether the journal is DOAJ Seal or No Seal. The data shows that Elsevier maintains a total of 323 journals indexed in the DOAJ, 167 of which are No Seal without article processing charges, while 154 are DOAJ No Seal with article processing charges; Springer maintains 176 DOAJ journals, most of which are certified as Seal (142), 78 of which apply article processing charges. It is worth reflecting on the significant number of open access journals maintained by commercial publishers that do not apply article processing charges, which is likely a temporary option until the journal establishes a reputation among researchers in the field or occupies emerging markets.

Table 2 shows the number of open access journals by article processing charges, according to the publisher of DOAJ Seal and No Seal journals.

In Table 2 we find the open access journals maintained by traditional commercial publishers from central countries, such as Elsevier and Springer. Morrison (2017) noted that, in 2016, according to publishers’ websites: Elsevier was the largest open access publisher in the world with 511 journals available in this model; De Gruyter was the second, with 435 journals; Hindawi was the third with 405. The information gathered by Morrison corroborates the results of this study. Even if the publishers are no longer ranked in the same order, they remain the three publishers with the largest number of journals: Hindawi Publishing Corporation with 5.8% of journals; Elsevier with 3.6%; and De Gruyter Open with 3.5% (Table 2).

Table 2 also shows that most publishers are based in advanced economies (central countries) and that publishers from emerging markets and developing economies (peripheral countries), according to the classification of the International Monetary Fund (2018), feature in smaller numbers and with fewer journals.

Hindawi Publishing began its activities in 1997 under a subscription business model, expanding by acquiring other publishing houses. In 2007, intending to make all of its journals open access (eliminating subscription and hybrid journals), it sold four subscription journals to the Oxford University Press and converted the remaining journals into an open access model (Shaw, 2006; Peters, 2007; Loy 2011).

In 2015, Elsevier along had 24% of the scientific journals market share, while Springer and Wiley-Blackwell maintained a 12% share each. These three companies thus accounted for half of the scientific publishing market. Scientific publishing is a profitable market with global revenues of over £19 billion. In 2010 Elsevier registered a profit-margin of 36%, corresponding to £724 million of revenue of £2 billion, out-performing companies such Apple, Google, and Amazon (Larivière et al., 2015a; Noorden, 2015; Buranyi, 2017).

The Public Science Library (PLOS) is a publisher which does not feature in Table 2 since it only has 7 journals in the DOAJ database. It is thus not a publisher with a large number of journals. However, if we analyse the number of articles, it emerges as the publisher with the largest number of articles published with a DOAJ Seal (215,812 articles). PLOS is based in San Francisco, USA, and all of its journals are open access, certified DOAJ Seal, with article processing charges of between US$ 1,495 and US$ 2,900 exemplified by the Web site of the PLOS (https://plos.org/).

Table 2 also shows that the number of journals is not uniformly distributed between publishers, since 52.7% of them publish between 1 and 5 journals. This tendency had already been noted by Ware and Mabe (2015), who observed that 95% of publishers publish one or two journals and that 67% of journals are published by commercial presses. Johnson et al. (2017) observed that 50% of all scientific articles published in 2013 were in journals maintained by Reed-Elsevier, Springer Nature, Wiley-Blackwell, Taylor & Francis, and SAGE. Commercial presses which apply article processing charges for open access also have revenues from subscription, big deals, and double-charging in hybrid journals (Pinfield et al., 2016; Shamashi, 2016).

Open access supported by article processing charges is detailed in Table 3, including the correlation article processing charges between DOAJ journals and publishers.

The study identified 679 DOAJ journals (626 = No Seal and 53 = Seal) from university publishers that did not apply article processing charges and 3,991 DOAJ journals (3,932 = No Seal and 59 = Seal) from publishers with up to five journals without article processing charges.

The publishers that presented the largest average rates for article processing charges are, in descending order: Nature Publishing with a US$ 2,122 average; Wiley, with a US$ 2,078 average; Frontiers Media, with a US$ 1,996 average; Dove Medical Press, with a US$ 1,879 average; and SAGE Publishing, with a US$ 1,514 average (Table 3).

Commercial publishers show the highest averages of article processing charge rates and their editorial groups maintain 17.9% (1,615) of journals that apply article processing charges (Table 3): Hindawi Publishing Corporation, 513 (19.3%); Springer Nature, 399 (15%); Elsevier 156 (5.9%); and Taylor & Francis, 130 (4.9%). Thus, 13 out of the 4,156 publishers included in the study are commercial presses and maintain the greatest number of journals that show a significant statistical relation (p-value) to article processing charges, which attests to the important role of the commercial scientific publishing industry in the open access movement.

Table 3 shows that DOAJ Seal journals on average higher article processing charges than DOAJ No Seal journals for the following publishers: Nature Publishing (US$ 4,000 - DOAJ Seal and US$ 2,122 - DOAJ No Seal); PAGE Press Publications (US$ 596- DOAJ Seal and US$ 425 DOAJ No Seal); university (US$ 364 - DOAJ Seal and US$ 224 - DOAJ No Seal); scientific societies and research institutes (US$ 1,143 - DOAJ Seal and US$ 493 - DOAJ No Seal); and for the varied others category (US$ 1,059 - DOAJ Seal and US$ 520 DOAJ No Seal). Furthermore, the p-value for these publishers is <0.05; that is, there is a significant statistical correlation for the higher rates of article processing charges for DOAJ Seal journals from the publishers under consideration.

Publishers linked to university show lower averages for article processing charge rates (US$ 224 for DOAJ No Seal and US$ 364 for DOAJ Seal – Table 3) or else do not apply article processing charges (53 DOAJ Seal journals and 626 DOAJ No Seal journals); and, when compared to their commercial counterparts, publishers linked to scientific societies and research institutes show lower averages for applying article processing charges and significant statistical relation to higher values in DOAJ Seal journals (average US$ 1,143).

The transformation of scientific communication toward an open access model is consolidated and commercial publishers have adapted to what was originally intended to open up the market to new publishers and to shift away from existing publishing houses.

Now, however, commercial publishers can rely on subscriptions from readers and charge article processing charges from authors in open access articles in so-called hybrid journals i.e., ‘double dipping’. Furthermore, new large publishing houses were created, maintaining exclusively open access titles with article processing charges in comparison to 47% of presses linked to learned societies and 25% of to universities that maintain open access titles that apply article processing charges (Beasley, 2016; Abadal, 2017; Annemark, 2017; Björk, 2017).

From 2006 to 2013, commercial presses profited with gold open access publications. At the end of this interim, 87% of them had options for open access publication with article processing charges in comparison to a rate of 47% of presses linked to learned societies and 25% of universities (Gadd et al., 2018). High article processing charges should hinder the large-scale shift to open access in academic publishing since a system that is based on fees continues to favour authors and organizations that can pay.

Thus, the open access movement, motivated by the ideal of minimizing the problems of social and scientific inequality caused by restricting access to academic research, assumes a new form. It now needs to affirm that ‘economic sustainability is a means to an end, not an end in itself. The end we seek is social justice’ (Beasley, 2016, pp. 1).

An example of the budgetary power of the providers of informational products and services can be seen in the sale of Thomson Reuters to the Onex Corporation and Baring Private Equity Asia in 2016 for US$ 3,55 billion, after which it was renamed Clarivate Analytics in 2017. Its business remains focused on science, such as the analysis of patents and regulatory norms; protection of registered brands; biopharma intelligence; and protection of domain brands; among other services and products (Clarivate Analytics, 2016).

Another change in scientific informational services was the change in Elsevier, belonging to the RELX Group, from a provider of scientific content to an information analytics business with the legal remit to host informational tools such as Altmetrics, Mendeley, and Scopus, as well as continuing to function as a publisher. This business expansion has ‘increased the dependency of researchers and institutions’ and generated unequal growth in global knowledge (Posada and Chen, 2017, pp. 2).

Elsevier’s influence in the cycle of research is a challenge for researchers and readers. Institutions often demand that their researchers publish in journals maintained by commercial publishers, which hampers the growth of independent information services of the information industry. As a reaction to these market pressures, the last years have witnessed the emergence of various resistance movements spearheaded by academics, to boycott and challenge the oligopoly of the large publishers (Posada and Chen, 2017).

In this landscape, different paths coexist with the open access model. This study has considered gold open access, characterised by article processing charges, and platinum/diamond open access, which does not demand payment. Journals with platinum/diamond models, free of charge for both readers and authors, represent 70% (6,319) of journals in the present study. Scholars such as Haschak (2007), Crawford (2011), Björk (2017), and Allahar (2017) claim that platinum/diamond journals are a relatively unexplored alternative, but an important possibility in open access publishing.

Academic journals are irreplaceable products for authors and researchers, and platinum/diamond open access journals help reduce monopolies within a discipline or field, widening readership and publication possibilities. Career paths that reward publication in established journals with a high Impact Factor (IF) strengthens the dominant position of large publishers, which maintain journals with the highest IF (Johnson et al., 2017; Gasparyan et al., 2018; Okada, 2018).

Impact Factor defines the concentration of scientific capital through the reputation of researchers as measured by citations, in a way that displays the Mathew Effect since these researchers have advantages such as funding, in contradistinction to those authors who do not meet the indicators recognized in a given field (Bourdieu, 1976; Cintra et al., 2017), ‘in such a way that the hyperbolic distributions that characterize bibliometric laws are the expression of a process of accumulated advantages’ in the market of commercial scientific publishing (Silva et al., 2011, pp. 116).

The averages displayed in Table 4 show the open access journals that are present in indexes such as the Journal Citation Reports and the SCImago Journal Rank.

Table 4 is organized in descending order of the concentration of journals in open access indexed in the Web of Science database, which provides the Journal Citation Reports index, and the Scopus database, which provides the SCImago Journal Rank index. The Journal Citation Reports averages of publishers are higher than the SCImago Journal Rank average, a result of the larger number of journals in the SJR index. Furthermore, commercial publishers (BioMed Central Journal Citation Reports=2.779 and SCImago Journal Rank=1.173 from Springer Nature; Elsevier Journal Citation Reports=2.912 and SCImago Journal Rank=0.688; Wiley Journal Citation Reports=4.102 and SCImago Journal Rank=1.733; Nature Publishing Group Journal Citation Reports=6.366 from Springer Nature) have the highest Journal Citation Reports and SCImago Journal Rank values, which allows us to infer a relation between indexation and article processing charge values.

Journals that are indexed with citation indexes gain scientific reputation, but:

Whether the communication system and the reputational system of science should be the same is the fundamental question that must be brought to light, debated, and ultimately replaced by a different structure. Science needs two independent layers. In the first one, the optimal dissemination of scientific knowledge can be allowed to take place freely. Call it the "net neutrality" of the Internet of the mind. In the second layer, the process of evaluation can proceed, as it should, i.e., on the values and objectives of the research communities themselves, not on the manipulated metrics favoured by publishers. (Guédon, 2017, pp. 38).

Thus, the value of metrics and the value of article processing charge rates determined by commercial presses maintain in place ‘the economic forces pressure for norms increasingly geared toward ensuring the monopoly of dominant groups’ (Santos and Rebouças, 2016, pp. 102).

The differential article processing charges rates that some publishers, such as PLOS and Springer Nature, offer some countries and institutions are insufficient to widen access, since the US dollar values remain inaccessible to many. In contrast, platinum/diamond open access makes no economic distinction, since it does not depend on the lists of benefits provided by publishers that determine which countries and institutions qualify for exemption or reduction of publication fees, hence making them all equal.

The results of this research show that commercial publishers emerged from a near-invisibility in the open access field to become behemoths of the model in the editorial market. The growing presence of commercial publishers in open access, supported by the high article processing charges values and consorted subscription deals in various countries (Björk, 2016; Morrison, 2017) reveals that:

the whole debate between publishers, libraries and funding agencies had been re-organized as if its most important point was no longer the communication system of science, but rather the well-being of the publishers. The message conveyed by William Garvey, Communication: the essence of science had been surreptitiously transformed into Commercial publishing: the essence of science.” (Guédon, 2017, pp. 21).

Hence, ‘to control access to scientific literature is, for all effects and purposes, to control science’ (Buranyi, 2017, pp. 2). To map which countries have more journals indexed in the DOAJ, Table 5 was elaborated according to the classification if journals and whether they apply article processing charges fees or not.

Table 5 shows that the 8 countries with most journals in the DOAJ are: Brazil (10.9%), the UK (9.7%), Indonesia (7.7%), Egypt (6.1%), Spain (5.9%), the USA (5.5%), Poland (4.8%), and Italy (3%). These countries publish over half (53.5%) of the journals featured in this study.

It is noteworthy that Brazil is the country with the highest number of open access journals in the DOAJ that do not apply article processing charges fees. In consonance with the findings of the Science-Metric report National Science Foundation's and National Center for Science and Engineering Statistics (2018), Brazil is the country with the highest number of open access journals. Data from the present study complement this information by showing that Brazil is the country with the highest number of open access journals in the DOAJ (978 = 10.9%). Furthermore, we registered an increase in the collection of Brazilian journals in the DOAJ: according to Kieńć (2017) this was at 9.6% in 2016, and at 10.9% in 2017 (Table 5).

Table 5 also shows that Europe is the continent with the largest concentration of DOAJ journals (47.65%), particularly when we consider only those certified Seal (84.39%). The UK, with its long publishing tradition and a high number of commercial publishers, also stands out among countries publishing open access journals. Asia is represented by 18.08% of the journals in the DOAJ and South America by 17.13%. South America is the continent with the highest number of No Seal journals (18.92%).

The emphasis of Brazil in terms of the number of open access journals can be explained by the development and consolidation of open access throughout Latin America, due to the low number of commercial presses, which made possible the establishment of a significant number of free journals edited by universities, scientific societies and research institutes (Rodrigues and Abadal, 2014; Costa and Leite, 2016; Shen, 2017). Furthermore, SciELO began to operate in Brazil in 1997, providing an incentive for Brazilian journals to follow on the path of open access even before the Budapest, Bethesda, and Berlin declarations.

The percentage of journals from Europe with DOAJ Seal in this investigation (84.39% - Table 5) allows us to infer that there was an accommodation of the criteria for DOAJ Seal certification to the demands of European science.

The Science Europe Working Group on Open Access, in 2013, published the principles for the transition to green and gold open access research, which are being adopted by member countries since 2015. The principles condemn publication in hybrid journals, demanding transparency regarding the sums charged by publishers for immediate open access options, and the application of reduced sums depending on paying institutions, regions, or countries. Publications should be high-quality, featuring (Science Europe Working Group on Open Access, 2015):

1. indexation – journals should be indexed in PubMed Central, DOAJ, Web of Science or Scopus, and books and other academic publications must have peer-review procedures and be directly available on the website of the publishing;
2. copyright and (re)use rights – authors should retain the copyright of their publications without restrictions, and these should have open licenses, preferably CC BY or another license which conforms to the requisites of the Berlin Declaration;
3. sustainable archiving – editors should enable articles to be deposited in repositories or third-party websites immediately after publication through a permanent link through which all content can be accessed, read, and downloaded;
4. machine readability – the complete text of the publication, metadata, supporting data, citations, and the status of the open access publication should be made available in standardized, machine-readable form, and editors must notify of any alterations to authors.

The quality principles listed above are similar to those of the DOAJ Seal, which came to be part of the DOAJ criteria in 2014 (DOAJ, 2019).

Members of the Council of Europe that adhere to open access include the United Kingdom and Germany, which are undergoing a transition from open access to article processing charges fees. This transition was influenced by national open access policies, with specific funding for this type of publication and support from universities and research institutes. The United Kingdom and Germany are undergoing a transition to open access with article processing charges fees, influenced by national open access policies, with a specific budget for this type of publication, supported by universities and research institutes (Jahn and Tullney, 2016, Schimmer, 2017; Gadd et al., 2018).

Authors from peripheral countries continue on the classic cycle of science, publishing the results of their research abroad; that is, in European and North American journals, since publications in peripheral countries remain marginal to the core of the science world, not being recognized as relevant by assessing entities (Gibbs, 1995; Chavarro et al., 2017).

Table 6 shows relations between countries and article processing charges fees.

Among countries present in the two classifications of DOAJ journals, we find a significant statistical margin (p-value <0.05) for dollar values with higher value article processing charges fees for No Seal DOAJ journals from the UK, Switzerland, and the Netherlands, which have a greater average of Seal certified journals. These countries host some of the largest commercial publishers (Springer, MDPI AG, Elsevier, Hindawi) which turn out the greatest profits in the scientific publishing industry. We also find a significant dollar value statistical difference (p-value <0.05) in DOAJ journals from Canada and Italy, which charge higher fees for DOAJ Seal journals.

Germany (Table 6) had an average article processing charges value of US$ 1,185 for No Seal DOAJ journals, but the UK has the highest value for DOAJ No Seal journals, averaging an article processing charges of US$ 1,474. With an article processing charges of US$ 1,860, Canada had the highest average of DOAJ Seal journals. From 2005 to 2015, Germany spent € 9,627,537 in open access publishing fees for 7,417 articles published in this model, with an article processing charges average of € 1,298 (Jahn and Tullney, 2016); the UK had article processing charges costs at US$ 18.6 million per year, paying for 140,000 open access articles from 29 Institutions in the country for 2013/2014 (Johnson et al., 2016).

Open access with article processing charges fees is linked to capital made up of directly financed resources, making a profit for commercial publishers. In contrast, platinum/diamond models are characterized by the exchange of symbolic goods without financial gain. Platinum/diamond open access journals give authors greater opportunity since they are free of article processing charges fees, thus benefitting both authors from central and peripheral countries.

When compared to the United Nations Educational, Scientific and Cultural Organization, (Unesco, 2015) science report , the countries included in the study (Table 6) are equivalent to some G20 countries, representing 66% of the world population: European Union, Germany, Argentina, Australia, Brazil, United States of America, France, Indonesia, Italy, Japan, United Kingdom, Russia, and South Africa (G20, 2018).

The fact that countries from Europe and the United States (Table 6) have a large number of DOAJ journals with article processing charges fees is evidence of an existing oligopolist economic model for internationalization, wherein Europe and the United States are responsible for the transmission of scientific knowledge for other parts of the globe.

Other regions, with the particular histories of their peoples and unique cultural, political, economic, and social styles, have started to distinguish themselves scientifically from central countries. Despite these distinctions, these regions still need to publish in journals controlled by central countries (Silva, 2014).

In the model of open access journals, the political, social, and economic relations between central and peripheral countries remain one of dependency. It thus coexists alongside the power hierarchies within which the epistemic colonization of systems of scientific communication by commercial publishers persists, sustained by initiatives that stem from the central countries which remain the headquarters of the large commercial publishers that have adopted an increasing number of open access titles, possibly as a way to maintain their profits and market presence.

Most open access journals from peripheral countries adopt platinum/diamond models, diminishing the financial barriers that affect authors in the (Nkuyubwatsi, 2017, 2018; Weller, 2014) ‘free and open dissemination of knowledge’ (Appel et al., 2018, pp. 11). In this model, ‘the costs of editing, peer review, and publication or hosting are covered by an academic institution or fund’ which makes its infrastructure available to this end, and academics do the work of editing and revising in all capacities – as authors, editors or copy editors (Fecher and Wagner, 2016, pp. 6). The institutions responsible for the journal determine the type of access and the value of article processing charges fees.

Analysing the scientific space of journals allows us to understand the scientific field competitively and, simultaneously, as a scene for social integration, with publishers and countries that apply article processing charges fees and those that do not.

The competitive space is understood as being ‘structured around specific challenges and interests’ (Ragouet, 2017, pp. 68) that unfold with no rupture of ties with commercial publishers and central countries. There is thus a critique of hierarchical relations and social and academic institutions that, modeled on the cultural field, keep their (Nascimento and Bufrem, 2017) ‘perspectives on the status quo, occasioning changes that are internal and external to the…’ scientific field (Passiani and Arruda, 2017, pp. 73).

The relations of science revealed in this study through data on countries and the publishers responsible for journals registered in the DOAJ,

like the economic world, knows relations of power, phenomena that concentrate capital and power or even monopoly [or oligopoly], social relations of domination that imply an appropriation of the means of production and reproduction … [and control over them] (Bourdieu, 1997, pp. 34).

Symbolic power is the basis of the production and reproduction of the means for perpetuating a structured and invisible hegemony, and it is constituted from the construction of the social world and it is

exercised with the complicity of those who do not wish to know that they are subject to it or even that they exercise it … [by the] social function that reproduces the order of relations with the interests of the dominant class presented as common interests (Bourdieu, 1989, pp 7-10).

Thus, the culture that unites is the same that renders asunder and makes legitimate the distinctions of the field of symbolic production, feeding the domination of economic capital by the production of the dominated party.

In this way, it is noteworthy that there is a small number of open access journals from smaller publishers and peripheral countries, a fact which legitimizes the dominant power of the large publishers and central countries which tend to apply article processing charges fees. This allows us to claim that open access has ceased to be an alternative phenomenon struggling against the high costs of journal subscriptions for libraries, particularly university libraries, to become a commercial alternative for the market as a whole. Publishers have refashioned themselves according to the pre-requisite of providing accessible information at no cost to their readers, making it into a sustainable business by, for example, increasing article processing charges values. But

we do not need more central journals or journal rankings, we need to think about how to recognize ourselves, to identify ourselves, without necessarily passing through systems of prestige and visibility, particularly when these are constructed with the double aim of facilitating the production of knowledge and intensifying the generation of profits (Guédon and Loute, 2017, pp. 14).

Open access benefits academics, populations, government and non-government organizations, and society at large. The right to information as a common good is included in Article 27 of the United Nations Declaration of Human Rights. Furthermore, public funding of research through taxation reveals that science and technology are social demands (Tennant et al., 2016; Cupani, 2018). Social demands are realities that contribute to maintaining the ecosystem of scientific communication as a means for disseminating science’s answers to the problems of society.

Conclusions

On analysing the DOAJ, we found that, of the 4,156 publishers included in the study, 42.3% are universities; 12.9% scientific societies and research institutes; 41.8% are classified as others; and only 3% are commercial publishers. These data indicate that there is a significant number of small publishers with up to 5 journals and a few commercial publishers with a large concentration of open access journals, the main ones being: Hindawi Publishing Corporation; Elsevier; De Gruyter Open; BioMed Central; and Springer; and publishers that maintain mega-journals, such as PLOS, with a large number of published articles.

The study has also shown that the context of open access journals in the DOAJ is made up of 45% of so-called peripheral countries, such as Brazil and Argentina; 35% of central countries, such as the United States and Germany; and 20% of semi-peripheral countries. Of the 978 Brazilian journals, 92% do not apply article processing charges fees (most journals follow the platinum model). Brazil (972), Indonesia (692), Egypt (546), Spain (530), and the UK (489) are the countries with the largest number of No Seal journals, while the countries with the largest number of DOAJ Seal journals are the UK (384), Germany (127), Switzerland (121), and the United States (54). These are the countries that host the large commercial oligopolies, now publishing in open access and applying article processing charges fees.

Commercial publishers maintain the largest number of journals that apply article processing charges and the highest rates (Nature Publishing – average US$ 2,122; Wiley – average US$ 2,078; SAGE Publishing – average US$ 1,514). By analysing interrelations, we conclude that universities stand out for not applying article processing charges fees. When they do, they charge lower values (US$ 244 for DOAJ No Seal journals; US$ 364 for DOAJ Seal journals; and 679 journals with no charges).

Open access with no article processing charges fees amounts to 70% of journals. The relationship between article processing charges fees and Seal certification is more evident in Europe (84.39%), which suggests that countries with a developed economy and commercial publishers adopt article processing charges fees as a means to sustain open access publishing.

Latin American countries are characterized by lower or no article processing charges fees since most of their journals are from universities. Commercial publishers are those that have the largest average in the citations indexes, and more Seal certifications, even though the DOAJ requirements for Seal certification are neither particularly difficult nor expensive to put into practice.

Journals that have Seal certification feature more frequently in commercial citation indexes and have more international prestige because of the (Journal Citation Reports) impact factor of the Web of Science and the SCImago Journal Rank of Scopus, as we can see from their averages: BioMed Central – Journal Citation Reports (2.779) and SCImago Journal Rank (1.17); Springer – Journal Citation Reports (2.797) and SCImago Journal Rank (0.66); and Wiley – Journal Citation Reports (4.102) and SCImago Journal Rank (1.73) are larger than the averages for university (Journal Citation Reports = 0.576 and SCImago Journal Rank = 0.25) and scientific societies and research institutes (Journal Citation Reports = 1.226 and SCImago Journal Rank = 0.51).

University maintains few journals (one to five) per press, which atomizes their efforts. Commercial publishers gain strength within open access through the creation of new journals, by acquiring journals from small presses, and by increasing the scale of technological structures and teams.

Beyond the dichotomy of platinum open access without article processing charges fees and gold open access with high article processing charges rates, the DOAJ criteria do not prevent or inhibit the oligopoly of large commercial publishers that shift high costs to authors, with the justification of allowing free competition between publishers. Thus the movement follows the directives and recommendations contained in declarations such as those of Budapest, Bethesda, and Berlin, which each journal or institution can take on board or not, but the oligopoly of scientific commercial publishers remains in the open access model of scientific journals.

The decision to adopt the directives and recommendations results from the economic, social, and cultural policies of the editorial committee and the aims of the journal which are supported by government and institutional policies.

An example of a government and institutional open access policy is the Horizon 2020 of the European Commission, which establishes that all research funded by the European Research Council should be made immediately open access (not be published in hybrid journals or those with embargo periods), as reported by Enserink (2018). We see government interest in minimizing barriers to access and the costs of scientific publication which must be considered as part of scientific research. However, this commitment needs to be institutional, professional, and personal, as a guiding thread in the aims of all stakeholders in the scientific ecosystem.

The data show that open access has made possible the emergence of new journals and that the oligopoly has not been altered. The publishing industry continues to maintain the largest number of journals, through commercial publishers such as Elsevier, Springer, and Taylor & Francis, in central countries and providing information services such as the Web of Science and Scopus which make available the Journal Citation Reports and SCImago Journal Rank citation indexes, highly valued by the systems of auditing researchers and institutions. The system of scientific communication thus feeds back on itself, with informational services and products provided by the large, traditional publishers.

This study has analysed the characteristics of scientific journals in the DOAJ globally, allowing us to identify models of publishers and widening the diversity of participating countries. Open access journals are increasing access through many ways while maintaining the hegemony and oligopoly of the commercial publishers which continue to maintain the largest number of titles.

Acknowledgements

The authors thank the reviewers for their helpful comments on the original paper.

About the authors

Suênia Oliveira Mendes is a librarian/archivist at the Universidade Federal do Maranhão, Brazil. Suênia received her Ph.D. in information science at the Universidade Federal de Santa Catarina and holds several professional certifications in scientific communication and information management. Her research focuses on scientific journals, open access, open science, scientific production, innovation, and sources of information. She can be contacted at: suenia.mendes@ufma.br.
Rosângela Schwarz Rodrigues is a professor at the Department of Information Science at the Universidade Federal de Santa Catarina, Brazil. She has MPhil and DPhil studies in industrial engineering at the Universidade Federal de Santa Catarina. Her research topics include open access, scientific journals, and scientific output. She can be contacted at: rosangela.rodrigues@ufsc.br.

How to cite this paper