How Collective Action Shaped Open Access in Latin America

In 1997, Abel Packer helped found SciELO, the publishing platform and aggregator of open access journals, books, preprints, and data. Today, SciELO holds the world’s largest collection of Latin American scholarly articles alongside content from South Africa, Mexico, Spain and Portugal. Last month, Packer joined Michael Upshall on the Against the Grain podcast to discuss SciELO: its history, its present-day operations, and visions for its future.

Their conversation has been edited for length and clarity.

I understand you started out as a librarian. You have a master’s in library science.

I originally studied engineering and computer programming and worked for many years with statistics data—processing demographic data, surveys, and censuses. About 50 years ago, I started working to develop bibliographic databases. That experience ultimately led me to do a master’s degree in library science, in 1996.

How did you get from bibliographic databases to SciELO, a full-text collection?

The first extensive bibliographic database dissemination solution, which we developed in Brazil, at BIREME (The Latin American and Caribbean Center on Health Sciences Information), with the support of the Pan American and World Health Organizations (PAHO/WHO), used CD-ROM. We started with a health science database called Latin America and Caribbean Health Sciences Literature Database (LILACS), in 1988. To my knowledge, the LILACS/CD-ROM was the first bibliographic CD-ROM title in the developing world. And then we produced and distributed the National Library of Medicine MEDLINE database.

But that was not enough. People kept asking for the full text. So, in 1997, we finally started a pilot project to develop a methodology and technology to run full-text databases on the early Web.

That was SciELO—Scientific Electronic Library Online—which has been continuously updated for the last 28 years.

We started the project with ten journals, the ten best journals edited in Brazil at that time from different disciplines. We created a focal group with the editors-in-chief of those journals to help frame and shape the SciELO publication model. They assisted in the development of a model centered on research communication. We decided from the very beginning to have the full text marked up, so the metadata and interoperability could be extracted from the full text and not be copied. A year later, the National Commission for Scientific and Technological Research of Chile (CONICYT) adopted the SciELO model and installed the system in Santiago, Chile. Thus began the fully decentralized SciELO network.

Within seven or eight years, we had 15 countries in the network. Now we have 18 countries, including South Africa, Portugal, and Spain, running a decentralized model adapted to the context of each country. From the very beginning, SciELO has operated as a networked platform for leading journals from research communities across various disciplines, governed by national research organizations and supported by national public policies.

Did you have any expectation when you started with those ten journals that SciELO would grow to the size it is today?

The field of scientific communication is in some ways tremendously conservative. When you introduce any innovation, you need to immediately show that it has advantages, that you will gain in prestige, visibility and impact. But we were lucky because at that time Anurag Acharya was launching Google Scholar. He worked with us to index SciELO perfectly, resulting in an incredible number of accesses. This visibility helped us to convince the editors of journals to adopt SciELO’s online publishing platform, which offered significant advantages compared to the limitations of print publishing. At the time, the print editions typically produced only 300 to 400 copies per issue, with high costs of production and distribution by mail, while reaching a very small fraction of users of the web version.

So, from the beginning, SciELO was successful in terms of efficiency and visibility. We established, and still maintain, very strict quality control. If you go to OpenAlex, a comprehensive database, there are around 7,000 journals from Latin America. SciELO indexes only about 25 percent of them, because we have indexing criteria that require peer review, metadata in English, and the full text marked up with JATS/XML.

What you’re describing is not dissimilar to the kind of quality checks that Directory of Open Access Journals (DOAJ) performs.

We follow the same global criteria as DOAJ, Web of Science (WoS), Scopus, and so on. But in many ways, SciELO is stricter than Scopus and Web of Science, because we have scientific committees that are familiar with national research and review what the journals communicate, their representativeness, and so on.

Why is it that SciELO’s qualifications for admission are more restrictive but Web of Science and Scopus index far fewer Latin American journals than SciELO does?

It depends on the field. For example, if you take the agricultural journals from SciELO, most of them are also indexed in Scopus and Web of Science. But if you take social science and humanities journals, they are not, because those indexes are not inclusive of languages other than English, nor research on locally centered topics.

SciELO hosts preprints and data. Does it host other content, like books?

Yes, SciELO manages several libraries of research communication objects. The first is the SciELO Preprints Server, which operates the collection of moderated manuscripts not yet evaluated by journals. The SciELO Books network includes indexed books and book chapters from academic presses. The SciELO Data repository collects research data associated with preprints, articles, and book chapters. In addition, journals are beginning to publish articles’ peer review approval reports.

SciELO is, by far, the largest collection of open access Latin American content, which is remarkable because there’s no similar, non-commercial equivalent in the US or Europe. Why do you think it became so successful?

SciELO’s decentralized model requires only one national collection per country, which needs to be implemented through a public policy that funds their journal collection following a common methodology. In each country, a scientific committee selects the journals. Since 2020, we have considered SciELO—including the SciELO Preprints Server, the SciELO Data repository, and the academic SciELO Books collections—to be a complete open science platform. Promoting open science improves quality and confidence in research.

In many countries there are institutional repositories through which individual institutions collect and provide access to that content. Is that their closest equivalent to SciELO?

Yes, definitely. But some of that published research might not meet our quality standard. For example, in Brazil, SciELO currently indexes 328 journals from 180 institutions, which is an impressive national network. But many of these institutions publish other journals that are still trying to meet the requirements to be indexed and published by SciELO. Meanwhile, many of these journals that do not comply with SciELO criteria are indexed by Scopus or WoS, which means they tend to be well evaluated by national research evaluation systems, regardless of the improvement of the research in the field.

The solution is to follow the San Francisco Declaration on Research Assessment (DORA) and other declarations that say not to evaluate research based on the journal in which it appears and instead to evaluate the impact of the article.

Who owns SciELO?

That is a good question that requires a long answer.

SciELO is an open science research communication program guided by five principles and a common publication model. The first principle is scientific knowledge as a public good. The second is to work by networking at all levels. The third is to work with the state of the art in research ethics and good practice and to promote innovation. These three basic principles drove our work for 25 years. In 2023, we adopted two additional principles: The DEIA (diversity, equity, inclusion and accessibility) principle, and the FAIR (findable, accessible, interoperable and reusable) principle.

Our publication model encompasses the policies, methodologies, and technologies necessary to operate collections—web libraries—of research communication objects such as preprints, articles, academic books and book chapters, research data, and peer review reports.

SciELO is also a decentralized, networked organization with a public policy objective: to develop national capacity and infrastructure on research communication. Within that general objective, SciELO aims to maximize the visibility and the impact of research communicated in nationally edited journals. Every five years, network representatives meet to evaluate performance, identify challenges, and update our five-year plan, which is focused on professionalization, internationalization, and sustainability.

So, there is no single owner. There is no signed agreement between countries, no transference of money, and, in most of the countries, there is no signed agreement with the selected journals. Everything is voluntary, an incredible collective construction.

How is SciELO funded? Does each nation contribute based on the proportion of its content?

It’s funded by the government of each participating nation, but there is no transfer of money between the countries. The contributions between countries are the research articles and other content that goes into the SciELO common space.

SciELO can best be conceived as a metapublisher space, a vast virtual journal or platform, with each individual journal communicating research in the common space. Fully decentralized—that’s important to repeat, because that makes it sustainable, equitable, and capable of addressing and overcoming asymmetries.

What’s next for SciELO’s open science model—a publish-review-curate process of the type that’s just coming into use in Europe and the US?

In my view, SciELO Open Science should move in the same direction as eLife—that is, toward integrated preprints and open peer review. We would also love to include F1000 in SciELO, but it’s currently private.

We are working to convince journals to start their publishing workflow with preprints. To this end, we are in the process of reserving a space for each SciELO journal in the Preprints Server, which allows an author to create a preprint upon submission of their manuscript. We conduct a preliminary desk review of the format as well as several editorial and quality control checks with AI support. The main criterion is the feasibility of the manuscript being accepted for peer review. The Preprints Server communicates to the journal editors that a manuscript was received. The editors decide the next steps. If rejected, the manuscript goes to a common space in the Preprints Server, allowing the author to submit it to other journals, or for other journals to discover and express interest in that piece.

The SciELO journal publishing process currently takes a median of one year from submission to publication. Using preprints is a way for us to speed up the communication of research, while still employing the quality control necessary to guarantee that what is being loaded is a genuine research article.

I read that although article processing charges (APCs) are used in Latin America, the average APC is in the region of $300, a tiny proportion of the typical cost in the US or Europe. How is this achieved?

SciELO journals have institutional support and do not need to pay editors, who are professors from these institutions. The platform is free for journals to use. It is developed and maintained with national government funding. Articles in SciELO cost less than $400 while still including all the high-standard processes, as well as being maintained and preserved.

SciELO’s low publishing costs contribute to an overall reduction of research communication costs. Because governments in the region are starting to establish transformative agreements with commercial publishers, high APCs are becoming increasingly visible. These agreements publicly reveal that scholars are paying over $2,000 per article. If SciELO can produce state-of-the-art article editing for only $300 to $400, why are scholars paying so much more to commercial publishers?

If SciELO were to publish more articles, our countries could spend less with commercial publishers. So SciELO is also a potential component of our countries’ strategies to decrease the financial burden of transformative agreements and to participate proactively in the global research communication flow.

What about the future of SciELO? Are you confident that SciELO will still be around in 10, 15 years?

Yes, I am confident, because it’s the very nature of SciELO to value research that is not addressed or prioritized by commercial publishers.

What we see happening on the Web is a tendency toward consolidation, which is predicted by network theory. In the future, we expect to see further consolidation toward a smaller number of very large publishers. SciELO has already proven it can survive within this concentrated model.

Our strategy will continue to be based on communicating high-quality research oriented toward solving our countries’ critical problems, something that will never be the priority for international commercial publishers. For the specific, applied part of research to be truly global, it needs to be published with high quality and high visibility. That is the irreplaceable role of SciELO.