Open Science Discussion
In recent weeks, relevant discussions on open science have been taking place. Regarding this topic, the National Council for Scientific and Technological Development (CNPq) published a letter, which was subsequently responded to by the Research Reproducibility Network. Aiming to increase the visibility of this debate, we present below the Network's response, which offers pertinent arguments and important reflections in relation to the letter sent by CNPq. Below is the transcription of the letter.
---
**Open Letter to the National Council for Scientific and Technological Development (CNPq)**
We are writing this letter to express our concerns regarding the text titled *“Open Science: A Dispassionate View”*, published on January 14, 2025, by the President and the Director of Results Analysis and Digital Solutions of CNPq.
We appreciate the agency’s openness to public debate on transparency in Brazilian science and recognize the legitimacy of the challenges mentioned. That said, we disagree with the overall tone of the text. Therefore, we have decided to write this response to continue the discussion with CNPq and the Brazilian scientific community.
---
The reservations expressed in CNPq’s text about open science focus on two main points:
1. The exorbitant article processing charges (APCs) in the current open-access model.
2. The challenges of data sharing, given the lack of infrastructure, resources, and qualified personnel to implement this practice.
We do not disagree with either point. However, the original text treats these obstacles as more significant than they truly are, failing to address simple, low-cost actions that could make the knowledge generated in Brazil more open.
---
**Point 1: Article Processing Charges (APCs)**
The discussion on open access focuses on APCs, which are indeed often prohibitive. However, the price range cited in the text is significantly inflated compared to available estimates. Of the approximately 21,000 open-access journals indexed in the *Directory of Open Access Journals (DOAJ)*, less than 35% charge publication fees, with a median of $820. In *SciELO*, this number was 31% in 2024, with an average of around $300.
More importantly, the text does not address the possibility of solving the access problem through the green route, which involves depositing a version of the article in an open repository. This practice is *permitted by the vast majority of scientific journals*, can be done in *numerous free repositories* – including in Brazil – and, contrary to what the text suggests, does not necessarily include an embargo period. On the contrary, the green route even includes the publication of *preprints* as a *simple and low-cost* strategy to ensure broad accessibility of science.
The article also fails to mention that the current subscription model costs *around 500 million reais per year* for the country, providing access to a limited portion of the population to publishers’ archives for a limited time – a solution that is more exclusionary and less sustainable than any open-access model.
---
**Point 2: Data Sharing**
On this point, the text provides several examples of data whose sharing is challenging due to their volume (such as in physics examples with CERN and LIGO), sensitivity (such as health data from Fiocruz), or strategic value for the country. However, much of the data generated by Brazilian research does not face these barriers, and focusing on complex cases may overshadow viable solutions for simpler cases.
Curation, specialized support, and institutional repositories are desirable but are not necessary conditions for sharing simple data. In particular, we highlight the existence of numerous open data repositories accessible to any researcher free of charge, both in Brazil and internationally.
We also find it puzzling that the text emphasizes the need for data embargoes and claims that data sharing weakens researchers. Since this “weakening” is not explained in the text, our interpretation is that it refers to copyright protection or concerns about data reuse. However, such issues can be resolved through the use of open licenses (such as *Creative Commons* models) that address these concerns.
Moreover, as a public agency, we believe CNPq should advocate for open research for the benefit of society as a whole, rather than promoting data protectionism that serves the personal interests of researchers. Therefore, it is the agency’s responsibility to act precisely to prevent authorship concerns from hindering scientific progress, establishing policies to incentivize and reward data sharing and reuse.
Additional discussions about the original text are addressed in *this document* (and *here*), which includes further clarifications for the Brazilian scientific community on the issues raised in the original text.
---
**Final Considerations and Recommendations**
By listing (and emphasizing) difficulties related to the “implementation of Open Science” – symptomatically capitalized by the authors – the text portrays this process as a major challenge belonging to a distant future and dependent on overcoming numerous obstacles. However, there is no utopian entity called “Open Science” to be implemented. According to *UNESCO’s definition*, open science is a set of actions aimed at enabling open access to scientific knowledge and its creation processes, which can and should be taken step by step.
In this sense, several practices related to open science are within immediate reach of CNPq and the scientific community, many of which come at low or no cost to the agency. Among the immediately implementable actions are:
1. **Commit to evaluating researchers’ output based on the content rather than the publication venue**, as advocated by international initiatives such as *DORA* and *CoARA*. CAPES has already taken *steps in this direction*, and a statement from CNPq in the same vein could eliminate incentives that lead researchers to submit to abusive publication fees.
2. **Create specific fields in the Lattes curriculum** to record open science practices, such as data sharing, code sharing, or preprints. Currently, researchers committed to these practices are forced to include them as “other technical productions,” which does not help highlight or legitimize such contributions.
3. **Require the sharing of scientific articles resulting from agency funding**, using the green route if necessary, as has been done for years by *numerous agencies* such as *FAPESP* and the *NIH*.
4. **Make data sharing the default expectation for projects funded by the agency** – allowing necessary exceptions, such as for sensitive or strategic data and large datasets – in line with what has been advocated by *funders* and *governments* worldwide.
None of these measures entail significant direct costs for CNPq, and although training researchers and support staff to implement open practices may require resources from the community, it is the agency’s responsibility to generate demand for these steps. In the long term, this will strengthen Brazilian research, placing the country at the forefront of an inevitable process in international science, much like the support from funding agencies for SciELO made Brazil a global reference in open access and, more recently, open science.
Finally, we do not agree that advocating for such measures is a passionate or naive view, as suggested by the title of the text. On the contrary, recognizing the need for more open science is a view grounded in evidence showing that the *low level of transparency* in academic research has not ensured reliable and *reproducible* science – which delays the advancement of knowledge and generates *potentially greater* waste of resources than would occur if the problem were addressed.
If CNPq views open science as a meritorious process, as stated in the text, it is the agency’s responsibility to acknowledge its role in the transparency issues faced by contemporary science and take proactive steps to address them. Starting the discussion by listing difficulties does not encourage the academic community to do what is already possible.
It is time to think about the measures that everyone – researchers, institutions, and funding agencies – can implement to make our science more open, one step at a time. The Brazilian Reproducibility Network is available to CNPq to discuss concrete measures that can be taken in this direction.
---
To become a signatory of this open letter, click here.
Source: Brazilian Reproducibility Network (2025). Available at: https://www.reprodutibilidade.org/post/ci%C3%AAncia-aberta-outra-vis%C3%A3o-desapaixonada
