The Brazilian Reproducibility Initiative was a research project that sought to estimate the reproducibility of Brazilian biomedical science.
​Learn about the project
Publications
How were the experiments selected?
​A systematic review initially identified the laboratory methods most frequently used by researchers in Brazil, and from this list we searched for laboratories that could perform them in our project. The most common methods identified were the MTT assay, which assesses cell viability, PCR, which assesses gene expression, and the elevated plus maze, which assesses behavior in rats and mice. A sample of the literature was then randomly selected, and after going through inclusion criteria, we arrived at 20 experiments with each of the techniques being selected.
How were the replications performed?
How do you know if a result has been successfully replicated?
​Each experiment was assigned to 3 different labs to replicate, and we were able to complete 143 replicates of 56 experiments. Of these, 3 experiments were performed by only 1 lab, 19 were performed by 2 labs, and 34 were able to be completed by all 3 labs as planned. Each replicate was designed using the methods of the original paper, but each lab could adapt the protocol independently of the other.
​Replication success can be understood in many different ways. We used 5 criteria: original effect within the 95% prediction interval of a meta-analysis of replications, effect of a meta-analysis of replications within the 95% confidence interval of the original result, result of a meta-analysis of replications being significant and in the same direction as the original, half of the individual replications having results that are significant and in the same direction as the original, and finally, a subjective assessment made by the replicating laboratories themselves.
Why focus on Brazilian articles?
​Because knowing how reproducible the research produced in the country is is essential for improving our science. Since this type of survey is not usually carried out by funding agencies, we believe that the scientific community itself should face this task as a collective challenge.
We conducted the first systematic study of scientific reproducibility at a national level that we have ever seen in the world. This data allows us to reflect on our individual and collective practices, helping to place Brazil at the forefront of the development of more reliable science worldwide.
​Learn more about the project through our publications:
Abrindo a caixa preta (2018)
http://cienciahoje.org.br/artigo/abrindo-a-caixa-preta/
Science Forum: The Brazilian Reproducibility Initiative (2019)
https://doi.org/10.7554/eLife.41602
Two years into the Brazilian Reproducibility Initiative: reflections on conducting a large-scale replication of Brazilian biomedical science (2020)
https://doi.org/10.1590/0074-02760200328
Reproducibility: expect less of the scientific paper (2021)
https://doi.org/10.1038/d41586-021-02486-7
Estimating the replicability of Brazilian biomedical science (2025)
https://doi.org/10.1101/2025.04.02.645026
Science Forum: Addressing selective reporting of experiments through predefined exclusion criteria (2020)
https://doi.org/10.7554/eLife.56626
Are Most Published Research Findings False In A Continuous Universe? (2022)
https://doi.org/10.1371/journal.pone.0277935