A 2025 preprint from the Brazilian Reproducibility Initiative reported replication success rates between 20% and 44% across 56 Brazilian biomedical labs.1
But the most revealing finding was not the replication rate itself. It was how often reproducibility broke down through day-to-day operational issues, including unclear protocols, inconsistent documentation, and missing process controls.
In the webinar The Reproducibility Revolution: Making Science Transparent, Dr Olavo Amaral, founder of the Brazilian Reproducibility Network, explained what those numbers looked like in practice. The project took 7 years, cost about $240,000, and started 143 replications. 53 were later invalidated because of protocol deviations, bad documentation, insufficient sample size, or other process problems.
This is what matters for your lab. Reproducibility often breaks in ordinary work.