Opinions: Reproducibility and revalidation: are they the same thing?
Much has been said on the reports from pharmaceutical industries indicating that a high proportion (from 60 to 80%) of published fundamental research cannot be reproduced. If reproducibility is indeed so low, the consequences on the cost associated with new drug development and translational research is obvious.
In order to limit the impact of this, steps must be taken at the project selection stage to ensure that good research is being moved forward.
In the translational medicine field, revalidation is a necessary and more invasive process than reproduction. In fact, the original finding is tested for robustness by increasing the number of points, subjects and models tested. However, the lack of reproducibility is the sum of a chronic lack of robustness of academic published findings arising from a complex variety of causes.
The high volume of scientific publications has been for long the only standard for career and funding in academia. The lack of “revalidation”, although not meaning fraudulence or incompetence, does stress the limits of what is currently considered an acceptable “peer review” process. It is time for academia and funders to add more relevant criteria to career assessment if innovation has to be an important result of the science in Europe.
EATRIS is organizing a session on Reproducibility in the Biomedical Sciences at the EuroScience Open Forum in Manchester, July 23 – 27. Experts representatives of the research infrastructures and scientists, funding agencies, policy makers and editors will meet to discuss the current situation and possible solutions to ensure responsible and effective research, restore the trust of society towards science and inspire the next generation of scientist.