Use of data
Issues around the acquisition, analysis and publication of data are part and parcel of science, but while these debates traditionally remained in the academic realm, they are now spilling out and affecting public trust in research and the reporting of modern science.
At the same time, analytical capabilities are expanding and the ability to study massive datasets is unlocking new frontiers of understanding and predictive capabilities. Publications in physiology often rest on real-world data collected from animals and humans, with all the variability and uncertainty this entails.
In order for experimental conclusions to be solid and reproducible, publications should include a full report of data gathering including a priori experimental design, uncertainty and caveats. Analysis should be detailed in full with justification for the methods utilised. Full experimental data should be made available to readers as a matter of course, unless exceptional reasons can be justified. New technologies and online hosting services mean this should also apply in Big Data studies.
Journals have a responsibility to ensure they have the capability to review the accurate and effective use and presentation of data in submitted manuscripts, and should use editorial policies to promote unrestricted access to experimental data. The skills of editors and reviewers should be supplemented by statistical experts to balance the reviewing workload.
The generation of knowledge through scientific efforts is fruitless unless this knowledge is accessed and acted upon. Policymakers have long been encouraged to base their decisions on available evidence, but these decisions should also be able to be scrutinised. When providing evidence or expertise to policymakers, they should be encouraged to give a transparent summary of the evidence used.
For more information, please read our full policy briefing on data and physiology.