Henry Lovett, Policy and Public Affairs Officer, The Physiological Society

https://doi.org/10.36866/pn.111.16

Research revolves around the acquisition and interpretation of data. New developments are continually enhancing capabilities in both of these activities, and the complexity of data that can be presented in scientific investigation is increasing massively. The onus is on the researcher to ensure that their data is up to scratch in terms of accuracy, completeness and accessibility in order to demonstrate their conclusion is reliable and make their research reproducible.

The Physiological Society has recently released a policy position statement suggesting how its Members can follow best practice to ensure good data. It also covers how its peer organisations can ensure the best standards in publishing, and how future audiences can use data effectively as evidence in policymaking.

Robust and reliable data gathering starts in experimental design. Studies should be large enough to ensure conclusions are valid, with statistical methods determining sample sizes. This avoids the waste of time and effort and, in in vivo experiments, follows the principle of reduction in animal use to the minimum number necessary to generate a statistically valid conclusion. Accuracy can also be dependent on the equipment and techniques used; these concerns should be addressed in any publication using the data.

Subsequent analysis of experimental data should be transparent, using appropriate statistical protocols set out in the initial discussion of the hypothesis. If experimental data is being used to derive a predictive test or scale, care must be taken to understand the range of ‘normal’ values across the potential diversity of subjects the test may be applied to in future.

Awareness is growing throughout science and the public concerning the issues of reproducibility affecting published research. The necessity of accurate and detailed description of experimental techniques and conditions is well stated by former President of The American Physiological Society Peter Wagner, who writes:

‘It is not a problem per se for two purportedly identical studies to disagree, because if both are adequately described, the reason(s) for discrepancies can usually be found, and this alone may considerably advance the field. It is when the outcomes are in doubt because of poor descriptions or statistical errors that we waste our time, resources, and dollars and put future research (and even clinical care) at risk.’

As ensuring reproducibility is such a critical aspect of research, The Society recommends training for early career researchers in how to design reproducibility into their experimental techniques and demonstrate it with transparency in their write-ups.

When publishing, it is becoming more common for funders and/or journals to require that experimental data is made available to readers. Some go further and mandate particular techniques when using and reporting statistics. We recommend that other journals replicate the model of The Society’s journals by appointing a Statistics Editor, who is involved in the review process of submitted manuscripts to ensure statistical accuracy and appropriateness. This means both that the accuracy of data analysis can be checked by an expert, and that the reviewers can concentrate on the design of the study and the conclusions.

These and other recommendations are explored in more detail in the position statement, which can be found in the Policy section of the Society website.