Posted: November 7, 2014
Back in 1958, the prominent British statistician Sir David Cox summarized five requirements for "good" experimental design. An experimental plan must be unbiased, be precise, have a wide range of applicability, be simple (or at least not too complicated), and have the ability to calculate uncertainty.1
The same requirements apply today, but in many cases, the frenetic pace and shifting demands of modern biomedicine have eclipsed attention to these important details so integral to the scientific method.
This week, as co-reported by the journals Science and Nature , a coalition of publishers announced a common set of Principles and Guidelines in Reporting Preclinical Research. The guidelines promote rigor and transparency in general, and recommend how to describe experiments more completely by noting subject characteristics, including sex. The announcement follows a June 2014 meeting of editors representing over 30 major journals, representatives from funding agencies, and scientific leaders, convened by the National Institutes of Health (NIH) with Science and Nature.
In 2014, examining sex and gender influences in health and disease has come to light as an important focal point throughout NIH . Studying both sexes is good science, and it is essential for advancing the health of women and men. Studying both sexes fulfills several aspects of good experimental design: mitigating bias and broadening applicability. Last, but not least, studying both sexes is a powerful tool for turning discovery into health. This is the NIH mission distilled into just a few words.
Why does all this matter?
Consider an all-too-common scenario in biomedicine. A scientist publishes a paper reporting aggregated data: the results of experiments with males and females combined or sex is unreported. Lumping results together can mask differences between the sexes. In the worst-case scenario, a positive response in one group (sex) combined with a negative response in another group (the other sex) results in "no response," due to the effects essentially cancelling each other out. Without complete reporting, important data can be left hiding in plain sight.
How can we uncover thought-provoking experimental data and learn the most we can from investigations? One way is to ensure that researchers report what they do very clearly and thoroughly. Scientific truths that are masked by incomplete reporting or suboptimal experimental design can't inform health if they never see the light of day. And colleagues can't repeat and build upon experiments without complete information. The publishing world is a critical link in translating discovery into health. The new guidelines offer an important opportunity to go back to basics.
1Cox, DR. Planning of Experiments, John Wiley and Sons, New York, 1958