Last week, the same editorial was published on the two top scientific journals, Nature and Science1,2, describing the intent of the scientific community to address reproducibility and transparency of scientific results.
Last June, a group of editors of scientific journals, members of funding agencies, and scientific leaders have met to discuss guidelines and principles for future publication policy to guarantee reliable scientific methods. The meeting resulted in the publication of a list of guidelines that the journal should follow to report preclinical studies (http://www.nih.gov/about/reporting-preclinical-research.htm). According to the guidelines, the journals should carefully check accuracy of statistical analysis use a checklist to assure a complete report of the methodology (standards, replicates, statistics, sample size, blinding, inclusion and exclusion criteria); the journal should encourage sharing datasets and software in public domains, be responsible of refutations, and guarantee the accurate description of all sources as well as check for image manipulation.
Sometimes it is difficult to incorporate all the information on the methodology in a paper, especially because of the word count limit. To guarantee a careful and accurate description of all the methodologies, some journal should revise the word count limit that induces the author to cut or to move a huge part of the methods section to the supplementary section. To encourage accuracy, some journals are already requesting a minimum word count for the methods section. To assure a good quality of the research published on both high and low impact journals, the new guidelines should be followed by all the journals and not only by a set of top scientific journals, otherwise affecting the whole research reliability.
These guidelines should not only be adopted by the journals but also by all researchers. Honestly, I thought that the principles outlined in the guidelines—statistics, blindness, sample size, etc.—were obvious steps when outlining an experiment. Obviously, they are not, since the scientific community had to meet and had to delineate them in an official document. The increasing number of retractions—the last most clamorous case of the Obokata’s papers published on Nature and then retracted previously this year—highlights the necessity to draft such a guide. To avoid future problems on science reliability, a training should be provided not only to the new generation of scientists but also to the old generation, who is mostly responsible of the recent scientific scandals. Unfortunately, scientists are not the only responsible for what is happening in science: everyone in the scientific community is guilty from the journals to the funding agencies.
I am still shocked that the people in the scientific community had to meet and draft these guidelines. Now it is time to slow down and to promote transparency and reliability instead than sensationalism.
1Journals unite for reproducibility. Nature. 2014 Nov 6;515(7525):7. doi: 10.1038/515007a.
2 Journals unite for reproducibility. McNutt M. Science. 2014 Nov 7;346(6210):679.
